"We are controlling transmission. If we wish to make it louder, we will bring up the volume. If we wish to make it softer, we will tune it to a whisper." -- from The Outer Limits
One of the greatest advantages of a high bit depth (eg, 24bit; 32bit, 64bit etc.) is the for any post-processing, the rounding errors become insignificant. Even digital faders become totally transparent when you use very high bit depths. In contrast, with a 16 bit fader you would easily hear the output going grainy as you lower it. With a 24bit, 32bit or 64bit fader any such grainyness disappears. Likewise, with AV processors, when you an apply EQ etc, it works much better when the input source music is using a very high bit depth.
Technically, the name Holland refers to just two of the 12 provinces that comprise The Netherlands, but people often use it colloquially to refer to the entire country. About a third of the population of the Netherlands lives in either North or South Holland.
From what I have observed over the years, many people forget about transients and peaks and think more about average level. When I compare the two, I’m not talking about seconds verses minutes, I’m talking about microseconds verses milliseconds. Whether it is your amplifier that can’t reproduce it or the recording chain that has compressed or limited those transients and peaks, once they are gone, they are gone. While a violin, may not have massive peaks, a drum kit sure does. To get the true nature and nuance of an instrument, you need those transients. That requires dynamic range. In most cases the ratio of average to peak is of several magnitude. In a professional recording studio (proper acoustics and sound isolation from the outside world) using world-class mics, properly set to record a drum kit, I can guarantee you will use every bit of dynamic range the mic pres and recording chain can muster to faithfully record what the player may be playing on that kit. During “tracking” of all instruments, getting a faithful recording of all nuances is important. You need dynamic range (among other things to do that). Keep in mind , when setting levels, you should be thinking about average max levels well below zero, or there will be transients well over zero. Keep the levels on the low side. During the mixing stage (depending on the music genre) compression will most likely be used on a track by track basis and again maybe on groups or stems. Hopefully, some of the transients and peaks are almost intact. If you think about the difference between a jazz quartet, a rock band and modern dance music, you get the idea. We still need dynamic range, but maybe not as much, as decisions have already been made. After mixing, the project will many times head to mastering. More eq, more compression and more peak limiting. After this stage it may be streamed, released for download, converted to MP3, burned to CD or possibly even pressed to vinyl. Hopefully they did multiple master sessions. As a side note, most broadcast FM radio stations have spent anywhere from $6000 - $20,000 for audio signal processors that do nothing more than alter the eq and dynamic range of their audio to “fit” their program material, target audience and make every song “sound the same” over the air. They have a government regulation that says they cannot exceed 100% modulation, peaks are their enemy. They limit or clip the peaks then compress everything else to raise the average level. This is done in a very sophisticated way where the audio is split, frequency- wise into 6-10 bands and each band receives different treatment (programable). Restoring dynamic range after it has already had the life squeezed out of it, is almost impossible (without coloring the sound). Having worked in live sound for many years, I have “Pro tools“ sessions that were recorded live that range from 20 something tracks up to about 60 tracks (per song) that have full dynamic range that I can mix, pan, eq and process as I see fit and listen to in my listening room. You have no idea how much I wish I had that for every artist I was interested in. Maybe someday we all could have that. Of course, If you were trying to listen to that late at night, in an apartment or with your kids trying to sleep in the next room……..you would be right back to FM radio……over headphones……..with a dynamic range of 6db. Seriously, don’t think about perceived loudness (average level) think about the nuances and peaks that you may not think you hear, but make a huge difference.
As a lover of classical music, we definitely need the abiility to portray a wide dynamic range. I was listening to a recent release of the Shostakovich 5th symphony and the magnificent slow movement Largo amply demonstrates such a need, with string tremolos barely audible one moment and triple forte outbursts the next.
I’m really glad Paul has been doing this channel. While there are times I don’t agree with him, I appreciate his passion and real world experience. It’s very rare that the top brass of a company is willing to get in front of the camera. Really proud of this guy… even bought his weight loss book.
From another Hans from the Netherlands: I remember having once sold a live record by Herman van Veen from the seventies because of the dynamic range being too big... I considered it unplayable in the living room with other people present. Most of the time one could not hear the soft parts and when the volume had been turned up sufficiently, people got scared by the loud parts
Yes, with music there’s a certain amount of dynamic range that sounds either “natural” (without obvious compression) or “kind of large”. If you go much beyond that, it just becomes bad music.
