The general consensus amongst audiophiles is that the quality of a router or Ethernet switch affects sound quality. Is it true? Have a question you want to ask Paul? Go to www.psaudio.co...
When streaming the data isn't being sent just in time as it is played, instead, it is being sent in chunks of data that get queued up on the streaming device (clock and jitter from buffer to DAC is only related to streaming hardware). Unless the network connection is so bad that it can't keep the buffer full then there is no difference. If for some reason you are having issues then you can enable the QOS settings in your current network hardware.
All TCP/IP based streaming systems buffer the data at the endpoint. There is no timing variation in the data stream, because there is no timing in the data stream; it’s just a sequence of bytes. It might as well be written down on post-it notes at this stage. Also, the transport protocols reject damaged packets, so noise at this stage is not a significant factor (it will reduce your download speeds, but even the most basic of networks have far more capacity than an audio stream needs). The timing is constructed at the DAC; and if you have a good DAC, then you will get good timing. Noise going into the DAC should be well below the digital switching threshold (which is the whole point of digital audio - each sample can be carried thru the system perfectly with zero variation) so spending inordinate amounts of money on “audiophile grade” digital connection hardware & cables is pointless... regular decent quality gear works just as well. Also, sorry, no... computers aren’t excessively prone to noise sensitivity. They’re actually quite resistant to it: typical line-level audio signals might be 1 volt peak to peak; but a digital signal - as used in the computer - will swing from a low level of less than 0.1 volts up to 3.2 volts or higher (older standards go to 4.8 volts or so typically). You can hear mili-volt noise in an analog signal quite clearly; but that noise level on a digital signal is going to be several thousand times less than what is needed to change the signal from a 0 to a 1 (or vice-versa). Use good quality, but regular digital components and you will be getting the best possible performance out of your system; because it will be 100% bit for bit perfect. Note that if you do have that kind of noise injection into a computer that will flip bits; the result won’t be limited to just a few low-order bits in an audio stream being flipped; the whole machine is going to crash (if a value is off by even one bit, the entire execution path goes completely off the rails). What digital gear can do, however, is create electromagnetic emissions that can couple into the analog side of your setup; so using balanced, shielded cables run at right angles to digital cables where they have to cross; and putting analog gear in appropriate metal enclosures is a good idea. Good analog gear will have power supplies that filter out noise coming in that way also; but if needed a line choke can be added. Most good gear will already come this way. Also, the cables don’t need to be ridiculously expensive either; just good ones of the same grade as is used in professional recording studios (after all, that’s how the music was made in the first place).
Nope you are wrong, it sounds right in theory but a PC is an almost endless amount of noise that needs to be taken care of. Switch is also a bit noisy, everything creates noise especially all out PSU and LPSU's. Bit theory only belongs in the network world. Science with numbers does only half belong in the audio world. Have you ever noticed why that most music systems in the world sound so mechanical? Because they believe in your theory and only treat it as a mechanical field. What I wrote is the truth. But that's the problem with the world, either you are the one or the other extreme, being in the middle partly agreeing partly with both sides of the equation is not popular. Audio is an art, and most engineers in this world treat it as a science, and that is why their sound is close to perfect, but has no soul. "Peter Qvortrup said it perfectly himself: You cannot measure good sound!"
Real HIFI Help - I am quite sure that you are going to disagree with me; and I’m also quite sure that I’m not going to change your mind with a CZcams comment. But I’m bored, so here goes: The path a sound takes from instrument being played by an artist to your cochleae sending nerve impulses to your brain can be divided into distinct stages; which will of course vary depending on the recording in question and on how you are listening to it. For the sake of discussion; let’s say it’s a classical guitar recording: strings & body vibrate, inducing vibration in the air, inducing vibration in the microphone capsule, which puts out a voltage proportional to the air pressure, which is amplified and run thru the recording equipment chain - and here I’m skipping a whole bunch of stages because, well, this is a CZcams comment, not a textbook 😉 - and ends up as a .wav file (again, just for the sake of discussion; AIFF, or any other uncompressed format would do as well here) on a server. At this point there is only the noise introduced by any stages in the recording chain from the microphone thru to the file on the server (and of course, this should be self-evident). Now, this is a digital file: the whole point of which is that it can be copied from one place to another exactly, with no variation. What is less obvious; and what I was trying to convey above in my first comment; is that there is NO timing in the file. None. Files cannot, by their very nature, have jitter - it’s a concept entirely foreign to them. They are simply sequences of numbers. The consequence of this is that as long as the copy is true (I.e, the sequence of numbers you put in on one end came out the other end identically - and there are mechanisms both in the hardware and software to make sure that happens 100% of the time); by definition you cannot add noise to the file... and the consequence of that is that it doesn’t matter what you send it thru, by definition you have the exact same thing you started with; with zero exceptions. The question was if spending extra on a fancy switch or router made a difference. Well, given that the network device in question operates on only the digital file stage of the chain, it should be self evident that as long as it’s not defective; then it cannot have an effect. Once it’s reached the stage where it leaves the digital realm and becomes an analog signal again; noise becomes an issue again - both in amplitude and in time. This is why your DAC should be in a separate box; and have a good, stable clock generator. Re: bit theory. This is fundamental to - and therefore applicable to - any place information exists in digital form. As long as the audio is digital, it applies. As for science with numbers, it is by definition concerned with, and fully applicable to reality in general... and last time I checked audio was very much a part of reality. Saying that a system “sounds mechanical” is too vague to consider... what, exactly, is meant by that? Yes, audio is an art. What it isn’t is magic: like all art, it is built on the fundamental principles of mathematics, physics, and engineering. There’s nothing going on that is non-understandable. You cannot truly master any given art until you fully master not only the what, but the how and why - and this applies universally; from as diverse arts as painting to ballet to hi-fi audio (and yes, I contend that the “great masters” of the past, and often the present; even though they are generally considered to be the best in history, are not true masters simply because they didn’t have access to the fundamental knowledge that underpins their art). As far as measuring universally “good sound” - you cannot measure it because in this context, it doesn’t exist: “good sound” means “sound that I like”... and people vary. What you can measure, in principle and to a large extent in practice, is what is actually physically going on; and from that extrapolate that any given setup’s characteristics are (or aren’t) ones that are popular.
I opened this thinking I was going to hear some snake oil about switches. I'm an IT professional and have been for over 30 years. I opened this thinking my blood pressure was going to go up, thank you for the realistic answer.
Same here. IT professional of 2+ decades, braced for some crap about having to use "audiophile grade" network "interconnects".... Thanks goodness some sanity prevails.
If there is someone I trust on all things audio, how it works, what make a difference and what does not, it’s you. You just seem incredibly knowledable and honest and you just seem to speak how it is without any bs. Thank you Paul.
What? You mean I also don't need to spend $$$ dollars network cable by Audioquest? Surely ONES and ZEROS need special wire with special shielding and special connectors to stop them from becoming TWOS and THREES.
Some network cabling (I know you're kidding, but I thought I'd mention it anyway) is heavily shielded. Not for getting data wrong per se, but for speed. The better shielding results in faster transmission due to lower packet loss.
@@aweidenhammer Yeah but you don't have to spend 50+ dollars per meter for "special" shielding. Here's a is the Audioquest's VP giving a seminar on computer audio... czcams.com/video/Gjsc_LTurl0/video.html Really instills confidence in their products, doesn't it? PS Audio says Audioquest's Dragon power cable is significantly better than their own, but nobody provides evidence it's better than a 40 dollar shielded generic power cable to justify its 4000+ dollar per meter price tag...
No, but we just outfitted our office with cat6a, which is more expensive for the shielding. Not saying it makes any difference to audio quality though. What Paul fails to mention is as long as the signal is in the digital realm, it has packet redundancy and error correction so the data is a perfect copy. It's once that data is set to a timing clock, sent to a dac and converted to analog where things get tricky.
If he's streaming audio through Comcast, Comcast is already the source of 99% of his jitter. At my last job, I supervised IT at small office sites which had dual WAN routers with both Comcast and Verizon FiOS. Here's what I found: The core of both ISP networks was fairly stable and comparatively low jitter (as measured by ping and traceroute). However, the last hop of Comcast increased the jitter greatly, while FiOS was low jitter all the way to the customer premises. I came to believe that this was because the Comcast last mile bandwidth was oversubscribed, while Verizon's last mile bandwidth was undersubscribed. That is, Comcast did not have enough bandwidth on their coax when lots of customers were using it, while Verizon had far more bandwidth, so all of their customers could be running at maximum and still have bandwidth left over. What I did was set the dual WAN router to divide up the traffic so that all the non-web traffic went through Verizon, which included VoIP where low jitter was necessary. All of the web traffic went out over Comcast, because they had better peering which led to faster throughput. So what about music streaming over IP? The streamer has a jitter buffer into which the audio data packets are collected, and they are clocked out with a stable, low jitter, local clock. So any jitter which occurred during transmission is removed by the jitter buffer at the streamer. In music streaming, the jitter buffer can be several seconds to 10s of seconds long, so it can even deal with situations where transmission is interrupted for several seconds. Now that's jitter! The reason VoIP needs a comparatively short jitter buffer is that a two way conversation cannot tolerate long delays. If much more than a quarter of a second jitter buffer is used, the callers need to start saying "over" so the other knows when to talk. I should also point out that part of the confusion is because audio and IP people are talking about different things when they use the word "jitter". Audio people are usually talking about waveform jitter, which is measured in nanoseconds or picoseconds. IP people are talking about packet jitter, which is measured in milliseconds or seconds. A streamer has to have a jitter buffer which is some number of seconds long to take out the packet jitter, but audio data needs to be clocked out by a local clock signal whose waveform jitter is measured in nanoseconds or picoseconds. So the answer is no, a low jitter ethernet switch will make no difference to streaming audio quality.
That's nothing to do with jitter in this context. Jitter is variability in the clock signal, and in this case the ADCs generate their own clock signals. What you are seeing is variability in the rate of data transmission, and providing their is a sufficiently large buffer it has no impact whatsoever on the quality of the musical signal. If it goes beyond that level, what happens is that the playing of the music stalls until more data arrives.
Only real time audio needs specific switches or protocols. Such as AES67, QLAN, etc. These switches depend on a clock to stay synced and have less than 1ms of latency. Streaming audio has no latency concerns and is buffered and reassembled to sound proper. The assembling codec is the most important.
Right on point. TCP/IP is robust and streaming services tolerate packet loss in transit. Each packet has an ID and a checksum. A switch that affects sound is one that is not functioning correctly. "jitter" is a complete non-issue with any properly working network. Cheers.
The "if it does make a difference then it isn't functioning correctly" argument is very interesting. It seems to be the anti-audiophile strategy of admitting you are wrong without admitting you are wrong. You start off saying it cannot possibly make a difference, then reluctantly agree that it does, but still maintain that the ways in which it does make a difference don't actually matter and that despite it making a difference you are correct in asserting that it doesn't.
I have a sotm connected to an Aurender N10 streamer. The difference was enough for me to spend 800eur +500 for a power supply. I’m pretty shure unbelievers never tried it. Or never tried in a proper system.
Aurender n10 has a buffer which coppy all the data it receives no matter where it comes. You are always playing from the buffer. I'm pretty shure You are imaging changes where it's only snake oil
Agreed a sensible answer everyone knows the only thing you need to worry about is making sure all your fly leads and patch cords are red to ensure the signal gets from point a to point b as fast as possible
Everyone, so far, seems to have missed the point. Bits are bits and networks accurately transmit them, otherwise our downloaded programs wouldn't work, because they MUST be bit perfect. So data accuracy is not the issue. Timing is. If there is jitter, if the pulses aren't perfectly regular, then that can affect sound quality. The trick is a reclocker, which fixes this and some are quite inexpensive. Just before your digital goes into your DAC, reclock it!
jonc113 The data on your Ethernet cables going to your audio player are entirely asynchronous to the DAC clock and any “reshaping” is total utter ignorant nonsense. The data is buffered in the player and as long as it has enough data (not starved) the player and DAC are happy. The player pulls the data from the server in blocks btw making this topic even more idiotic. This is scandalous snake oil nonsense.
When you are talking about switches or anything network related you do not have "sound quality" as there is no sound signal hence timing is irrelevant. Timing only comes into play the moment the digital buffer is being addressed for conversion which would be in the DAC.
You do realize that as a frame enters one port on switch and then exits another port on the switch it's regenerated. CAT5 at 10/100/1000Mbps is good for 328ft/100 yards. I can simply extend that by placing another switch and adding another 328ft. Rinse repeat. The reason this works is that the signal is regenerated. Switches also have buffers. At default they are mostly FIFO but we can change this with Weighted RED. Also our playback systems are not real-time and not delay sensitive. They are asynch. Data can be delivered ahead of playback.
Matt Noble Yes absolutely pure ignorant snake oil BS that deserves zero respect. People selling such stuff at an overpriced MSRP arguing it “reshapes” the timing of audio should be ashamed of themselves.
I'm a network administrator. 1. Those "audiophile switches" are a waste of money. 2. As long as you're cable modem/router combo is working correctly under a normal load you won't be able to tell a difference streaming audio. 3. Now if you're home network has a traffic problem , QOS ( quailty of service , aka traffic management ) settings will make a big difference streaming wise. I have a FiOS router ( I disabled the WiFi on it ) but I added a cheap 16 port unmanaged gbit switch ( not a hub ) + Google WIFI. Each node has an Ethernet link back to the switch, master node it's between the router and the switch. I have a lot of devices at home and I do a ton of HD video streaming over WiFi. I get wire level speed over wifi from anywhere in my house. That was like a $350 upgrade. Adding a managed switch would have only been $50 more but didn't need those extra options. I've worked with cheap netgear switches up to $30k+ Cisco switches. And yes I played with a few audiophile switches too, save your money. My home switch is Netgear prosafe jgs516 listed around $100 on Amazon.
I did something similar with my gig fiber connection- all into a Cisco Catalyst switch, then split into three separate WiFi access points which are located in different parts of the house. One is for me, one for the kids' gaming systems and the other is for the cameras and alarm system. Splits the loads nicely and I have QOS set up in the Catalyst to make sure I get preference over the kids since I paid for it all. We can have over forty devices on the WiFi and 11 devices wired and streaming still comes through with no discernable packet loss as far as audio goes.
There is certainly not a difference on the digital / signal side, but if you connect an (any) electrical device to your system it might introduce some analog garbage to the ground level. Even a lamp on the same socket as your system or touching a metal surface of any your devices might make a difference.
More importantly, if that was such a concern, it would make far more sense to run a fiber optic cable from your switch to your streamer to completely isolate it from any electrical noise going through a copper Ethernet cable.
@@go2yanks that is what drives me the most wild. I've seen people who make audiophile an identity rather than a hobby with tens of thousands USD in a rack for a single listening station and it''s all overpriced garbage using meaningless buzzwords. At the end of the day if they were *that* worried about electrical noise, they would use fiber (which can even be incredibly cheap to do!)
@@kumakell I don’t actually have any issue with spending a ton of money on components regardless of how they may measure. My big issue is when companies start actively lying and deceiving people with deep pockets who don’t know any better. Synergistic Research is the biggest offender in this regard but they’re far from the only company peddling that particular variety of snake oil.
@Dan My streamer is already galvanically isolated from the rest of my network using fiber optics. Going to their copper based switch would be a lateral change at best while costing several thousand dollars more than my implementation.
Paul, while your explanation about data transmission is correct, it does not address the issue of electrical or other noise generated by the router, switch etc…polluting sensitive audio components downstream. 2 quick ways to test/verify this…..1) Swap out the switch mode power supply on the network switch for a good quality linear power supply (LPS). 2) Try a network switch that has a fibre optic SFP input and isolate it from the noisy network by connecting it to a fibre media converter using a length of fibre optic cable (the media converter is also connected to the wifi router using an ethernet cable). Ideally all of these powered by LPSs. The proof is in the listening and the extent of the impact made me question why my very expensive music server (which already sounded fantastic) hadn’t/couldn’t fully address/resolve network noise issues. Please have one of your electrical engineers check this out and let us all know. Thanks.
I am sorry but I can not agree. In my system the switch made a huge difference. And after adding the dedicated power supply for the switch from the same manufacturer the improvement was just as big. Paul just says it does not make a difference, but does not tell if and what they have already tried. If your stereo is your hobby trying new stuff is part of the enjoyment. Most of the devices can be ordered risk free with a money back guarantee. So maybe try it for yourself. If you can not here an improvement you have assured yourself and can tell everybody "I told you so in the beginning" . If you can hear an improvement, keep the device if you like and enjoy.
i was just into the subject, have found several sources on the internet that it can make a difference (in audioquality) because a switch filters a lot of noise from cheap ac/dc devices and computers.
@@thepuma2012 just FYI I am using a Silent Angel Bonn N8 switch and the Forester F1 power supply. Streamer the switch is connected to is a Linn Akurat DSM
You indeed have to have decent equipment to Heat a difference. For what i read the music opens a bit more. My equipment is not high end. Bit i found a cheap one which could make a difference maybe. My audiolab amp is maybe Just revealing enough to try
Switches can diminish noise isn't it? They can cause noise too with bad power supplies. Yesterday I tested 2 ethernet cables and they sound different. The data will be correct in both so I guess they do something with noise
That either didn't happen, or you defied lottery-winning odds to have a cable that provides a solid link yet somehow drops packets. With cables, it's mostly works/doesn't work- although shielding makes a difference when in an electrically noisy environment. It's 1's and 0's and if they all (or mostly all) get to the destination, it sounds the same regardless of the cable.
@@4879daniel I always wondering why there are no "videophile" Ethernet cables (or switches) on the market. More contrast, clearer colors, smoother movements and even the actors would play better.... 🙂
Notorious truths: 1. $20 switch and $640 switch - no difference. Amir from ASR measured it with the most precise $100k measuring device on earth. No difference in the spectrum. 2. No difference means if one sounds good then both sound good. If one sounds bad then both sounds bad. 3. Noise alone does not matter for jitter (time distortion) IF BANDWIDTH IS INFINITE. 4. Bandwidth alone does not matter for jitter (time distortion) IF THERE IS ZERO NOISE. 5. BOTH noise (low frequency and high frequency) AND bandwidth constraint together impact jitter. 6. If you put any filter on the signal path you are going to constrain the bandwidth. Which is bad for jitter (read point 5). 7. The noise should be exterminated as much as possible in the music server, before it reaches the stremaer and DAC. For once it leaves the server it is all too late to deal with it later. 8. Bitrate matters: higher the bitrate the higher the impact on bandwidth due to more time needed for processing the stream
It’s very strange that everyone claiming the switch makes a difference, always have some anecdotal references. None of them are able to show the results in measured data.
The first part is correct. As long as we are talking network equipment, including the cables, up to the streamer, then it CANNOT affect the sound. The datapackets are indeed assembled in the streamer, if it misses a part, or fails a consistency check, then additional packets are retrieved. And btw, this is done BEFORE the Application, layer of the osi model. So it is even independent of which streamer you use. What happens after that is up to the streamer. A "noisy" switch, whatever that is supposed to be, cannot affect the completed file in a streamer. 'Hifi' switches and network cables are an absolute scam. Stay far away from them.
@@borisbobowski628 the point is, datapackets sent through a network protocol will arrive in your streamer exactly the same as they are stored 4000+ km away on a server in a datacenter. It's just how this works. Error correction, acknowledgements, reaquiring missing or corrupted packets etc. If this wouldn't work, we wouldn't have the internet like we do now. It would just crash. So, regardless of the network cable you use, or the router, if technically sound, it just doesn't matter. Once the file has been assembled in the streamer, then that part is over. It is then up to the the soft and hardware of the streamed. Btw, there have been quite a few people explaining and showing why the whole powercable etc story is just as much nonsense as hifi network cables. A LOT of placebo effect going on. Noise is filtered out once a powersupply transforms it. Those filters are very cheap and do the job. Of you want to filter 'noise' at 220v 1 amp etc. You need extremely powerful and expensive filters. Which are then rendered obsolete by the ones i mentioned prior. Its a waste of money really.... now analoge interlinks and powercables, that's a very different story
@@CyberBeep_kenshi so even on normal Websites you and i notice different Speed, in Terms of high res Files the Timing is absutly Key If you wanna have a Perfect Decoding going on. Tbh PS Audio is Not a really high End Brand, Their Streamer are Not good and they have No clue. Every source has a different quality and It makes a huge Difference If you Drive a dac with a Smartphone/MacBook or even desicaded Streamer. Ofcourse you Need a great Setup. Even my Girlfriend that is Not a active listener can Hear when a Switch is connected or Not.
@@borisbobowski628 the speed of a download has no effect on the final file. And so it cannot affect the sound. Now, if you go through wifi and do live streaming, that is a different matter. It uses udp instead of tcp ip. Which allows certain packets to be dropped. But when you use i.e. qobuz, it will copy a flac file onto your streamer, which is then played. How it gets there isn't relevant. The file is 100% the same. To give you a small idea of what packets go through (yes its my job :-) ) Storage center, (fiber, patchpanel, fiber patchpanel, fiber) = fpfpf, database server, fpfpf, application server and or loadbalancer or both, fpfpf, switch, fpfpf, coreswitch, firewall (at this point you are still in a customer cage for a customer in a datacenter. Then its routers, fiber or copper, router, router router, satellite or ocean fiber, the more routers and finally in your street. Then into your house, modem and THEN finally in your switch. We are talking easily 4000 km depending on where you live. But at the end its the network protocol that ensures the file is exactly the same :) internet really is a wonderful thing :-) Then to a central connection to the outside, riuter
@@borisbobowski628 regarding ps audio, i never heard it irl. But its a lot of talk and big boxes with tiny circuitboards. Never a good sign. To be fair, i have a blue node as a first go on set, but i use the Dac on my pre amplifier. So far no complaints:)
I dont see how it would. Its a digital signal, the bandwidth you get to your place is your bottleneck. Now with broadband so available its not much of an issue. Its not like an analog signal thats.. all contained in a local area to an extent. Thanks for clarification Paul!
I believe - without being an audio engineer - that these switches address potential noise being introduced by standard issued modems and routers, not how the data is delivered. Alvin at Vinshine says it best: "at the very least, it fixes audiophile OCD".
Alvin's SW-8 and I believe the SW-10 both use this Netgear switch, the GS108Ev3. They are $39 on Amazon right now, and boy do they work! I have since installed three of these switches in the house to decrapify my internet. The damn things work! On a side note. I do own the LHY SW-8, it is a fine switch. BUT! Save your money and get the $39 switch and a decent 12v power supply, like a Meanwell or one of Alvins LHY linear units. I cannot tell the difference between the SW8 and this less expensive combo. The improvements in PQ and SQ were quite profound, either way you go. Piggy back two of these, I have the $39 switch feeding Alvin's $500 switch, and there was even more of an improvement, subtle, but there. I have three of these $39 switches in the house, as they really cleaned up my noisy internet feed.
@@michaelwright1602 Somehow can't remember asking for financial advice, lol. The SW-8 is using a Linksys Chassis with upgraded components. No relations to a Netgear switch, which serves its purpose when achieving a noise-free signal is not a concern.
The only place where switches make a difference is in live sound where low latency can not tolerate large store and forward delays. If you run digital stage boxes with your digital audio console, with digital hearback personal mixers for the band, then yes, switches make a difference, and not "Audiophile Switches" but low latency switches for commercial sound such as sending 48 microphone channels and 24 monitor returns at 48KHZ 24 bit audio with less than 17mS of delay. Running a single stereo DAC for a streaming audio service with several minutes of stream buffered locally, won't care if your router store and forwards a packet 600mS late.
They don't matter from the audiophile point of view as long as the right ones are chosen to build a properly designed network that is able to carry the expected traffic with no congestion. And if that's not the case, then they matter as much for audio streaming as they do for any other kind of traffic.
Most audiophiles that have money have been stuck to the analog world so they think that digital audio works like old LP players were flutter and wow effected the sound. In digital realm if something is bit perfect, it's just that. And with a working network connection things are bit perfect. Even CD players have error correction algorithms which negate random errors in playback. The cd must be really damaged for the bit rate to drop low enough to cause a drop in bit-depth. To put things into perspective, imagine downloading a 1Gb movie or an ISO image for windows install for example. If even one bit transferred in that download is not bit perfect to the original, the end result is corrupt. Then ask yourself how many times your downloads came in corrupt? Never, unless you have literally broken hardware.
@@dingdong2103 Well, it’s because in the end, we listen to an analog signal that air-pressure waves are. It also doesn’t really matter whether a digital audio file as a whole can be transported without being corrupted, but how we sequentially feed data it contains to the D/A conversion process. All of a sudden, clocking and a constant data flow do matter. Providing all that costs a lot of effort, which in turn, introduces variable noise. How much of it propagates to the audio channel is another story, and Paul has elaborated on that to a great extent over the years. This is also where separating different components and processing stages becomes essential.
@@themarcinmm4840 Incorrect. I suggest you study the AD/DA process a bit further, you'll learn that also there are built-in error corrections and buffers which make the signal totally impervious to slight delays or noise. The end result is just as bit perfect as was the transmission.
@@themarcinmm4840 "Well, it’s because in the end, we listen to an analog signal that air-pressure waves are." Correct "It also doesn’t really matter whether a digital audio file as a whole can be transported without being corrupted, but how we sequentially feed data it contains to the D/A conversion process." Correct "All of a sudden, clocking and a constant data flow do matter. Providing all that costs a lot of effort, which in turn, introduces variable noise." Wrong It is very easy and inexpensive to ensure that the DAC is "feed" with constant data flow by adding some digital memory working as a buffer in front of it. Think about portable CD Players or CD Players in a car - if there is a heavy "bump" from outside the laser looses track and needs some time to position again but this does under normal circumstances not interrupts the playback as there is buffer which stores severals seconds of music in advance - given the laser enough time to reposition. And as the player reads the data with an much higher rate than needed for the playback the buffer is soon filled again - ready for the next "bump".
@@themarcinmm4840 "All of a sudden, clocking and a constant data flow do matter. " But for clocking we can use quarz oscillators and feeding an DAC with constant data rate using such a clock is no problem at all today there even an external USB Harddisk (not even an SSD) can easily transfer 100 MByte per second.
Packet loss - read up on it. In audio or video streams, between 1% and 5% is considered acceptable, by some. Paul's description hits the right note and understand that an audio stream may make a number of "hops" in long distance paths (open a command shell and run the "tracert" command - a variant of ping - to see how many hops are in the path of your audio stream.) Each hop involves an active component like a switch, gateway or router - where packets in your stream are buffered a fraction of time. When there is no packet loss and the receiving end has no issues, the active components in the path should not influence audio quality in themselves. The occasional packet loss, however, in some cases may cause the receiver to ask a resend to the sender and you have to wonder what happens to your stream - there may be dropouts. At some point, the audio stream may arrive in (A) your network computer that feeds your DAC, or (B) a DAC that is capable of receiving a stream itself. here, buffering plays a role again, essentially to prevent being bothered by dropouts from packet-loss/resends. If a difference could be audible, then where do we expect it? In case A in a DAC that gets fed its audio samples (TCP/IP packets unwrapped and the contents put into sequence in a buffer and next presented, likely over something like USB) in a synchronous way, where the PC or Mac determines the clock of the audio samples. This can be addressed by an "asynchronous" DAC, which does a form of bus-mastering in that it asks for samples and buffers these. That DAC needs to put the samples on a clock again. Here, wired networks give the best reliability with the least drop outs, as a general rule. But (I use shielded Ethernet) there likely is some ground connection between network components in a cabled situation so if the digital link to the DAC-core is not floating relative to ground, then stray electrical signals can seep through the DAC into the analog section. This, though, will never happen in a perfect world populated by perfect connections, like we can see inside PS Audio ;). In case B, the DAC integrates a PC-like network section and the problem under A needs to be solved between the stages in B's design. As to packet loss, the question is, when this happens, how the DAC handles this. The original CD playback system was designed to compensate lost bits that could not be retrieved from the CD by filling the blank with samples extrapolated from the previous ones. If the stream-receiving DAC does this with packet loss, then the occasional drop out is not heard. As to Quality of Service (QoS) which is a mode you can set on your switches or routers, it is implemented generally in such a bad way that it may be detrimental rather than beneficial. Yes, as Paul says, a switch is a computer. And like a Windows PC, its performance may already degrade above 3% CPU load when it has to do smart things (protocol conversion or QoS). So, in the network we have to look for packet loss and how this is handled. Then we have to look at the step from network to DAC as well as inside the DAC how this works. We could argue that wireless does not have the inherent grounding issue, which is true, but a component that is sensitive to some leakage from the digital front end to the analog backend is already inherently sensitive to this problem in itself (if it does not transfer between stages over floating links and the individual stages have grounding issues.) If we cannot imagine something, then this does not imply our conceptual models we use for explanation are valid, or relevant. It just means we cannot imagine it. And have to be careful with opinions. My challenge: try to refute these claims in triple-blind experiments, set up in a way that can demonstrate these points I make here. Use extremely revealing content (no studio mix, no multichannel recordings) with exceptional real spatial clues of a programme with width, depth and height. Invite a musical blind person who can tell you where the windows in your room, are to join in on the test.
SO well put, bravo! I'm a retired network engineer and designed networks for national security systems and for some of most cutting edge agencies in the world. You explained this very, very well.