An extremely quite listening room has a background noise level of at least 30db. Usual noise level is 40db plus. Listening even at the dynamic range a CD theoretically supports and still being able to hear the quietest of passages would blow your ear drums during the loud ones. That's all Paul should have answered instead of circumvavigating the question. You can't reasonably have music DELIVERED with a DR of more than 30-35db and even that would be oftentimes too much even for in house listening. How you record, mix and master is another matter altogether, but in the end you have to limit the DR. Just not too much. 10-15db of DR in the music content is enough.
@@JonAnderhub I am sure the recordings delivered by Octave Records adhere to the above principles, no matter what Paul says to his audience. I would be extremely surprised if there is more than 20db DR in them, even 15db would be a welcome surprise. There's no need for more.
This is a poor explanation that conflates bit depth used in the recording process with bit depth in the distribution format. High bit depth is used in the recording process to give headroom for editing. 16 bits gives a dynamic range of 96dB and 24 bits gives 144dB. The decibel is a logarithmic scale, +10 dB is a multiplier of 10. The quietest studio microphone has a s/n ratio of about 92dB due to the Browning Effect, the random movement of air molecules hitting the diaphragm. A full symphony orchestra has a dynamic range around 60dB (10 bits). No home audio equipment can handle 144dB. Ignoring all the above let us imagine you have obtained a true 24 bit track that has captured a dynamic range of 144dB. A quiet listening room may have 50dB background noise. You play the track and set volume so the quietest sound can just be heard above background. Then there is a sound at 144dB. If this does not instantly destroy your speakers then your hearing will be destroyed.50dB+144dB=194dB which is in the lethal range (not joking). It is fortunate 144dB dynamic range is impossible for any home system. Even a track with 96dB dynamics can potentially cause ear pain and damage. Increasing bit depth has no effect on audio fidelity other than lowering the noise floor which is already silent with 16 bits.Most music recordings require a sensible level of dynamic compression to be playable on a home system.
Sometimes there can be overkill. A perfect example is Christoper Herrick's "Organ Dreams 2". Some soft notes you can almost not hear and suddenly the loud notes blow the roof off. I could be a nightmare for people living in apartment buildings.
Love the channel I usually get a cup of coffee and start my day with one of your videos. As a former recording engineer if you want to blow listeners minds, Recording and Close-miking a rock or jazz band without any compression at all is amazingly dynamic from the softest passage to an explosive drum solo. Most people have never heard anything like it and it will sound alien to some people. Tom Jung from DMP used to do some of this with his recordings in the 80’s/90’s that were recorded on the Mitsubishi X80 digital recorders but even he used some compression. The Flim & the BB’s albums are some examples of this.
There seems to be some confusion on the term Dynamic Range. There are two basic camps, one that considers dynamic range to be the difference between Peak and Average levels. In this case, many rock recordings are in the 6 to 10dB range, 10dB being relatively great. The other camp is that of dynamic range being the difference between maximum and minimum recorded levels. This can be very large, but still is only about 8 bits worth (48dB). Aside from those two groups, something that gets lost is resolution. For resolution, I consider 16 bits to be adequate, but just barely. Personally, I think that 20 bits resolution is good and provides some headroom for most recording streams. Of course, in our computing world, this rounds up to 24 bits that can readily be manipulated by our systems. I don't think that there is ANY piece of equipment in our hobby that can resolve the full 144dB range of 24 bits, so 24 is a bit of overkill, but reasonable.
To appreciate dynamic range you need a quiet listening environment. One of my pet peeves about car stereo is that inside a vehicle at 70 mph is anything but a quiet place. When I mention this to car stereo salesmen the invariable answer is to "turn up the volume." But what about the quiet parts? Perhaps this doesn't matter to the average driver.
Dynamic range in a recording can itself be overkill at times, believe it or not. If the listener has to adjust the volume mid-track because something is too quiet or too loud, it ruins overall immersion. Simply more difference between the quietest and loudest parts is not always better. There's a golden sweet spot between the extremes of dynamic range and compression.