"So, in the network we have to look for packet loss and how this is handled." Packet loss were so rare today (except for Wifi under some circumstances) that you can stream for months before even a single package is lost. And it is well known how such an error is handled. The TCP/IP stack detects the error by verifing the checksum and requests a retransmission of that packet. So a packet loss does not mean that any data is lost. "My challenge: try to refute these claims in triple-blind experiments" Wouldn't it be more reasonable if the manufactors of such "audiophile" switches would proof that their devices really have an impact on the sound quality by either performing measurements or triple-blind tests? "The original CD playback system was designed to compensate lost bits that could not be retrieved from the CD by filling the blank with samples extrapolated from the previous ones." That is only half of the truth. The audio CD standard defines a special encoding scheme (based on Red Solomon codes) to ensure that nearly most wrong bits (either to reading errors or scraches on the CD or even errors which were already on the master used for pressing the CD) could be fixed perfectly. Even sequences of errors like unreadable bits in a sequence due to a scratch) could still be corrected to the original data. Only if there are so many errors that this error correction could not fix them anymore then samples are extrapolated (or even better intrapolated) so minimize the hearable effect.
Paul, anytime you’re going through DFW airport, you’re invited to my house about 15 minutes away to listen to my system streaming with and without an EtherREGEN switch. If you can’t tell the difference in your system, there’s something else going on. The difference is far from subtle, especially with percussion.
I was sceptical about audiophile switches too so I bought a cheap Teradak one with txco clock and linear power supply of Ali Express and immediately noticed a big improvement. Bigger soundstage, better texture,more bass definition and more sparkle up top all for around $300 well worth it. Sorry I can't post a link but everytime I post any good product that's not PS Paul deletes it.
2028: PS releases audiophile switch, "Hey we admit that we got it wrong, digital is more complex than we understood, and and our PSwitch is the best in the market. It'll elevate your system"
Thanks Paul, good video and you probably just saved some people a heap of money as those audio grade switches are hideously overpriced for what sounds like negligible to no benefit.
Not so fast Paul … surely the Paul on the video didn’t plug a SoTM switch on his system. A topical case of someone that didn’t do his homework. I advise anyone who has this same disbelief to test drive one of these switches before make a stupid comment or vidro about the subject 😉
@@crventura I’m an advocate of “ Don’t nock it until you’re tried it” so fair enough. I must confess that I rarely stream music from streaming services. But I have ripped my digital collection to a Drobo NAS and that would probably benefit from audio grade switches and network treatment. From CZcams I find Hans Beekhuyzen more an authority than Paul M on this subject. I do however only listen to vinyl or CD’s for critical listening. Thanks for your comment and since posting that comment I have learnt a lot more.
@@paulpavlou9294 "From CZcams I find Hans Beekhuyzen more an authority than Paul M on this subject. " Hans Beekhuyzen talks a lot but his understandig of digital networks and protocolls is very limitted and sometimes totally wrong. He never proof his claims by measurements or double blind tests but instead he believes uncritically what manufacturers claim. Ultimately, he is still trapped in the analog world where every noise in the signal chain had an impact on the quality which was true - but is no longer valid in the digital world.
I spy a SADIE audio workstation. Great to see hardware like this still in use. I use one all day/everyday even though it went native yonks ago. Amazing piece of kit.
It was hard for me to believe, but I found that using a Mac Mini as a source connected to a DAC by either optical or USB sounds better if I am connected to my wifi router using 5 ghz rather than 2.4 ghz. The only thing I can account for the difference is less interference on 5 ghz from other nearby wifi routers on the same channel, because 5 ghz has less range and doesn't penetrate walls very well. It is possible, if there is no other wifi than yours reaching the computer, that there would be no difference at all. My findings were vindicated when I later read in UHF Magazine that they found the exact same thing - 5ghz yielded better sound with their Mac mini. They had no explanation. I know this isn't an ethernet switch, but Paul did mention wifi as well, saying he heard no difference. It is very likely that no neighbouring wifi reaches PS Audio's listening room and therefore the signal is free of interference from another router on the same channels.
Wi-Fi is a whole layer of problems added onto wired ethernet. But, in theory, the problems of the network transmission channel are completely removed by the streamer's jitter buffer.
It's amazing. The presenter doesn't seem to have tried different cables/switches himself. The people commenting miss the key point of the video and most haven't made comparisons themselves. People, you are looking to hear something from youtube that supports what you feel must me true. Stop doing that and try it for yourself. In my situation a self made cat 8 cable vs a cat 5 cable was a clearly audible difference. Unexplainable and confirmed by blind listening test with audio buddy. Upgrading my network switch to a Bonn 8 was a huge difference, I mean literally huge. Perhaps it is due to eliminating noise entering the streamer, perhaps something else. Whatever way, it was the best money I spent on my system in a while. My message: please try it for yourself and stop all this online bitching, it's exhausting.
I had the same experience with Uptone Etherregen, the improvement in sound quality was clearly audible, and no, that is not a bias effect or anything like that. Clearly there are things going on here that we simply do not understand yet.
@@dfpurton you heard here first folks! Don't listen to hifi before you buy it. If it measures great it must be great! Daniel, you are missing out a lot with that mind set
"In my situation a self made cat 8 cable vs a cat 5 cable was a clearly audible difference." If you make the effort to build a self mde cat 8 cable when your expectation that this will sound better must be very high. That is known as expectation bias. "Unexplainable and confirmed by blind listening test with audio buddy. " Unexplainable things needs a little bit more than a self performed blind test with your body... How was the test protocoll? How many rounds did you perform? How high was the percentage of "hits" / correct guesses? Was in each round determined by a random generator which cable should be tested in the next run - for example using a coin. Were you and your buddy at any time while the complete test was performed in the same room at the same time? Was the test observed by a neutral person?
I'm in It Networking the last 24years and you are right, IPv4/v6 protocol is from a pure Hifi standpoint a bad transport protocol. At the end every streamer of course does buffering and therefore a switch for 100 or hifi-optimized 1.000$ makes no difference. Said that, i own a paul pong dual switch and if course we know power supply matters, for all digital Devices. Do i hear a differenfe? Technical: NO Psychological: YES :-)
I don't know about audiophile switches but introducing an optical network ( two media converters connected together with an optical fiber cable) between server and router brought about a big improvement in sound quality.
Hi Paul, I don't have an answer to that question... however, your arguments could easily be applied to the power network, and yet you find clear differences there.
No, the data correction of digital data is in no way analogous to power, apart from it travelling through a large network. They are completely different conceptually - I think you are very confused.
There will always be people who think that if it cost more, it sounds better. 1's and 0's sound the same, regardless of how much the equipment cost that delivered it to your audio equipment. There's no extra "noise" or anything, it's just a stream of 1's and 0's. A consumer grade, regular router can handle a LOT of simultaneous audio streams easily as long as it is configured properly and using reasonable, but not necessarily "fancy" cables. The snobbery in audio and pseudo-science in audio equipment sales is laughable to anyone with a technical skillset. Any other industry would get shut down for fraud, yet so many people buy into it (literally).
Paul’s absolutely correct on this one. We’ll 98% right. Regardless of how those packets got to that PS Audio DAC, and I don’t care how much ca ca, delay and jitter is in the packet stream, that DAC is really getting it’s content stream from an internal buffer. And I’ll assume the buffer(s) in a PS Audio DAC are HUGE!!! Think of it as a bucket of water. Poke a small hole in the side and bottom of the bucket and fill the bucket with water. The water is going to leak out of the bucket at basically a constant rate. The water filling the bucket comes in at a rate that’s all over the place. So what your hearing from the little stream of water is completely un-effected by the rate the bucket is being filled. Ya Ya what if the bucket is almost drained. Well most good circuits dealing with an uneven incoming data stream “double buffer”. And I’m sure you’ve heard that term before. The first bucket fills up, leaks out of a little hole, that stream fills a second bucket and the stream of water leaking out of that is your treasured music stream. Double buffer chips will set you back maybe, 25 cents?? And making the buckets 10,000 gallon storage tank(s) may cost you another 10 cents. BTW those one or two buckets are gonna initially fill up in less then a second with minutes worth of content. I can remember double buffered digital receiver circuits in the 60”s. Remember dial-up modems? They had double buffers on the incoming data streams.
It helps a lot if you are using your computer as the music server. Its not about errors in the actual digital infomation its all about reducing jitter by reclocking it when sending it to the dac. Its the same reason why the same file always sounds better directly played from a transport than a computer hard drive or streaming. Just burn a disc and try it yourself
What he could do that would benefit him if his router that he got from the provider supports it, which is likely that it does… Is an able quality of service or QOS. You can use QOS to provide guaranteed quality for specific types of data that are passing through your router (streaming audio, streaming video, etc.) and let it de-prioritize other things like regular downloading or email pushes or smart TVs that are scraping data. If you have a decent router you can classify all the different types of data that commonly pass over your network and tell the router which ones are the most important to you and which ones you don’t care about. They can help you a great deal if your son Starts playing halo while you’re trying to listen to the symphony on Tidal.
QOS only really matter in a heavy-congestion environment--where packets that are sensitive to latency might end up behind a "'sizeable" queue of packets waiting to transmit. It would allow them to skip the line, as it were. If their network environment isn't used particularly heavily, it doesn't really do anything as the queues are cleared as packets arrive.
Love your vids Paul but have to disagree on your point about wifi vs ethernet cable.I have a relatively high end system which I find quite revealing in that component changes and even cable changes usually result in subtle and not so subtle changes in sound.I found a definite improvement in sq going from wifi streaming to ethernet cable connection.I suspect the high end passive preamp in my system plays a role in the transparency of the system as every active preamp I have heard seams impart it's own colour on the sound making it harder to hear difference in different components/cables.Just a thought.
But did you actually try and listen in a capable stereo system? Or are you just rejecting the idea of audible differences because you are a network engineer? I'm working in IT myself for about 30 years and I can clearly hear differences between switches.
You are a network engineer and you're giving your opinion from a regular network usage point of view. That doesn't make you a audio engineer that tested audio quality switches in hi-end audio systems, in order to reach a conclusion on that subject. That's a totally different matter.
Finally the right answer! Or you must have a very cheap $5 switch from China, which isn't capable to do it's normal job. When a switch is capable to provide a stable connection with enough bandwidth, than it should be OK. If you have very haevy video streamers in your household, you could consider a switch that can do some traffic shaping to guarantee a minimum bandwidth for audio streaming. But even these 'smart switches' are very affordable.
This is correct. Unless you're seeing packet congestion on the switch (from over-utilization on the LAN side plus saturating your broadband connection), a better switch (not an "audiophile" grade ethernet switch) will not make a difference. And like Paul eludes to, a better switch (in that situation) would only help reduce packet retransmission and congestion *locally*, which whatever streaming device you're using already handles by buffering. The same logic applies to WiFi. If you have a lot of WiFi devices that are competing for bandwidth/spectrum on a single access point/router (or have other WiFi routers around you on the same frequency/channel), over-saturation can occur and packets can be dropped or delayed. Again, this does NOT affect sound quality: only the timing of when packets arrive to the device (which get put into a buffer).
Well said, I know sometimes when you try to break down something to simply if, some of the tech heads get upset, but really, job well done on this one. I think what would make streaming better, the the dac you offer, as I understand it, you buffer in the input data, the clock it out with an extremely stable clock, basically jitter free.
I think the problem could start at the point, where the signal is being processed from Ethernet to USB or inside your DAC if it has built-in network functionality. Ethernet chipsets (also wireless) can be problematic in terms of electrical noise. In my opinion, you should try to isolate Ethernet as good as possible from critical devices in you audio chain. But for streaming, you need a network. There is no way around it. I think the most elegant solution is to introduce optical Ethernet for the one connection of your streaming endpoint device. It provides galvanic isolation and therefore eliminates electrical noise. Every device before that, does not really matter. And of course, your streaming endpoint should have a highly optimized Ethernet chipset, that does not introduce noise again... For me, this solution is the opticalRendu.
I would bet 10 bottles of german beer that all streamers are using the same microcontrollers with integrated Ethernet controllers than they are used for other (non audio) purposes also. And I dont know any Ethernet chipset ever optimized for audio.
So if I undertsand correctly switches are not sensitive to noise because the IP is digital and it is either gets a packet or not. But how come that computers as such are very sensitive to noise ? The are also digital aren't they?
The ethernet switch is too far upstream from the business end of the audio for that to matter. When you stream music from the internet, it is not streamed in real-time. Soon as you hit play on a song, it most likely downloads the entire song within a few seconds, so the entire PCM stream gets queued up in the computer's buffer well ahead of when it is needed. Now how the PCM stream is delivered to your DAC is the next challenge. Usb? Optical? SPIDF? All of these require real-time streaming and jitter and noise come into the picture.
I didn't really understand what he meant by saying that computers are sensitive to noise. They actually are the ones that introduce noise into audio. For that reason sound cards on motherboards are isolated from the rest components on a separate circuit through electrical filters. Howerever this is not ideal as induced currents can still appears from the monterboard components. That is why people use DACs and audio interfaces that are physically separated from the computer.
I am a manger over IT networking and security. I thank you for your video. I personally prefer a wired over wireless as the protocol for wifi is more chatty than the traditional ethernet protocol, sort of like the old school hub versus a switch. My suggestion for users is to create a separate SSID for all 2.4Ghz devices and a separate SSID newer wireless AC supported devices. You DO have to be careful on wireless as NOT to oversaturate the access points. Even with high-end enterprise grade wireless access points, 10-20 devices is a good place to be but I do not recommend anything over 30 per wireless access point. purchase more access points if needed and spread the wireless love to your devices. If you have a 2 story house, depending on where the router is located, you can install an access point (wired back to your router for better results) and have this access point on a separate SSID; per careful of stepping on your wireless channels. This will help to improve performance. Wired is still best, but not alway feasible. I wire all of my components, tvs and use wireless for devices such as smart home devices, and for laptops/phones/tablets. A friend of mine is a wireless expert; meaning Cisco and other vendors reach out to him for testing enterprise wireless products. You can get more tips and hints on his home page. www.my80211.com/
There are a lot of experienced reviewer who describe heavy differences... Why should they lie? There are a lot (!) of things happen in electronics regarding HiFi presentation and our ears are that sensitive. So if cables matter, why should not switches should matter?
Digital audio is not my forte. Vut, if you are happy with the audio reproduction your current set up is providing, then I can’t see how putting another device in the signal path can improve things. Enjoy the music and remember, music was born analogue!
Over short distances cable doesn't matter but if you're doing any sort of long distances it can increase your losses and drops that can matter. Realistically I would stay away from cheap switches like anything that's like $10 because that's going to induced noise from whatever crappy power supply they're using but more than likely that's just going to cause connectivity issues not sound quality issues unless there isn't proper isolation in the dac.
I watched in complete astonishment some CZcams Audiofool claiming that he could detect the subtle nuances between several different “audiophile” switches…it’s beyond hysterical. 🤦♂️
In low quality routers you have low quality clocks very bad quality ac power, that results in jitter and bad power in rj45 going inside your server. Ok to normal computer work, but not for music or video. A good switch corrects everything. Don't compare yours Enterprise network with a normal house network.
"A good switch corrects everything. " How can a "good switch" correct the "low quality clock" build in a cheap streamer? A streamer writes the "perfectly timed" data bytes he receives other Ethernet into its buffer from there it is readen with a speed defined by the streamers clock and send to an DAC
Dear PaulThank you for great videos, most of them are very educating.I agree on many of your obviations, but I do not agree on this: Ethernet cables do not matter.For the last 22 years have I been working with IT and computers professionally, my hobby has been Hifi and Electronics for over 40 years.I have built my own DAC and I have been working on a Dual PC solution as Media Server for my DAC. 1) Ethernet cables from my fiber connection box to my Router/switch, does not matter. 2) I run Wifi between my Router to my Control PC (Media Server), so here does Ethernet cables also not matter. If I did connect my Control PC (Media Server) to my Router and also uses a NAS which is also connected to the Router, when would Ethernet cable matter. 3) My Control PC (Media Server) is connected to my Audio PC (here is the stream transformed from network to USB) here does Ethernet cable matter a lot!It all comes down to, how your system is designed if Ethernet cable matter or not.One have to have a huge knowledge to be able to analyze the need or not, so my recommendation to the average user, do not use money on Ethernet cables, your money is probably spend better elsewhere in your setup.Thanks! Jacob Land Jensen Denmark.
No doubt a switch with a good linear power supply will sound better. I've done the tests, it's all about the quality of power and lack of noise. Everything matters.
@@Iblis666ful Seriously, do your own tests. With my Naim ND555 steamer you can hear the difference between a switch with a switch-mode power supply vs. linear power. Noise is always the issue... not the packets.
if you know a bit about computer then it is more about whether you have good wifi signal or not. if you have weak wifi signal and the speed gets slow then the music can stop playing. As long as the player receives a data pack of music, it plays as well as a CD player! It is like transferring a picture via bluetooth! All routers do a good job as long as the wifi signal is good or you use the network cable!
I dont have the technical background to take a stand on whether switches matter or not and I have never tried an audiophile switch on my system, so i am happy to sit on the fence. However Pauls argument as to why switches can't make a difference ie the signal is being bounced around all over the place before it reaches your house so the last link in the chain cant make a difference is the same argument power cable skeptics use. Yet Paul has previously dismissed this and argues that power cables do matter because he can hear the difference and that the fact he can not explain it scientifically is irrelevant. This seems to be a case, with all due respect, of having your argument and eating it. In this case many people claim to hear an improvement with switches, like dear old Hans, but Paul does not, so therefore switches don't matter. As I said, I have no skin in this game, but I think, unusually for him, Pauls argument is resting on pretty shaky ground if you compare it to previous utterances. Either subjective listener experience takes precedence or it doesn't. Conversely the scientific consensus takes precedence or it doesn't. You can't have it both ways.
Hi Paul - from my PC using the built-in command tracert in a console window I see there are 10 hops from my PC to TuneIn.com - 4 hops to before I get into the internet from my provider alone! Takes up to 23ms to get here. Streaming 320 kbps gives *audible quality improvement* over 128 kbps. All the best from Switzerland, Rob
I added a basic Netgear 5 port switch between my Comcast modem/router and AV gear, and there was a very noticeably lowering of the noise floor plus all of the typically associated benefits. I think a switch (even a budget one) will be better than running ethernet straight from your noisy Comcast device directly to your sensitive AV gear. Worth a try folks, if you don't hear an improvement send back to Amazon.
the most ridiculous thing is that in most homes the internet connection arrives on a 0.2 euro / meter twisted telephone cable and even hundreds of meters long .....
Hi Paul, I maked the test with SOtM switch sNH-10G with externel Mutec Ref 10 Master Clock. Difference with commectial 8 Gigabits swith is BIG. Very low noice, great instrumental separation and like never leasten before sound images in my system. For sure problem isn''t in the IP data package but il the last digital segment from Router Power Supply (very noising switch), your position is very basic there is a lot of technical question mark inside any audio system at home that probably the audiograde switch can fix. PS Audio is big company so you can make the same test with your Directsctream and FOR SURE you will change your opinion on this. This also happened to me.
There is more to digital than zeros and ones. It will take some time for pure objectivist to accept this, and that is fine. We all need time to change our beliefs
@@FooBar89 whilst I appreciate your response TLS....was wanting to hear from Moussa about what he meant. It's just too easy to make a condescending comment to all readers without actually providing some detail It's a bit like me saying there's more to photosynthesis than the sun....we will all change our beliefs in time.
@@nostro1001 I am not an engineer, but I have been comparing digital gear in my system for years now. That includes USB cables, ethernet switches, servers etc. There is always a change in SQ. If it were only 0s and 1s then nothing should matter. Of course this is based on my subjective ears.
@@moussaobeid Oh right....so no qualitative evidence, or actual science, just your opinion. That's cool, your initial comments led me to believe you had some deep understanding about what you said, so I wanted to learn. Cheers!
Steve Fick you stick to your CDs bro, locally stored music in my personal experience is king, streaming is good, but doesn’t compare from a sound quality perspective.
Rami Bachar indeed, listen to what you want. I’m just pointing out Steve’s locally stored music is most likely to be of better quality than streamed music.
In my humble opinion, perhaps banking and financial industry shall be targeted for marketing with high grade router / ethernet switch rather than HiFi industry. New York Exchange are deailing with billions of dollar / stock transaction each day, visa and master are similar...Their data accuracy rate (associated with money) are far more important than music data (audiophiles). If banking and financial industry (even military centre or NASA) are comfortable using ordinary commercial grade router / switch for their routine data processing works. It seems that some audiophiles without reasonable logic for using so-call higher grade router/switch for routing their music data!!!
I would agree that using an audiophile switch for streaming Tidal or Qobuz won't make any difference, it's never going to sound great. However, streaming music files to your DAC from a NAS on your own internal network could very well be improved by better quality switches with better quality clocks and power supplies.
USB falls in to the same boat. It works till it doesn't. The cable can pick up some noise but if the gear was built correctly this noise will be easily filtered out before it effects any anallg audio. If the cable and gear was built to basic industry spec changing to a fancy USB cable shouldn't make any difference. If it does, just means that piece of gear wasn't well engineered or done on purpose so their partners in crime, cough "audioquest" can sell you a expensive USB cable. People that buy expensive usb cables probably aslo buy gold plated optical cables....lol
Sry i disagree going from MacBook Pro to Chord Qutest to Atoll AMP .. and i was skeptical but the 500 Euro cable is day n night between the Standard Audioquest Cable. Streaming via Quboz in Hi Res
@@borisbobowski628 Well precision testing and measurements done by AudioScienceReview and Audioholics say otherwise. If a USB cable is effecting say a DACs audible output then its pooy designed, shouldn't happen if engineered properly. Chord gear is nice but tends to measure poorly for reasons even most CHi-Fi (Topping, SMSL) gear has been able to overcome at a fraction of the price. I have had about 20 different DACs and and several cables, none have made a difference for me.
@@iowaudioreviews i think No Audioholic can denay that different Dac's Sound different or more transparent. I can clearly say and ofcourse measure that my old DAC has Not a slightly Chance against my chord DAC. I dont understand why U was buying Dac's If you didnt Hear a Difference after lets say the First 3 Dac's? Something Must be really unbalanced in your Setup.
@@borisbobowski628 Most were sent to me to try, some barrowed from friends and others bought and returned. Most places have a 30 day trial or return policy. Some sound slightly different and others can't be discerned from each other. Got boring so I just use a miniDSP, REW/mic, and EQ APO/Peace to fine tune my system. Its made the most noticeable improvement for me and has tons of flexibility. When the techy stuff gets annoying I just fire up the turn table and grab a Parts Express catalog and forget about it.
@@iowaudioreviews OK Hearing ni Difference or slightly differences is indeed a big Difference. Im blessed with really good ears, No Joke. And i hate IT that i really hear this tiny differences because to me its Like opening a Door...you cant Go Back Afterwards.
Exactly. The clock in the DAC will handle any potential jitter from a switch. Noone ever published a scientific a/b listening test proving anything. Or better still, a measurement on the signal coming out of the DAC comparing cheap versus audiophile network switches. That signal will be identical.
@@JustBrowsing777 that Dutch audio guy swears he hears a difference in an obscenely overpriced "audiophile" switch and gets pissed if that is ever questioned. No problem with me, it's his money, not mine.
@@nunofernandes4501 I like Hans channel, he taught me a lot, but as with anyone's claims you always have to question them and not just accept them as truths/facts. I've struggled with what is in fact impacting audio in the range I can discern better from worse. It made a huge difference to switch out my amp and streamer/dac but power cables, speaker cables and interconnects....not sure I can hear anything and if anything very little in my system. Switches no way😄
@@JustBrowsing777 yes, I like Hans's channel and learned a lot from him too, but I don't agree with that particular point about switches. As for interconnect cables, in my experience there was a clear improvement in sound when I upgraded from €10 but decent quality supermarket RCAs to AQ Evergreen for my DAC and AQ Tower for my CD player. These cost me around €30 each and the results were so good I doubt spending any more would be worth it for my personal pocket depth.
@@nunofernandes4501 I agree with the interconnects. A mid range pair is fine as long as you avoid the cheapest ones. I have seen/heard comparisons on these with a clear audible difference.
I found that moving the switch further away from the music gear made the biggest difference. Since the psu you receive with the switch is usually a piece of shit and more of an emi source than anything else.
Why don't you tell Hans Beekhuyzen that internet switches don't matter (and I agree they don't). He seems to thing a $600 internet switch made an enormous difference in his system ...
I'd rather listen to a seasoned and successful designer about anything, for that matter. Instead of a gadget lover/hobbyist with lots of money, who also started rather late, because of his son interest in the field...
Get an optical isolation i.e. a regular single or multimode network connection between your streamer and your otherwise noisy switch mode psu that powers the cheap modem/router or whatever delivers your Internet connectivity. Is there any streamers with an SFP Ethernet port? Or wait just simply go wireless, since that also isolates your streamer from any noise on the Ethernet side. But then let’s talk about that data integrity, how does streaming services like Tidal make sure that I get a bit perfect reception of that centrally stored audio file which also has to be identical to the production master. Do they rely on the CRC validation in the Flac format and actually discard and resend frames that do not compute or is that simply ignored causing dropped frames. Or even worse are they played back without validation? And is the flac frame footers CRC-16 enough to validate the various frame sizes? Also how does the streaming services store the music files that are sent? Can they be subject to silent data corruption and what mechanism are in place to guarantee those original zeros and ones that the master intended. That is an interesting topic, at least for me!
Ethernet is already electrically isolated. Magnetic, rather than optical - tiny coupling transformers at both ends. Streaming services for music will simply detect lost or corrupt packets via checksum and re-send. VoIP handles it differently because it's latency-critical - by the time a lost packet is re-sent it'll be too late to be useful, so they just attempt to interpolate the missing data. It works well enough for voice that the user shouldn't notice.
@@vylbird8014Yes, ethernet is galvanic isolated with those tiny coupled transformers. Still, that does not eliminate noise coming from the electronics. Nor induced noise from the cables. Also STP cabling could bridge the ground planes between the router and the streamer if used improperly.
"Do they rely on the CRC validation in the Flac format and actually discard and resend frames that do not compute or is that simply ignored causing dropped frames. Or even worse are they played back without validation?" Tidal uses TCP/IP as transport protocol for their streaming data. So they not have to handle transmission errors or lost packages on their own - all this is done by the TCP/IP software. "Also how does the streaming services store the music files that are sent? Can they be subject to silent data corruption and what mechanism are in place to guarantee those original zeros and ones that the master intended" Streaming provides are storing their music on hard discs (and maybe some stuff that is requested often on SSDs). Both are using also CRC checksums (build in already in the storage deviced hardware) to identify read errors (or data corruption on the media itself). Such harddiscs are typically bundled to RAID-5-Systems which offers another level of error detection and also error correction. The chance that such an error is not detected by the CRC checksum is about 1 in 4 billions
When streaming the data isn't being sent just in time as it is played, instead, it is being sent in chunks of data that get queued up on the streaming device (clock and jitter from buffer to DAC is only related to streaming hardware). Unless the network connection is so bad that it can't keep the buffer full then there is no difference. If for some reason you are having issues then you can enable the QOS settings in your current network hardware.
All TCP/IP based streaming systems buffer the data at the endpoint. There is no timing variation in the data stream, because there is no timing in the data stream; it’s just a sequence of bytes. It might as well be written down on post-it notes at this stage. Also, the transport protocols reject damaged packets, so noise at this stage is not a significant factor (it will reduce your download speeds, but even the most basic of networks have far more capacity than an audio stream needs).
The timing is constructed at the DAC; and if you have a good DAC, then you will get good timing. Noise going into the DAC should be well below the digital switching threshold (which is the whole point of digital audio - each sample can be carried thru the system perfectly with zero variation) so spending inordinate amounts of money on “audiophile grade” digital connection hardware & cables is pointless... regular decent quality gear works just as well.
Also, sorry, no... computers aren’t excessively prone to noise sensitivity. They’re actually quite resistant to it: typical line-level audio signals might be 1 volt peak to peak; but a digital signal - as used in the computer - will swing from a low level of less than 0.1 volts up to 3.2 volts or higher (older standards go to 4.8 volts or so typically). You can hear mili-volt noise in an analog signal quite clearly; but that noise level on a digital signal is going to be several thousand times less than what is needed to change the signal from a 0 to a 1 (or vice-versa). Use good quality, but regular digital components and you will be getting the best possible performance out of your system; because it will be 100% bit for bit perfect. Note that if you do have that kind of noise injection into a computer that will flip bits; the result won’t be limited to just a few low-order bits in an audio stream being flipped; the whole machine is going to crash (if a value is off by even one bit, the entire execution path goes completely off the rails).
What digital gear can do, however, is create electromagnetic emissions that can couple into the analog side of your setup; so using balanced, shielded cables run at right angles to digital cables where they have to cross; and putting analog gear in appropriate metal enclosures is a good idea. Good analog gear will have power supplies that filter out noise coming in that way also; but if needed a line choke can be added. Most good gear will already come this way. Also, the cables don’t need to be ridiculously expensive either; just good ones of the same grade as is used in professional recording studios (after all, that’s how the music was made in the first place).
Finally a nerdly answer that's 100% correct. Bits in = bits out, and the buffering and retiming of those bits is performed deep within the DAC.
Thank you for this explanation.
@@hwarnold123 So true!
Nope you are wrong, it sounds right in theory but a PC is an almost endless amount of noise that needs to be taken care of. Switch is also a bit noisy, everything creates noise especially all out PSU and LPSU's. Bit theory only belongs in the network world. Science with numbers does only half belong in the audio world.
Have you ever noticed why that most music systems in the world sound so mechanical?
Because they believe in your theory and only treat it as a mechanical field. What I wrote is the truth.
But that's the problem with the world, either you are the one or the other extreme, being in the middle partly agreeing partly with both sides of the equation is not popular.
Audio is an art, and most engineers in this world treat it as a science, and that is why their sound is close to perfect, but has no soul.
"Peter Qvortrup said it perfectly himself: You cannot measure good sound!"