Having asked this question that Paul addressed, I have read many of the comments below. It's been very helpful for me and thanks for that ! I referred to DSOTM for that's an album many of us know. I have it on CD's and DSD. According to Roon and Jriver the most dynamic track on that album is the track " the great gig in the sky " and it's about 19dB on both CD and DSD ( R128) . I'm not really interested in numbers but it just raised the question why botter to have 100 dB or more if we only have 15 dB or less recordings ? What I read from most of you, is that indeed recordings are often below 20 dB Dynamic range and hence we don't need more for it will be just to much for us humans to absorb. Is dynamic range than nonsense ? No it's not. more bits or DSD will provides you much more granularity ( provides space, generates more nuances, more details ) and these little details will pull you more into the music. If I'm taking a wrong turn here, don't hesitate to correct me 🙂
@Douglas Blake the second plane is something I started to think about to see if I can make the correlation between Dynamic range and distortion. Is it fair to say that when you have an amplifier with a high dynamic range the likeliness that when playing at medium volume the distortion of that amp is so low that it will not influence the added little nuances you will get when playing high res music? (High bit rate ) An amplifier with less dynamic range will likely have a higher distortion level at that same volume and therefore these nuances are lost ? Meaning if you want to have the benefit of listening to the little details you just need to have a amp with a high dynamic range ?
I really like high dynamic range so the tracks Paul has in mind can be lots of fun. However, if you want to listen to it, you need to play it always loud otherwise you can not hear the low passages, and that is not always possible. I have a CD with the Bolero on it. Living in an apartment I can not play it start to end at one volume, so I hear the music at the start and not annoy the neighbors at the end of the CD.
I think in rock/pop,etc., there has been a move to make everything sound near field, particularly vocals where for some reason they want you to hear the vocalist's tongue leaving the roof of their mouth.
dynamic range is important, and more than we need of it is great for master recordings and the future. however it can be pretty brutal for casual listening to music, depending on the system and person some recordings just are to intense and kick like a mule :)
Definitions vary, but "dynamic range" of a track generally means something like the difference in dB between the volume of the loudest passage and the volume of the softest passage in that track. But by "volume" we mean the sum of all the sounds going on at that moment. Our ears can totally hear a faint violin solo thats many dB below the noisefloor. The woodwinds may be tens of dB below the brass in the same passage but be very audible. Overtones that tell us timbre may be dozens of dB below the fundamental. The tiny cues we use to perceive the room/space can be _scores_ of dB down from the main signal. Automated "dynamic range" measurements miss all of these things. But you do need the dynamics to actually reproduce them, and so you do need the bit depth to store them. 24-bit might still be overkill (as it exceeds the pain threshold of the human ear), but as other commenters have mentioned it helps with things like rounding errors. As the inquirer points out, a dynamic range of 15 dB is a common score for a well-mastered song. If that were truly the full story on dynamics, you would need only 1.5/log10(2) = 5 bits to reproduce it - 8 bits would be overkill. But search for 8-bit music - we millennials grew up listening to it in video games - and it doesn't take an audiophile to notice it is limited.
I used to own a DBX 3BX dynamic range expander. Yes I could crank up the dynamic range and it made the music sound artificial but dynamic. When used reasonably it would lower the noise floor of my cassettes and VCR tapes. You have to remember that not only does dynamic range affect the louder portions of music but can also lower the softer sounds like noise or certain instruments making it harder to hear the more subtle sounds. Probably the reason why I like vinyl.
While I do agree that modern music is unnecessarily dynamic compressed, I do not think there is anything lacking with the dynamic range currently offered by CDs. At 16bit CDs offer excellent dynamics. It's not the fault of the pcm format that the dynamics are not utilised properly and that the music is compressed in the mastering.
The best example for me, of an album with great dynamics is the Moody Blues' "Days Of Future Passed"...............the opening gong crescendo is amazing, as is the rest of the recording! Not bad for 1969!