Real HIFI Help - I am quite sure that you are going to disagree with me; and I’m also quite sure that I’m not going to change your mind with a CZcams comment. But I’m bored, so here goes:
The path a sound takes from instrument being played by an artist to your cochleae sending nerve impulses to your brain can be divided into distinct stages; which will of course vary depending on the recording in question and on how you are listening to it. For the sake of discussion; let’s say it’s a classical guitar recording: strings & body vibrate, inducing vibration in the air, inducing vibration in the microphone capsule, which puts out a voltage proportional to the air pressure, which is amplified and run thru the recording equipment chain - and here I’m skipping a whole bunch of stages because, well, this is a CZcams comment, not a textbook 😉 - and ends up as a .wav file (again, just for the sake of discussion; AIFF, or any other uncompressed format would do as well here) on a server.
At this point there is only the noise introduced by any stages in the recording chain from the microphone thru to the file on the server (and of course, this should be self-evident). Now, this is a digital file: the whole point of which is that it can be copied from one place to another exactly, with no variation. What is less obvious; and what I was trying to convey above in my first comment; is that there is NO timing in the file. None. Files cannot, by their very nature, have jitter - it’s a concept entirely foreign to them. They are simply sequences of numbers.
The consequence of this is that as long as the copy is true (I.e, the sequence of numbers you put in on one end came out the other end identically - and there are mechanisms both in the hardware and software to make sure that happens 100% of the time); by definition you cannot add noise to the file... and the consequence of that is that it doesn’t matter what you send it thru, by definition you have the exact same thing you started with; with zero exceptions.
The question was if spending extra on a fancy switch or router made a difference. Well, given that the network device in question operates on only the digital file stage of the chain, it should be self evident that as long as it’s not defective; then it cannot have an effect.
Once it’s reached the stage where it leaves the digital realm and becomes an analog signal again; noise becomes an issue again - both in amplitude and in time. This is why your DAC should be in a separate box; and have a good, stable clock generator.
Re: bit theory. This is fundamental to - and therefore applicable to - any place information exists in digital form. As long as the audio is digital, it applies.
As for science with numbers, it is by definition concerned with, and fully applicable to reality in general... and last time I checked audio was very much a part of reality.
Saying that a system “sounds mechanical” is too vague to consider... what, exactly, is meant by that?
Yes, audio is an art. What it isn’t is magic: like all art, it is built on the fundamental principles of mathematics, physics, and engineering. There’s nothing going on that is non-understandable. You cannot truly master any given art until you fully master not only the what, but the how and why - and this applies universally; from as diverse arts as painting to ballet to hi-fi audio (and yes, I contend that the “great masters” of the past, and often the present; even though they are generally considered to be the best in history, are not true masters simply because they didn’t have access to the fundamental knowledge that underpins their art).
As far as measuring universally “good sound” - you cannot measure it because in this context, it doesn’t exist: “good sound” means “sound that I like”... and people vary. What you can measure, in principle and to a large extent in practice, is what is actually physically going on; and from that extrapolate that any given setup’s characteristics are (or aren’t) ones that are popular.
I opened this thinking I was going to hear some snake oil about switches. I'm an IT professional and have been for over 30 years. I opened this thinking my blood pressure was going to go up, thank you for the realistic answer.
Exactly this, only difference is that I'm 18 years in this profession.
This was my exact thoughts when I saw this at first. He explained it perfectly.
Same 👍👌
Same. I'm am IT professional as well
Same here. IT professional of 2+ decades, braced for some crap about having to use "audiophile grade" network "interconnects".... Thanks goodness some sanity prevails.
If there is someone I trust on all things audio, how it works, what make a difference and what does not, it’s you. You just seem incredibly knowledable and honest and you just seem to speak how it is without any bs. Thank you Paul.
What? You mean I also don't need to spend $$$ dollars network cable by Audioquest? Surely ONES and ZEROS need special wire with special shielding and special connectors to stop them from becoming TWOS and THREES.
Some network cabling (I know you're kidding, but I thought I'd mention it anyway) is heavily shielded. Not for getting data wrong per se, but for speed. The better shielding results in faster transmission due to lower packet loss.
@@aweidenhammer
Yeah but you don't have to spend 50+ dollars per meter for "special" shielding.
Here's a is the Audioquest's VP giving a seminar on computer audio...
czcams.com/video/Gjsc_LTurl0/video.html
Really instills confidence in their products, doesn't it? PS Audio says Audioquest's Dragon power cable is significantly better than their own, but nobody provides evidence it's better than a 40 dollar shielded generic power cable to justify its 4000+ dollar per meter price tag...
Oh, the tired old "ones & zeros" argument. How original.
No, but we just outfitted our office with cat6a, which is more expensive for the shielding. Not saying it makes any difference to audio quality though. What Paul fails to mention is as long as the signal is in the digital realm, it has packet redundancy and error correction so the data is a perfect copy. It's once that data is set to a timing clock, sent to a dac and converted to analog where things get tricky.
@@arthurott4561 Yet so accurate.
If he's streaming audio through Comcast, Comcast is already the source of 99% of his jitter. At my last job, I supervised IT at small office sites which had dual WAN routers with both Comcast and Verizon FiOS. Here's what I found: The core of both ISP networks was fairly stable and comparatively low jitter (as measured by ping and traceroute). However, the last hop of Comcast increased the jitter greatly, while FiOS was low jitter all the way to the customer premises. I came to believe that this was because the Comcast last mile bandwidth was oversubscribed, while Verizon's last mile bandwidth was undersubscribed. That is, Comcast did not have enough bandwidth on their coax when lots of customers were using it, while Verizon had far more bandwidth, so all of their customers could be running at maximum and still have bandwidth left over. What I did was set the dual WAN router to divide up the traffic so that all the non-web traffic went through Verizon, which included VoIP where low jitter was necessary. All of the web traffic went out over Comcast, because they had better peering which led to faster throughput.
So what about music streaming over IP? The streamer has a jitter buffer into which the audio data packets are collected, and they are clocked out with a stable, low jitter, local clock. So any jitter which occurred during transmission is removed by the jitter buffer at the streamer. In music streaming, the jitter buffer can be several seconds to 10s of seconds long, so it can even deal with situations where transmission is interrupted for several seconds. Now that's jitter! The reason VoIP needs a comparatively short jitter buffer is that a two way conversation cannot tolerate long delays. If much more than a quarter of a second jitter buffer is used, the callers need to start saying "over" so the other knows when to talk.
I should also point out that part of the confusion is because audio and IP people are talking about different things when they use the word "jitter". Audio people are usually talking about waveform jitter, which is measured in nanoseconds or picoseconds. IP people are talking about packet jitter, which is measured in milliseconds or seconds. A streamer has to have a jitter buffer which is some number of seconds long to take out the packet jitter, but audio data needs to be clocked out by a local clock signal whose waveform jitter is measured in nanoseconds or picoseconds.
So the answer is no, a low jitter ethernet switch will make no difference to streaming audio quality.
That's nothing to do with jitter in this context. Jitter is variability in the clock signal, and in this case the ADCs generate their own clock signals. What you are seeing is variability in the rate of data transmission, and providing their is a sufficiently large buffer it has no impact whatsoever on the quality of the musical signal. If it goes beyond that level, what happens is that the playing of the music stalls until more data arrives.
Only real time audio needs specific switches or protocols. Such as AES67, QLAN, etc. These switches depend on a clock to stay synced and have less than 1ms of latency. Streaming audio has no latency concerns and is buffered and reassembled to sound proper. The assembling codec is the most important.
Right on point. TCP/IP is robust and streaming services tolerate packet loss in transit. Each packet has an ID and a checksum. A switch that affects sound is one that is not functioning correctly. "jitter" is a complete non-issue with any properly working network.
Cheers.
Has nothing to do with jitter or data integrity and has everything to do with noise.
The "if it does make a difference then it isn't functioning correctly" argument is very interesting. It seems to be the anti-audiophile strategy of admitting you are wrong without admitting you are wrong. You start off saying it cannot possibly make a difference, then reluctantly agree that it does, but still maintain that the ways in which it does make a difference don't actually matter and that despite it making a difference you are correct in asserting that it doesn't.
I have a sotm connected to an Aurender N10 streamer. The difference was enough for me to spend 800eur +500 for a power supply. I’m pretty shure unbelievers never tried it. Or never tried in a proper system.
Aurender n10 has a buffer which coppy all the data it receives no matter where it comes. You are always playing from the buffer.
I'm pretty shure You are imaging changes where it's only snake oil
Make sure your network cables are connected in the right direction for signal flow according to those little arrows that all network cables have.😂😁🤣😎
Agreed a sensible answer everyone knows the only thing you need to worry about is making sure all your fly leads and patch cords are red to ensure the signal gets from point a to point b as fast as possible
Everyone, so far, seems to have missed the point. Bits are bits and networks accurately transmit them, otherwise our downloaded programs wouldn't work, because they MUST be bit perfect. So data accuracy is not the issue. Timing is. If there is jitter, if the pulses aren't perfectly regular, then that can affect sound quality. The trick is a reclocker, which fixes this and some are quite inexpensive. Just before your digital goes into your DAC, reclock it!
jonc113 The data on your Ethernet cables going to your audio player are entirely asynchronous to the DAC clock and any “reshaping” is total utter ignorant nonsense. The data is buffered in the player and as long as it has enough data (not starved) the player and DAC are happy. The player pulls the data from the server in blocks btw making this topic even more idiotic. This is scandalous snake oil nonsense.
When you are talking about switches or anything network related you do not have "sound quality" as there is no sound signal hence timing is irrelevant. Timing only comes into play the moment the digital buffer is being addressed for conversion which would be in the DAC.
Still waiting for somebody to tell me what 'jitter' actually sounds like.
You do realize that as a frame enters one port on switch and then exits another port on the switch it's regenerated. CAT5 at 10/100/1000Mbps is good for 328ft/100 yards. I can simply extend that by placing another switch and adding another 328ft. Rinse repeat. The reason this works is that the signal is regenerated.
Switches also have buffers. At default they are mostly FIFO but we can change this with Weighted RED.
Also our playback systems are not real-time and not delay sensitive. They are asynch. Data can be delivered ahead of playback.
Matt Noble Yes absolutely pure ignorant snake oil BS that deserves zero respect. People selling such stuff at an overpriced MSRP arguing it “reshapes” the timing of audio should be ashamed of themselves.
I'm a network administrator. 1. Those "audiophile switches" are a waste of money. 2. As long as you're cable modem/router combo is working correctly under a normal load you won't be able to tell a difference streaming audio. 3. Now if you're home network has a traffic problem , QOS ( quailty of service , aka traffic management ) settings will make a big difference streaming wise. I have a FiOS router ( I disabled the WiFi on it ) but I added a cheap 16 port unmanaged gbit switch ( not a hub ) + Google WIFI. Each node has an Ethernet link back to the switch, master node it's between the router and the switch. I have a lot of devices at home and I do a ton of HD video streaming over WiFi. I get wire level speed over wifi from anywhere in my house. That was like a $350 upgrade. Adding a managed switch would have only been $50 more but didn't need those extra options. I've worked with cheap netgear switches up to $30k+ Cisco switches. And yes I played with a few audiophile switches too, save your money. My home switch is Netgear prosafe jgs516 listed around $100 on Amazon.
Jaakan Shorter so you’re the one my computer wants me to contact when my internet goes down!
I did something similar with my gig fiber connection- all into a Cisco Catalyst switch, then split into three separate WiFi access points which are located in different parts of the house. One is for me, one for the kids' gaming systems and the other is for the cameras and alarm system. Splits the loads nicely and I have QOS set up in the Catalyst to make sure I get preference over the kids since I paid for it all. We can have over forty devices on the WiFi and 11 devices wired and streaming still comes through with no discernable packet loss as far as audio goes.
There is certainly not a difference on the digital / signal side, but if you connect an (any) electrical device to your system it might introduce some analog garbage to the ground level. Even a lamp on the same socket as your system or touching a metal surface of any your devices might make a difference.
I have never heard of ground noise making it through the ethernet adapter all the way to the analog audio out of your device.
More importantly, if that was such a concern, it would make far more sense to run a fiber optic cable from your switch to your streamer to completely isolate it from any electrical noise going through a copper Ethernet cable.
@@go2yanks that is what drives me the most wild. I've seen people who make audiophile an identity rather than a hobby with tens of thousands USD in a rack for a single listening station and it''s all overpriced garbage using meaningless buzzwords. At the end of the day if they were *that* worried about electrical noise, they would use fiber (which can even be incredibly cheap to do!)
@@kumakell I don’t actually have any issue with spending a ton of money on components regardless of how they may measure.
My big issue is when companies start actively lying and deceiving people with deep pockets who don’t know any better. Synergistic Research is the biggest offender in this regard but they’re far from the only company peddling that particular variety of snake oil.
@Dan My streamer is already galvanically isolated from the rest of my network using fiber optics. Going to their copper based switch would be a lateral change at best while costing several thousand dollars more than my implementation.
Paul, while your explanation about data transmission is correct, it does not address the issue of electrical or other noise generated by the router, switch etc…polluting sensitive audio components downstream. 2 quick ways to test/verify this…..1) Swap out the switch mode power supply on the network switch for a good quality linear power supply (LPS). 2) Try a network switch that has a fibre optic SFP input and isolate it from the noisy network by connecting it to a fibre media converter using a length of fibre optic cable (the media converter is also connected to the wifi router using an ethernet cable). Ideally all of these powered by LPSs. The proof is in the listening and the extent of the impact made me question why my very expensive music server (which already sounded fantastic) hadn’t/couldn’t fully address/resolve network noise issues. Please have one of your electrical engineers check this out and let us all know. Thanks.
I am sorry but I can not agree. In my system the switch made a huge difference. And after adding the dedicated power supply for the switch from the same manufacturer the improvement was just as big. Paul just says it does not make a difference, but does not tell if and what they have already tried.
If your stereo is your hobby trying new stuff is part of the enjoyment. Most of the devices can be ordered risk free with a money back guarantee.
So maybe try it for yourself. If you can not here an improvement you have assured yourself and can tell everybody "I told you so in the beginning" .
If you can hear an improvement, keep the device if you like and enjoy.
i was just into the subject, have found several sources on the internet that it can make a difference (in audioquality) because a switch filters a lot of noise from cheap ac/dc devices and computers.
@@thepuma2012 just FYI I am using a Silent Angel Bonn N8 switch and the Forester F1 power supply. Streamer the switch is connected to is a Linn Akurat DSM
You indeed have to have decent equipment to Heat a difference. For what i read the music opens a bit more. My equipment is not high end. Bit i found a cheap one which could make a difference maybe. My audiolab amp is maybe Just revealing enough to try
Switches can diminish noise isn't it? They can cause noise too with bad power supplies. Yesterday I tested 2 ethernet cables and they sound different. The data will be correct in both so I guess they do something with noise
"Yesterday I tested 2 ethernet cables and they sound different." Yeah OK.