Paul, The number of bits does not increase the amount the sound can go loud as any 8 bit, 12 bit, 16 bit, or 24 bit signal can be amplified to 0DB FS. (DB Full Scale) What changes is the resolution. It's like taking a battery voltage measurement with a 2 digit multimeter and seeing 12 volts, but a 4 digit multi-meter on the same scale can read up to 99 volts, one reads 12 volts, the other reads 12.57 Volts, a much more precise number. Bitrate on the other hand, limits the maximum frequency that is able to be recorded without aliasing. For example, CD quality can have a nyquist frequency up to 22 KHZ, but that would result in generation of many beat frequencies in much lower frequencies, so a low pass filter is used to cut the high frequencies that create aliasing. Experiment for the audiophiles. Use Audacity to make a pure digital recording of the following frequencies at -18DB. 18 KHZ, 19 KHZ, 19.5 KHZ, and 19.8 KHZ. Play it back on your favorite high quality DAC and sound system. Use a spectrum anlyzer to see the pure sine wave played back while reproduced at the 44.1 KHZ stereo sample rate. Now you know the reason Pink Floyd sounds so different between CD and an LP on the Dark Side of the Moon album. The coins in Money and the clocks in Time don't have a chance to sound correct on a CD. Most of the aliasing would be removed if recorded at 96KHZ with a 40KHZ Low Pass Filter. Paul, if you have the time, post a video with digitally generated test tones sampled at CD quality. Not sure what CZcams would do to the recording in addition to the original aliasing. A perfect DAC can't fix this problem as the aliasing is well into the audiable range of a few KHZ. This can only be cured by filtering out the high frequencies in the recording, so they are not recorded to produce aliasing products.
All music to some degree goes through compression. modern music much more then classical. personally too much dynamic range makes me crank up the volume for the small subtleties and then when a big ensemble hits it knocks me off the couch. I can't under my normal listening deal with that much dynamic range. some compression is needed so you can hear the subtleties.
Hello Paul. Just a quick feedback message to say that I genuinely enjoy these videos that you generate. I realize that you spend quite a bit of time creating these videos and it appears that you are honestly dedicated to what you are doing here. I want to tell you in clear terms that I appreciate your efforts and your dedication towards creating said videos, and I really enjoy listening to you, as you work your way through the various topics which you speak on. While I do not agree with every, single, one of your positions (what a boring world it would be if we were all exactly the same), I do very much respect you and your efforts in putting yourself out here, in CZcams land, spending a portion of your valuable time with us through these videos. You are appreciated, no matter what, and thank you very much for sharing some of your passion of audio with us! I genuinely look forward to your next educational video Paul! Thank You! Fred
That database seems reliant on voluntary contributions so not a good start in my opinion. Dark Side of the Moon, recorded at Abby Road Studio 3 would have used 30ips TBR 2 tape machines with a (for that time) clean 40db of dynamic range. Later it was digitally remastered to PCM. Old masters can be tweaked quite well with a competent engineer. There is also a large difference in the dynamic range of an LP depending on the equipment used for playback. Some analogue techniques used back then are almost impossible to replicate in the digital domain such as the tape phase effect on The Small Faces - Itchycoo Park that if you wear headphones (Sigh-I know) will contract and expand to the point you think your head is getting crushed. Never heard that effect work to that extent on a digitally sourced recording to date. While I agree modern recordings are much cleaner all round with better dynamic range, the actual useful amount of loudness is around the threshold of human hearing (About 96db). As others here have also explained there are problems associated with going loud. Even momentarily as could be needed for classical music. In the 70s we were more concerned about crosstalk from one magnetic track being picked up by its neighbouring track than overall dynamic range which was beyond the engineers control. (There were other problems as well such as the monitors were not much better than the tape machines and few studios had mics as good as the a Neumann U87). Transferring that to an LP was more about groove depth vs time duration vs bass depth (frequency) to stop styluses from bouncing out of the groove. Hence the introduction of the 12" single where both bass and volume could be increased. Looking at the site I can see one other fundamental flaw and that is placing their meter into a daw using at least a 32 bit floating point audio engine and placing a recording which at most will be 24 bit and measuring the dynamic output. To say I'm sceptical about the results on that site would be an understatement.
Average dynamic range is different than the total (or absolute) dynamic range of a program. While "Dark Side" may have a DR of 14, that's the average "peak to trough" range.... but the total range of the recording is measured from the noise floor of the silence to the absolute peak, which may be around 70 dB. But don't go thinking you only need 12 bits for 70dB of dynamic range. Do you really want your noise floor and quiet parts to be encoded by only 2-3 bits of resolution? That sounds pretty grainy.