Maybe they wre different CAT types or untwisted wires picking up EMF from lights or something?
And can you also see differences in picture quality. Would love a screenshot.....
That either didn't happen, or you defied lottery-winning odds to have a cable that provides a solid link yet somehow drops packets. With cables, it's mostly works/doesn't work- although shielding makes a difference when in an electrically noisy environment. It's 1's and 0's and if they all (or mostly all) get to the destination, it sounds the same regardless of the cable.
@@4879daniel I always wondering why there are no "videophile" Ethernet cables (or switches) on the market.
More contrast, clearer colors, smoother movements and even the actors would play better.... 🙂
For once, somebody willing to call BS on this issue! Thank you, Paul, for being a voice of reason.
I think the SoTM switch mentioned at the start is good if you have a big house and use the fibre cable over long runs
Notorious truths:
1. $20 switch and $640 switch - no difference. Amir from ASR measured it with the most precise $100k measuring device on earth. No difference in the spectrum.
2. No difference means if one sounds good then both sound good. If one sounds bad then both sounds bad.
3. Noise alone does not matter for jitter (time distortion) IF BANDWIDTH IS INFINITE.
4. Bandwidth alone does not matter for jitter (time distortion) IF THERE IS ZERO NOISE.
5. BOTH noise (low frequency and high frequency) AND bandwidth constraint together impact jitter.
6. If you put any filter on the signal path you are going to constrain the bandwidth. Which is bad for jitter (read point 5).
7. The noise should be exterminated as much as possible in the music server, before it reaches the stremaer and DAC. For once it leaves the server it is all too late to deal with it later.
8. Bitrate matters: higher the bitrate the higher the impact on bandwidth due to more time needed for processing the stream
"Audiophile Ethernet Switch"... that part just killed me.
It’s very strange that everyone claiming the switch makes a difference, always have some anecdotal references. None of them are able to show the results in measured data.
The first part is correct. As long as we are talking network equipment, including the cables, up to the streamer, then it CANNOT affect the sound. The datapackets are indeed assembled in the streamer, if it misses a part, or fails a consistency check, then additional packets are retrieved. And btw, this is done BEFORE the Application, layer of the osi model. So it is even independent of which streamer you use. What happens after that is up to the streamer.
A "noisy" switch, whatever that is supposed to be, cannot affect the completed file in a streamer.
'Hifi' switches and network cables are an absolute scam. Stay far away from them.
Even the Main electricity source in your Wall can affect your whole Performance.
@@borisbobowski628 the point is, datapackets sent through a network protocol will arrive in your streamer exactly the same as they are stored 4000+ km away on a server in a datacenter. It's just how this works. Error correction, acknowledgements, reaquiring missing or corrupted packets etc. If this wouldn't work, we wouldn't have the internet like we do now. It would just crash. So, regardless of the network cable you use, or the router, if technically sound, it just doesn't matter.
Once the file has been assembled in the streamer, then that part is over. It is then up to the the soft and hardware of the streamed.
Btw, there have been quite a few people explaining and showing why the whole powercable etc story is just as much nonsense as hifi network cables. A LOT of placebo effect going on. Noise is filtered out once a powersupply transforms it. Those filters are very cheap and do the job. Of you want to filter 'noise' at 220v 1 amp etc. You need extremely powerful and expensive filters. Which are then rendered obsolete by the ones i mentioned prior.
Its a waste of money really.... now analoge interlinks and powercables, that's a very different story
@@CyberBeep_kenshi so even on normal Websites you and i notice different Speed, in Terms of high res Files the Timing is absutly Key If you wanna have a Perfect Decoding going on. Tbh PS Audio is Not a really high End Brand, Their Streamer are Not good and they have No clue. Every source has a different quality and It makes a huge Difference If you Drive a dac with a Smartphone/MacBook or even desicaded Streamer. Ofcourse you Need a great Setup. Even my Girlfriend that is Not a active listener can Hear when a Switch is connected or Not.
@@borisbobowski628 the speed of a download has no effect on the final file. And so it cannot affect the sound. Now, if you go through wifi and do live streaming, that is a different matter. It uses udp instead of tcp ip. Which allows certain packets to be dropped. But when you use i.e. qobuz, it will copy a flac file onto your streamer, which is then played. How it gets there isn't relevant. The file is 100% the same.
To give you a small idea of what packets go through (yes its my job :-) ) Storage center, (fiber, patchpanel, fiber patchpanel, fiber) = fpfpf, database server, fpfpf, application server and or loadbalancer or both, fpfpf, switch, fpfpf, coreswitch, firewall (at this point you are still in a customer cage for a customer in a datacenter. Then its routers, fiber or copper, router, router router, satellite or ocean fiber, the more routers and finally in your street.
Then into your house, modem and THEN finally in your switch. We are talking easily 4000 km depending on where you live.
But at the end its the network protocol that ensures the file is exactly the same :) internet really is a wonderful thing :-)
Then to a central connection to the outside, riuter
@@borisbobowski628 regarding ps audio, i never heard it irl. But its a lot of talk and big boxes with tiny circuitboards. Never a good sign.
To be fair, i have a blue node as a first go on set, but i use the Dac on my pre amplifier. So far no complaints:)
I dont see how it would. Its a digital signal, the bandwidth you get to your place is your bottleneck. Now with broadband so available its not much of an issue. Its not like an analog signal thats.. all contained in a local area to an extent. Thanks for clarification Paul!
I believe - without being an audio engineer - that these switches address potential noise being introduced by standard issued modems and routers, not how the data is delivered. Alvin at Vinshine says it best: "at the very least, it fixes audiophile OCD".
Alvin's SW-8 and I believe the SW-10 both use this Netgear switch, the GS108Ev3. They are $39 on Amazon right now, and boy do they work! I have since installed three of these switches in the house to decrapify my internet. The damn things work!
On a side note. I do own the LHY SW-8, it is a fine switch. BUT! Save your money and get the $39 switch and a decent 12v power supply, like a Meanwell or one of Alvins LHY linear units. I cannot tell the difference between the SW8 and this less expensive combo. The improvements in PQ and SQ were quite profound, either way you go. Piggy back two of these, I have the $39 switch feeding Alvin's $500 switch, and there was even more of an improvement, subtle, but there. I have three of these $39 switches in the house, as they really cleaned up my noisy internet feed.
@@michaelwright1602 Somehow can't remember asking for financial advice, lol. The SW-8 is using a Linksys Chassis with upgraded components. No relations to a Netgear switch, which serves its purpose when achieving a noise-free signal is not a concern.
The only place where switches make a difference is in live sound where low latency can not tolerate large store and forward delays. If you run digital stage boxes with your digital audio console, with digital hearback personal mixers for the band, then yes, switches make a difference, and not "Audiophile Switches" but low latency switches for commercial sound such as sending 48 microphone channels and 24 monitor returns at 48KHZ 24 bit audio with less than 17mS of delay. Running a single stereo DAC for a streaming audio service with several minutes of stream buffered locally, won't care if your router store and forwards a packet 600mS late.
They don't matter from the audiophile point of view as long as the right ones are chosen to build a properly designed network that is able to carry the expected traffic with no congestion. And if that's not the case, then they matter as much for audio streaming as they do for any other kind of traffic.
Most audiophiles that have money have been stuck to the analog world so they think that digital audio works like old LP players were flutter and wow effected the sound. In digital realm if something is bit perfect, it's just that. And with a working network connection things are bit perfect. Even CD players have error correction algorithms which negate random errors in playback. The cd must be really damaged for the bit rate to drop low enough to cause a drop in bit-depth. To put things into perspective, imagine downloading a 1Gb movie or an ISO image for windows install for example. If even one bit transferred in that download is not bit perfect to the original, the end result is corrupt. Then ask yourself how many times your downloads came in corrupt? Never, unless you have literally broken hardware.
@@dingdong2103 Well, it’s because in the end, we listen to an analog signal that air-pressure waves are. It also doesn’t really matter whether a digital audio file as a whole can be transported without being corrupted, but how we sequentially feed data it contains to the D/A conversion process. All of a sudden, clocking and a constant data flow do matter. Providing all that costs a lot of effort, which in turn, introduces variable noise. How much of it propagates to the audio channel is another story, and Paul has elaborated on that to a great extent over the years. This is also where separating different components and processing stages becomes essential.
@@themarcinmm4840 Incorrect. I suggest you study the AD/DA process a bit further, you'll learn that also there are built-in error corrections and buffers which make the signal totally impervious to slight delays or noise. The end result is just as bit perfect as was the transmission.
@@themarcinmm4840 "Well, it’s because in the end, we listen to an analog signal that air-pressure waves are."
Correct
"It also doesn’t really matter whether a digital audio file as a whole can be transported without being corrupted, but how we sequentially feed data it contains to the D/A conversion process."
Correct
"All of a sudden, clocking and a constant data flow do matter. Providing all that costs a lot of effort, which in turn, introduces variable noise."
Wrong
It is very easy and inexpensive to ensure that the DAC is "feed" with constant data flow by adding some digital memory working as a buffer in front of it.
Think about portable CD Players or CD Players in a car - if there is a heavy "bump" from outside the laser looses track and needs some time to position again but this does under normal circumstances not interrupts the playback as there is buffer which stores severals seconds of music in advance - given the laser enough time to reposition. And as the player reads the data with an much higher rate than needed for the playback the buffer is soon filled again - ready for the next "bump".
@@themarcinmm4840 "All of a sudden, clocking and a constant data flow do matter. "
But for clocking we can use quarz oscillators and feeding an DAC with constant data rate using such a clock is no problem at all today there even an external USB Harddisk (not even an SSD) can easily transfer 100 MByte per second.
Packet loss - read up on it. In audio or video streams, between 1% and 5% is considered acceptable, by some. Paul's description hits the right note and understand that an audio stream may make a number of "hops" in long distance paths (open a command shell and run the "tracert" command - a variant of ping - to see how many hops are in the path of your audio stream.) Each hop involves an active component like a switch, gateway or router - where packets in your stream are buffered a fraction of time. When there is no packet loss and the receiving end has no issues, the active components in the path should not influence audio quality in themselves. The occasional packet loss, however, in some cases may cause the receiver to ask a resend to the sender and you have to wonder what happens to your stream - there may be dropouts. At some point, the audio stream may arrive in (A) your network computer that feeds your DAC, or (B) a DAC that is capable of receiving a stream itself. here, buffering plays a role again, essentially to prevent being bothered by dropouts from packet-loss/resends. If a difference could be audible, then where do we expect it? In case A in a DAC that gets fed its audio samples (TCP/IP packets unwrapped and the contents put into sequence in a buffer and next presented, likely over something like USB) in a synchronous way, where the PC or Mac determines the clock of the audio samples. This can be addressed by an "asynchronous" DAC, which does a form of bus-mastering in that it asks for samples and buffers these. That DAC needs to put the samples on a clock again. Here, wired networks give the best reliability with the least drop outs, as a general rule. But (I use shielded Ethernet) there likely is some ground connection between network components in a cabled situation so if the digital link to the DAC-core is not floating relative to ground, then stray electrical signals can seep through the DAC into the analog section. This, though, will never happen in a perfect world populated by perfect connections, like we can see inside PS Audio ;). In case B, the DAC integrates a PC-like network section and the problem under A needs to be solved between the stages in B's design.
As to packet loss, the question is, when this happens, how the DAC handles this. The original CD playback system was designed to compensate lost bits that could not be retrieved from the CD by filling the blank with samples extrapolated from the previous ones. If the stream-receiving DAC does this with packet loss, then the occasional drop out is not heard. As to Quality of Service (QoS) which is a mode you can set on your switches or routers, it is implemented generally in such a bad way that it may be detrimental rather than beneficial. Yes, as Paul says, a switch is a computer. And like a Windows PC, its performance may already degrade above 3% CPU load when it has to do smart things (protocol conversion or QoS).
So, in the network we have to look for packet loss and how this is handled. Then we have to look at the step from network to DAC as well as inside the DAC how this works. We could argue that wireless does not have the inherent grounding issue, which is true, but a component that is sensitive to some leakage from the digital front end to the analog backend is already inherently sensitive to this problem in itself (if it does not transfer between stages over floating links and the individual stages have grounding issues.)
If we cannot imagine something, then this does not imply our conceptual models we use for explanation are valid, or relevant. It just means we cannot imagine it. And have to be careful with opinions.
My challenge: try to refute these claims in triple-blind experiments, set up in a way that can demonstrate these points I make here. Use extremely revealing content (no studio mix, no multichannel recordings) with exceptional real spatial clues of a programme with width, depth and height. Invite a musical blind person who can tell you where the windows in your room, are to join in on the test.
SO well put, bravo! I'm a retired network engineer and designed networks for national security systems and for some of most cutting edge agencies in the world. You explained this very, very well.
@@skrutinizr9372 - thank you sincerely
"So, in the network we have to look for packet loss and how this is handled."
Packet loss were so rare today (except for Wifi under some circumstances) that you can stream for months before even a single package is lost. And it is well known how such an error is handled. The TCP/IP stack detects the error by verifing the checksum and requests a retransmission of that packet.
So a packet loss does not mean that any data is lost.
"My challenge: try to refute these claims in triple-blind experiments"
Wouldn't it be more reasonable if the manufactors of such "audiophile" switches would proof that their devices really have an impact on the sound quality by either performing measurements or triple-blind tests?
"The original CD playback system was designed to compensate lost bits that could not be retrieved from the CD by filling the blank with samples extrapolated from the previous ones."
That is only half of the truth. The audio CD standard defines a special encoding scheme (based on Red Solomon codes) to ensure that nearly most wrong bits (either to reading errors or scraches on the CD or even errors which were already on the master used for pressing the CD) could be fixed perfectly. Even sequences of errors like unreadable bits in a sequence due to a scratch) could still be corrected to the original data.
Only if there are so many errors that this error correction could not fix them anymore then samples are extrapolated (or even better intrapolated) so minimize the hearable effect.
A shielded switch with a good power supply might help a bit, depending on the location of the switch.
It won’t make any difference. It’s a digital signal.
Paul, anytime you’re going through DFW airport, you’re invited to my house about 15 minutes away to listen to my system streaming with and without an EtherREGEN switch. If you can’t tell the difference in your system, there’s something else going on. The difference is far from subtle, especially with percussion.
The EtherREGEN from Uptone Audio makes a different, I noticed. It’s not huge, but still a somewhat better sound because of noise isolation.
I was sceptical about audiophile switches too so I bought a cheap Teradak one with txco clock and linear power supply of Ali Express and immediately noticed a big improvement. Bigger soundstage, better texture,more bass definition and more sparkle up top all for around $300 well worth it.
Sorry I can't post a link but everytime I post any good product that's not PS Paul deletes it.
2028: PS releases audiophile switch, "Hey we admit that we got it wrong, digital is more complex than we understood, and and our PSwitch is the best in the market. It'll elevate your system"
Thanks Paul, good video and you probably just saved some people a heap of money as those audio grade switches are hideously overpriced for what sounds like negligible to no benefit.
People are paying hundreds of dollars for a $20 switch.
Not so fast Paul … surely the Paul on the video didn’t plug a SoTM switch on his system. A topical case of someone that didn’t do his homework.
I advise anyone who has this same disbelief to test drive one of these switches before make a stupid comment or vidro about the subject 😉
@@crventura I’m an advocate of “ Don’t nock it until you’re tried it” so fair enough. I must confess that I rarely stream music from streaming services. But I have ripped my digital collection to a Drobo NAS and that would probably benefit from audio grade switches and network treatment. From CZcams I find Hans Beekhuyzen more an authority than Paul M on this subject. I do however only listen to vinyl or CD’s for critical listening. Thanks for your comment and since posting that comment I have learnt a lot more.
@@paulpavlou9294 "From CZcams I find Hans Beekhuyzen more an authority than Paul M on this subject. "
Hans Beekhuyzen talks a lot but his understandig of digital networks and protocolls is very limitted and sometimes totally wrong.
He never proof his claims by measurements or double blind tests but instead he believes uncritically what manufacturers claim.
Ultimately, he is still trapped in the analog world where every noise in the signal chain had an impact on the quality which was true - but is no longer valid in the digital world.
I spy a SADIE audio workstation. Great to see hardware like this still in use. I use one all day/everyday even though it went native yonks ago. Amazing piece of kit.
It was hard for me to believe, but I found that using a Mac Mini as a source connected to a DAC by either optical or USB sounds better if I am connected to my wifi router using 5 ghz rather than 2.4 ghz. The only thing I can account for the difference is less interference on 5 ghz from other nearby wifi routers on the same channel, because 5 ghz has less range and doesn't penetrate walls very well. It is possible, if there is no other wifi than yours reaching the computer, that there would be no difference at all. My findings were vindicated when I later read in UHF Magazine that they found the exact same thing - 5ghz yielded better sound with their Mac mini. They had no explanation. I know this isn't an ethernet switch, but Paul did mention wifi as well, saying he heard no difference. It is very likely that no neighbouring wifi reaches PS Audio's listening room and therefore the signal is free of interference from another router on the same channels.
Wi-Fi is a whole layer of problems added onto wired ethernet. But, in theory, the problems of the network transmission channel are completely removed by the streamer's jitter buffer.
It's amazing. The presenter doesn't seem to have tried different cables/switches himself. The people commenting miss the key point of the video and most haven't made comparisons themselves. People, you are looking to hear something from youtube that supports what you feel must me true. Stop doing that and try it for yourself. In my situation a self made cat 8 cable vs a cat 5 cable was a clearly audible difference. Unexplainable and confirmed by blind listening test with audio buddy. Upgrading my network switch to a Bonn 8 was a huge difference, I mean literally huge. Perhaps it is due to eliminating noise entering the streamer, perhaps something else. Whatever way, it was the best money I spent on my system in a while. My message: please try it for yourself and stop all this online bitching, it's exhausting.
I had the same experience with Uptone Etherregen, the improvement in sound quality was clearly audible, and no, that is not a bias effect or anything like that. Clearly there are things going on here that we simply do not understand yet.
@@blakeyisgod Nonsense. Expectation bias. Checkout Audio Science Review.
@@dfpurton you heard here first folks! Don't listen to hifi before you buy it. If it measures great it must be great! Daniel, you are missing out a lot with that mind set
Nice comment, GD.
"In my situation a self made cat 8 cable vs a cat 5 cable was a clearly audible difference."
If you make the effort to build a self mde cat 8 cable when your expectation that this will sound better must be very high.
That is known as expectation bias.
"Unexplainable and confirmed by blind listening test with audio buddy. "
Unexplainable things needs a little bit more than a self performed blind test with your body...
How was the test protocoll?
How many rounds did you perform?
How high was the percentage of "hits" / correct guesses?
Was in each round determined by a random generator which cable should be tested in the next run - for example using a coin.
Were you and your buddy at any time while the complete test was performed in the same room at the same time?
Was the test observed by a neutral person?
I'm in It Networking the last 24years and you are right, IPv4/v6 protocol is from a pure Hifi standpoint a bad transport protocol. At the end every streamer of course does buffering and therefore a switch for 100 or hifi-optimized 1.000$ makes no difference. Said that, i own a paul pong dual switch and if course we know power supply matters, for all digital Devices. Do i hear a differenfe? Technical: NO Psychological: YES :-)
"we know power supply matters, for all digital Devices."
Fully agree - nearly all digital devices stop working if they not have a power supply.
I don't know about audiophile switches but introducing an optical network ( two media converters connected together with an optical fiber cable) between server and router brought about a big improvement in sound quality.
Ah, honesty. Anyone who understands packet switching knows it can't affect sound quality.
Hi Paul, I don't have an answer to that question... however, your arguments could easily be applied to the power network, and yet you find clear differences there.
No, the data correction of digital data is in no way analogous to power, apart from it travelling through a large network.
They are completely different conceptually - I think you are very confused.
Hi Paul, seems you are locked down in your office :). Love the frequent uploads. Stay safe.
ive found that my cheap switch connected on the same powerboard as my hifi makes alot of noise
Switches dont matter but cables do?
There will always be people who think that if it cost more, it sounds better. 1's and 0's sound the same, regardless of how much the equipment cost that delivered it to your audio equipment. There's no extra "noise" or anything, it's just a stream of 1's and 0's. A consumer grade, regular router can handle a LOT of simultaneous audio streams easily as long as it is configured properly and using reasonable, but not necessarily "fancy" cables.
The snobbery in audio and pseudo-science in audio equipment sales is laughable to anyone with a technical skillset. Any other industry would get shut down for fraud, yet so many people buy into it (literally).
Paul’s absolutely correct on this one. We’ll 98% right. Regardless of how those packets got to that PS Audio DAC, and I don’t care how much ca ca, delay and jitter is in the packet stream, that DAC is really getting it’s content stream from an internal buffer. And I’ll assume the buffer(s) in a PS Audio DAC are HUGE!!!
Think of it as a bucket of water. Poke a small hole in the side and bottom of the bucket and fill the bucket with water. The water is going to leak out of the bucket at basically a constant rate. The water filling the bucket comes in at a rate that’s all over the place. So what your hearing from the little stream of water is completely un-effected by the rate the bucket is being filled.
Ya Ya what if the bucket is almost drained. Well most good circuits dealing with an uneven incoming data stream “double buffer”. And I’m sure you’ve heard that term before. The first bucket fills up, leaks out of a little hole, that stream fills a second bucket and the stream of water leaking out of that is your treasured music stream. Double buffer chips will set you back maybe, 25 cents?? And making the buckets 10,000 gallon storage tank(s) may cost you another 10 cents.
BTW those one or two buckets are gonna initially fill up in less then a second with minutes worth of content.
I can remember double buffered digital receiver circuits in the 60”s. Remember dial-up modems? They had double buffers on the incoming data streams.
It helps a lot if you are using your computer as the music server. Its not about errors in the actual digital infomation its all about reducing jitter by reclocking it when sending it to the dac. Its the same reason why the same file always sounds better directly played from a transport than a computer hard drive or streaming. Just burn a disc and try it yourself
What he could do that would benefit him if his router that he got from the provider supports it, which is likely that it does… Is an able quality of service or QOS. You can use QOS to provide guaranteed quality for specific types of data that are passing through your router (streaming audio, streaming video, etc.) and let it de-prioritize other things like regular downloading or email pushes or smart TVs that are scraping data.
If you have a decent router you can classify all the different types of data that commonly pass over your network and tell the router which ones are the most important to you and which ones you don’t care about. They can help you a great deal if your son Starts playing halo while you’re trying to listen to the symphony on Tidal.
QOS only really matter in a heavy-congestion environment--where packets that are sensitive to latency might end up behind a "'sizeable" queue of packets waiting to transmit. It would allow them to skip the line, as it were. If their network environment isn't used particularly heavily, it doesn't really do anything as the queues are cleared as packets arrive.
Love your vids Paul but have to disagree on your point about wifi vs ethernet cable.I have a relatively high end system which I find quite revealing in that component changes and even cable changes usually result in subtle and not so subtle changes in sound.I found a definite improvement in sq going from wifi streaming to ethernet cable connection.I suspect the high end passive preamp in my system plays a role in the transparency of the system as every active preamp I have heard seams impart it's own colour on the sound making it harder to hear difference in different components/cables.Just a thought.
Agreed completely. I'm a network engineer. As long as a switch is minimally functional, sound quality will be unaffected by the switch in use.
But did you actually try and listen in a capable stereo system? Or are you just rejecting the idea of audible differences because you are a network engineer? I'm working in IT myself for about 30 years and I can clearly hear differences between switches.
You are a network engineer and you're giving your opinion from a regular network usage point of view. That doesn't make you a audio engineer that tested audio quality switches in hi-end audio systems, in order to reach a conclusion on that subject. That's a totally different matter.
@@andru2625 So true! 👍🏻
dont forget to ONLY, use a cat5e cable from the home depot in alvin texas, or you wont get the best audio quaility
Nah, you can use cat6 or even better cat7. Cat5e is fine though.
He is correct. Digital signal. Not matter. Well said pual.
Finally the right answer!
Or you must have a very cheap $5 switch from China, which isn't capable to do it's normal job. When a switch is capable to provide a stable connection with enough bandwidth, than it should be OK.
If you have very haevy video streamers in your household, you could consider a switch that can do some traffic shaping to guarantee a minimum bandwidth for audio streaming. But even these 'smart switches' are very affordable.
what computer parts are made in the US? just saying
@@ixfxi That could be a consideration. But isn't it more important to support your local supplier instead of giants like Amazon?
Gerrit, the subject matter is not predicated on what is "OK". I think we need better videos and not just a bunch of talk.
This is correct. Unless you're seeing packet congestion on the switch (from over-utilization on the LAN side plus saturating your broadband connection), a better switch (not an "audiophile" grade ethernet switch) will not make a difference. And like Paul eludes to, a better switch (in that situation) would only help reduce packet retransmission and congestion *locally*, which whatever streaming device you're using already handles by buffering. The same logic applies to WiFi. If you have a lot of WiFi devices that are competing for bandwidth/spectrum on a single access point/router (or have other WiFi routers around you on the same frequency/channel), over-saturation can occur and packets can be dropped or delayed. Again, this does NOT affect sound quality: only the timing of when packets arrive to the device (which get put into a buffer).
Well said, I know sometimes when you try to break down something to simply if, some of the tech heads get upset, but really, job well done on this one. I think what would make streaming better, the the dac you offer, as I understand it, you buffer in the input data, the clock it out with an extremely stable clock, basically jitter free.
Why is nobody discussing audiophile grade oxygen-free air?
The video which could have been a single word: no
Well... I think you missed an important feature while commenting right at .41, clock input!....
I think the problem could start at the point, where the signal is being processed from Ethernet to USB or inside your DAC if it has built-in network functionality. Ethernet chipsets (also wireless) can be problematic in terms of electrical noise. In my opinion, you should try to isolate Ethernet as good as possible from critical devices in you audio chain. But for streaming, you need a network. There is no way around it. I think the most elegant solution is to introduce optical Ethernet for the one connection of your streaming endpoint device. It provides galvanic isolation and therefore eliminates electrical noise. Every device before that, does not really matter. And of course, your streaming endpoint should have a highly optimized Ethernet chipset, that does not introduce noise again... For me, this solution is the opticalRendu.
I would bet 10 bottles of german beer that all streamers are using the same microcontrollers with integrated Ethernet controllers than they are used for other (non audio) purposes also. And I dont know any Ethernet chipset ever optimized for audio.
So if I undertsand correctly switches are not sensitive to noise because the IP is digital and it is either gets a packet or not. But how come that computers as such are very sensitive to noise ? The are also digital aren't they?
The ethernet switch is too far upstream from the business end of the audio for that to matter. When you stream music from the internet, it is not streamed in real-time. Soon as you hit play on a song, it most likely downloads the entire song within a few seconds, so the entire PCM stream gets queued up in the computer's buffer well ahead of when it is needed. Now how the PCM stream is delivered to your DAC is the next challenge. Usb? Optical? SPIDF? All of these require real-time streaming and jitter and noise come into the picture.
I didn't really understand what he meant by saying that computers are sensitive to noise. They actually are the ones that introduce noise into audio. For that reason sound cards on motherboards are isolated from the rest components on a separate circuit through electrical filters. Howerever this is not ideal as induced currents can still appears from the monterboard components. That is why people use DACs and audio interfaces that are physically separated from the computer.
What's worse is that there are people who can even consider asking this question!
I am a manger over IT networking and security. I thank you for your video. I personally prefer a wired over wireless as the protocol for wifi is more chatty than the traditional ethernet protocol, sort of like the old school hub versus a switch. My suggestion for users is to create a separate SSID for all 2.4Ghz devices and a separate SSID newer wireless AC supported devices. You DO have to be careful on wireless as NOT to oversaturate the access points. Even with high-end enterprise grade wireless access points, 10-20 devices is a good place to be but I do not recommend anything over 30 per wireless access point. purchase more access points if needed and spread the wireless love to your devices. If you have a 2 story house, depending on where the router is located, you can install an access point (wired back to your router for better results) and have this access point on a separate SSID; per careful of stepping on your wireless channels. This will help to improve performance. Wired is still best, but not alway feasible. I wire all of my components, tvs and use wireless for devices such as smart home devices, and for laptops/phones/tablets. A friend of mine is a wireless expert; meaning Cisco and other vendors reach out to him for testing enterprise wireless products. You can get more tips and hints on his home page. www.my80211.com/
Hi Paul, I would love to know what switches you have personally tried feeding your Streamer/DAC. Much appreciated. Paul
There are a lot of experienced reviewer who describe heavy differences... Why should they lie? There are a lot (!) of things happen in electronics regarding HiFi presentation and our ears are that sensitive. So if cables matter, why should not switches should matter?
About time there was a bit of truth around this topic and not more snake oil. Well done Paul, and thank you.
Digital audio is not my forte. Vut, if you are happy with the audio reproduction your current set up is providing, then I can’t see how putting another device in the signal path can improve things.
Enjoy the music and remember, music was born analogue!
A switch is just moving packets and frames it's not analog audio it's 1's and 0's.
Over short distances cable doesn't matter but if you're doing any sort of long distances it can increase your losses and drops that can matter. Realistically I would stay away from cheap switches like anything that's like $10 because that's going to induced noise from whatever crappy power supply they're using but more than likely that's just going to cause connectivity issues not sound quality issues unless there isn't proper isolation in the dac.
I watched in complete astonishment some CZcams Audiofool claiming that he could detect the subtle nuances between several different “audiophile” switches…it’s beyond hysterical. 🤦♂️
In low quality routers you have low quality clocks very bad quality ac power, that results in jitter and bad power in rj45 going inside your server. Ok to normal computer work, but not for music or video. A good switch corrects everything. Don't compare yours Enterprise network with a normal house network.
So true! 👍🏻
You let me know when the low quality clock and ac power matters: czcams.com/video/u6mYM8Xy0H8/video.html
"A good switch corrects everything. "
How can a "good switch" correct the "low quality clock" build in a cheap streamer?
A streamer writes the "perfectly timed" data bytes he receives other Ethernet into its buffer from there it is readen with a speed defined by the streamers clock and send to an DAC
Dear PaulThank you for great videos, most of them are very educating.I agree on many of your obviations, but I do not agree on this: Ethernet cables do not matter.For the last 22 years have I been working with IT and computers professionally, my hobby has been Hifi and Electronics for over 40 years.I have built my own DAC and I have been working on a Dual PC solution as Media Server for my DAC.
1) Ethernet cables from my fiber connection box to my Router/switch, does not matter.
2) I run Wifi between my Router to my Control PC (Media Server), so here does Ethernet cables also not matter.
If I did connect my Control PC (Media Server) to my Router and also uses a NAS which is also connected to the Router, when would Ethernet cable matter.
3) My Control PC (Media Server) is connected to my Audio PC (here is the stream transformed from network to USB) here does Ethernet cable matter a lot!It all comes down to, how your system is designed if Ethernet cable matter or not.One have to have a huge knowledge to be able to analyze the need or not, so my recommendation to the average user,
do not use money on Ethernet cables, your money is probably spend better elsewhere in your setup.Thanks!
Jacob Land Jensen
Denmark.
Jacob Land Jensen probably try and get an authorised audioquest dealer in Denmark to loan you a pair of diamond ethernets and give it a shot...
Yeah, let's hope this myth doesn't get any traction. On the surface it seems obvious why it shouldn't matter.
Never underestimate human stupidity
jcat.eu/product/m12-switch-gold/
Marc Ozor Shame on the people selling that snake oil product dude. It’s a crime.
Too late. It's all over Audio Asylum digital forum with full support of the golden ears, science be damned!
Andy Delle It doesn’t matter if 90% of the world believe in audiophile Ethernet switch snake oil. It’s still snake oil.
No doubt a switch with a good linear power supply will sound better. I've done the tests, it's all about the quality of power and lack of noise. Everything matters.
Bullshit. Show us the results of your tests.
@@Iblis666ful Do your own tests.
@G Guest With very revealing hi-fi components you can hear the differences. Mid-Fi and lower grade you won't.
@@spotshooter1
www.audiosciencereview.com/forum/index.php?threads/silent-angel-bonn-n8-audio-grade-ethernet-switch.12360/
www.audiosciencereview.com/forum/index.php?threads/uptone-audio-etherregen-switch-review.10232/
@@Iblis666ful Seriously, do your own tests. With my Naim ND555 steamer you can hear the difference between a switch with a switch-mode power supply vs. linear power. Noise is always the issue... not the packets.
if you know a bit about computer then it is more about whether you have good wifi signal or not. if you have weak wifi signal and the speed gets slow then the music can stop playing. As long as the player receives a data pack of music, it plays as well as a CD player!
It is like transferring a picture via bluetooth!
All routers do a good job as long as the wifi signal is good or you use the network cable!
I dont have the technical background to take a stand on whether switches matter or not and I have never tried an audiophile switch on my system, so i am happy to sit on the fence.
However Pauls argument as to why switches can't make a difference ie the signal is being bounced around all over the place before it reaches your house so the last link in the chain cant make a difference is the same argument power cable skeptics use. Yet Paul has previously dismissed this and argues that power cables do matter because he can hear the difference and that the fact he can not explain it scientifically is irrelevant.
This seems to be a case, with all due respect, of having your argument and eating it.
In this case many people claim to hear an improvement with switches, like dear old Hans, but Paul does not, so therefore switches don't matter.
As I said, I have no skin in this game, but I think, unusually for him, Pauls argument is resting on pretty shaky ground if you compare it to previous utterances. Either subjective listener experience takes precedence or it doesn't. Conversely the scientific consensus takes precedence or it doesn't.
You can't have it both ways.
Thank you, Paul! Great video!
Hi Paul - from my PC using the built-in command tracert in a console window I see there are 10 hops from my PC to TuneIn.com - 4 hops to before I get into the internet from my provider alone! Takes up to 23ms to get here. Streaming 320 kbps gives *audible quality improvement* over 128 kbps. All the best from Switzerland, Rob
I added a basic Netgear 5 port switch between my Comcast modem/router and AV gear, and there was a very noticeably lowering of the noise floor plus all of the typically associated benefits. I think a switch (even a budget one) will be better than running ethernet straight from your noisy Comcast device directly to your sensitive AV gear. Worth a try folks, if you don't hear an improvement send back to Amazon.
Did someone say "BHK Preamp"?. Paul will change his mind on this eventually. Just wait and see.
the most ridiculous thing is that in most homes the internet connection arrives on a 0.2 euro / meter twisted telephone cable and even hundreds of meters long .....
audiophools can cook perfect Beef bourguignon out of rotten meat just by adding audiophool grade salt.
Hi Paul, I maked the test with SOtM switch sNH-10G with externel Mutec Ref 10 Master Clock. Difference with commectial 8 Gigabits swith is BIG. Very low noice, great instrumental separation and like never leasten before sound images in my system. For sure problem isn''t in the IP data package but il the last digital segment from Router Power Supply (very noising switch), your position is very basic there is a lot of technical question mark inside any audio system at home that probably the audiograde switch can fix. PS Audio is big company so you can make the same test with your Directsctream and FOR SURE you will change your opinion on this. This also happened to me.
Couldn’t you same about incoming power?
There is more to digital than zeros and ones. It will take some time for pure objectivist to accept this, and that is fine. We all need time to change our beliefs
@Moussa....whilst I'm not disagreeing, just making a blanket statement means nothing
Please explain. Thank you!
@@FooBar89 whilst I appreciate your response TLS....was wanting to hear from Moussa about what he meant. It's just too easy to make a condescending comment to all readers without actually providing some detail
It's a bit like me saying there's more to photosynthesis than the sun....we will all change our beliefs in time.
There literally isn't.
@@nostro1001 I am not an engineer, but I have been comparing digital gear in my system for years now. That includes USB cables, ethernet switches, servers etc. There is always a change in SQ. If it were only 0s and 1s then nothing should matter. Of course this is based on my subjective ears.
@@moussaobeid Oh right....so no qualitative evidence, or actual science, just your opinion. That's cool, your initial comments led me to believe you had some deep understanding about what you said, so I wanted to learn. Cheers!
In the bowels of Isengard, where uruk hai and PS Audio products are crafted...
So he's got "what" attached to "what" with "what kind of cables" to a "what"? Boy, am I behind the times. I still listen to albums and CD's. LOL!
Steve Fick you stick to your CDs bro, locally stored music in my personal experience is king, streaming is good, but doesn’t compare from a sound quality perspective.
Rami Bachar indeed, listen to what you want. I’m just pointing out Steve’s locally stored music is most likely to be of better quality than streamed music.
I think you mean “long playing” records (or LPs for the racy types).
In my humble opinion, perhaps banking and financial industry shall be targeted for marketing with high grade router / ethernet switch rather than HiFi industry.
New York Exchange are deailing with billions of dollar / stock transaction each day, visa and master are similar...Their data accuracy rate (associated with money) are far more important than music data (audiophiles).
If banking and financial industry (even military centre or NASA) are comfortable using ordinary commercial grade router / switch for their routine data processing works.
It seems that some audiophiles without reasonable logic for using so-call higher grade router/switch for routing their music data!!!
I think you will find that they don't use consumer grade ethernet switches. Most use very expensive managed switches.
Is that a Sadie hardware?
I would agree that using an audiophile switch for streaming Tidal or Qobuz won't make any difference, it's never going to sound great. However, streaming music files to your DAC from a NAS on your own internal network could very well be improved by better quality switches with better quality clocks and power supplies.
Cisco Catalyst?
An $18 switch improved my system.
TCP/IP FTW :) Plus any playback device worth its salt will have a sizable buffer too.
USB falls in to the same boat. It works till it doesn't. The cable can pick up some noise but if the gear was built correctly this noise will be easily filtered out before it effects any anallg audio. If the cable and gear was built to basic industry spec changing to a fancy USB cable shouldn't make any difference. If it does, just means that piece of gear wasn't well engineered or done on purpose so their partners in crime, cough "audioquest" can sell you a expensive USB cable. People that buy expensive usb cables probably aslo buy gold plated optical cables....lol
Sry i disagree going from MacBook Pro to Chord Qutest to Atoll AMP ..
and i was skeptical but the 500 Euro cable is day n night between the Standard Audioquest Cable. Streaming via Quboz in Hi Res
@@borisbobowski628 Well precision testing and measurements done by AudioScienceReview and Audioholics say otherwise. If a USB cable is effecting say a DACs audible output then its pooy designed, shouldn't happen if engineered properly. Chord gear is nice but tends to measure poorly for reasons even most CHi-Fi (Topping, SMSL) gear has been able to overcome at a fraction of the price. I have had about 20 different DACs and and several cables, none have made a difference for me.
@@iowaudioreviews i think No Audioholic can denay that different Dac's Sound different or more transparent. I can clearly say and ofcourse measure that my old DAC has Not a slightly Chance against my chord DAC. I dont understand why U was buying Dac's If you didnt Hear a Difference after lets say the First 3 Dac's? Something Must be really unbalanced in your Setup.
@@borisbobowski628 Most were sent to me to try, some barrowed from friends and others bought and returned. Most places have a 30 day trial or return policy. Some sound slightly different and others can't be discerned from each other. Got boring so I just use a miniDSP, REW/mic, and EQ APO/Peace to fine tune my system. Its made the most noticeable improvement for me and has tons of flexibility. When the techy stuff gets annoying I just fire up the turn table and grab a Parts Express catalog and forget about it.
@@iowaudioreviews OK Hearing ni Difference or slightly differences is indeed a big Difference. Im blessed with really good ears, No Joke. And i hate IT that i really hear this tiny differences because to me its Like opening a Door...you cant Go Back Afterwards.
If you own PS Music Direct
you really need a switch
Firewire & optical 🤘
This podcast was so noisy due to all those routers I had to add an ethernet switch between my laptop and headphone amp in order to hear it. (j/k!)
We don't hear the stream of bits from the switch, if anyone can they must be a cyborg. What matters is the output from the DAC.
Exactly. The clock in the DAC will handle any potential jitter from a switch. Noone ever published a scientific a/b listening test proving anything. Or better still, a measurement on the signal coming out of the DAC comparing cheap versus audiophile network switches. That signal will be identical.
@@JustBrowsing777 that Dutch audio guy swears he hears a difference in an obscenely overpriced "audiophile" switch and gets pissed if that is ever questioned. No problem with me, it's his money, not mine.
@@nunofernandes4501 I like Hans channel, he taught me a lot, but as with anyone's claims you always have to question them and not just accept them as truths/facts. I've struggled with what is in fact impacting audio in the range I can discern better from worse. It made a huge difference to switch out my amp and streamer/dac but power cables, speaker cables and interconnects....not sure I can hear anything and if anything very little in my system. Switches no way😄
@@JustBrowsing777 yes, I like Hans's channel and learned a lot from him too, but I don't agree with that particular point about switches.
As for interconnect cables, in my experience there was a clear improvement in sound when I upgraded from €10 but decent quality supermarket RCAs to AQ Evergreen for my DAC and AQ Tower for my CD player. These cost me around €30 each and the results were so good I doubt spending any more would be worth it for my personal pocket depth.
@@nunofernandes4501 I agree with the interconnects. A mid range pair is fine as long as you avoid the cheapest ones. I have seen/heard comparisons on these with a clear audible difference.
I found that moving the switch further away from the music gear made the biggest difference. Since the psu you receive with the switch is usually a piece of shit and more of an emi source than anything else.
Why don't you tell Hans Beekhuyzen that internet switches don't matter (and I agree they don't). He seems to thing a $600 internet switch made an enormous difference in his system ...
Hans did not conduct a blind test. If somebody exchanged $20 for $640 Hans would not hear difference IF NOBODY TOLD HIM ABOUT EXCHANGE.
So, Paul...What You're saying is...it doesn't matter what kind of conveyer belt delivers my garbage to the dump? Lol
The Hans has videos on Switches if you are more interested audio quality over ethernet.
Hans is full of it and hasn't heard a tone over 16kHz for the last 20 years.
@@Iblis666ful I would bet no more than 12k ....🙉
I'd rather listen to a seasoned and successful designer about anything, for that matter. Instead of a gadget lover/hobbyist with lots of money, who also started rather late, because of his son interest in the field...
Get an optical isolation i.e. a regular single or multimode network connection between your streamer and your otherwise noisy switch mode psu that powers the cheap modem/router or whatever delivers your Internet connectivity. Is there any streamers with an SFP Ethernet port? Or wait just simply go wireless, since that also isolates your streamer from any noise on the Ethernet side.
But then let’s talk about that data integrity, how does streaming services like Tidal make sure that I get a bit perfect reception of that centrally stored audio file which also has to be identical to the production master. Do they rely on the CRC validation in the Flac format and actually discard and resend frames that do not compute or is that simply ignored causing dropped frames. Or even worse are they played back without validation? And is the flac frame footers CRC-16 enough to validate the various frame sizes?
Also how does the streaming services store the music files that are sent? Can they be subject to silent data corruption and what mechanism are in place to guarantee those original zeros and ones that the master intended.
That is an interesting topic, at least for me!
Ethernet is already electrically isolated. Magnetic, rather than optical - tiny coupling transformers at both ends.
Streaming services for music will simply detect lost or corrupt packets via checksum and re-send. VoIP handles it differently because it's latency-critical - by the time a lost packet is re-sent it'll be too late to be useful, so they just attempt to interpolate the missing data. It works well enough for voice that the user shouldn't notice.
@@vylbird8014Yes, ethernet is galvanic isolated with those tiny coupled transformers. Still, that does not eliminate noise coming from the electronics. Nor induced noise from the cables. Also STP cabling could bridge the ground planes between the router and the streamer if used improperly.
"Do they rely on the CRC validation in the Flac format and actually discard and resend frames that do not compute or is that simply ignored causing dropped frames. Or even worse are they played back without validation?"
Tidal uses TCP/IP as transport protocol for their streaming data. So they not have to handle transmission errors or lost packages on their own - all this is done by the TCP/IP software.
"Also how does the streaming services store the music files that are sent? Can they be subject to silent data corruption and what mechanism are in place to guarantee those original zeros and ones that the master intended"
Streaming provides are storing their music on hard discs (and maybe some stuff that is requested often on SSDs).
Both are using also CRC checksums (build in already in the storage deviced hardware) to identify read errors (or data corruption on the media itself).
Such harddiscs are typically bundled to RAID-5-Systems which offers another level of error detection and also error correction.
The chance that such an error is not detected by the CRC checksum is about 1 in 4 billions
When I eat chips it sounds better. Seems to do something with my hearing.