Rock / Pop has always been compressed due to stage amplifier/speakers limitations, genre dependant, from Peak-to-Average of 20dB for Queen to 8dB for Metallica. The typical P2A for [mastered] acoustic recordings is ~25dB. Live [acoustic] recordings often have P2A around 40dB and beyond (Verdi operas), with an average of around ~80dBSPL. To record and reproduce them faithfully, we need 110+dB of dynamic range and we are nowhere near that point. Rare top-notch loudspeakers have distortions > 0.1%(60dB) on 90dBSPL, much worse on > 110 dBSPL. As of today, only a few premium headphones are capable of 80dB (0.01%) S2D, and people rave about them.
Does increased bit depth lead to an overall wider dynamic range or finer gradations within a more or less fixed overall dynamic range? One of Paul's early comments seemed to imply the latter. I can see the value of finer gradations that should lead to more accurate coding of the signal. A wider overall dynamic range does me little good. There is a limit to how loud I can COMFORTABLY listen to music (mainly classical). If the overall dynamic range is too great I have to turn up the volume on the quieter passages to hear them well enough to enjoy them. Judiciously applied compression has its place.
I agree with the compression comment, it indeed can kill the dynamic life....Side chain compression is however very valuable and has its place in high fidelity sound engineering
There is something that bothers me when I hear discussions about amount of bits and dynamic. I've always thought that if we know that, for example, the output voltage of a standard CD player is usually between 0 and 2V, that means that whatever digital word we choose for the sampling amplitude, the maximum (ie. all 1s) will always be 2V. So, the dynamic should always be the same. For example, with 16 bits, the maximum possible value is "65535", which is then mapped to 2V output. With 24 bits, the maximum value is now "16777215", which I assume is again mapped to 2V output (if it would be proportional, it would be 256V which I don't think is the case). And so on with 32 bits, etc... To me it seems that what varies with the number of bits is not the dynamic range, but the amount of levels/nuances that the player can output within the same dynamic range, which is always between 0 and 2V. The lower and the higher possible levels of a signal are always the same regardless of the amount of bits. What am I missing here?
@Douglas Blake thanks for the explanation! Now it is more clear. Am I right then if I say that the same volume level when put into a bigger digital quantization word is placed in an less coarser area in terms of available step levels? If I remember well, the lower is the digital number, the coarser are the number of steps (thus definition) available. That's the reason why usually in digital photography there is the concept of "exposure to the right", to make maximum usage of the finer definition present in the hight numbers.
I think you are missing one part: resolution. Those 50dB of dynamic range must be represented by a number, so if we were using only one bit at reproduction only square waves will come out. Like the pixels of a digital image.
A long time ago, over 50 years, another {at the time!} young sailor described how digital works with music. His father was doing very well in the computing industry. At, say. 44.1 kHz, the digital converter, looks at the analog signal, say, a sine wave, samples, `Looks,' at the signal position / values, & says, "Ah Ha! The value of the signal is, HERE!" And so, it is! THEN, the Digital sampling device, GOES STRAIGHT ACROSS, TO THE RIGHT, for a fraction of a second, & then looks for the next value. If the sine wave value is rising, the digital reader value goes straight up, to the new value, & says, "O. K., now we're HERE!" This sounds great, but what we have in the digital conversion, IS A SERIES OF SQUARE WAVES! These last can be, `Rounded off, smoothed off,' with filters, chips - but this last provides a better - sounding approximation! And consider: While 44.1kHz may sound fast, if you have good hearing, & can hear, say, 20,000 cycles, a cymbal crash, other high frequency instruments, electronic instruments, 20 k Hz is, `Only,' being sampled, slightly over twice, two times, during its complete cycle! You hear some highs, but the resulting signal sounds, mushy, indistinct. Studio reel - to - reel, while having its own distortion issues, at 15 to 30 ips, can sound much more realistic, precise, & onto a record, especially with a fine turntable, tone arm, cartridge. At the very least, where Paul's DSD 256, at 11.2 Mega Cycles, helps a lot, is a much higher Sampling Rate, & improved over PCM {at, perhaps, `Only,' 352 k Hz!} Also, inertia in speaker components, I feel, also helps to, `Round - off,' the, any, `Square Waves.' Listen, & see what you think! `One - Bit Technology, seems to help ~
Thats right we don't need that amount of compression anymore. Movie and game industries have been increasing their content size for over a decade now to deliver higher quality visuals. Music industry seem to be still lacking...
Data compression and dynamic compression are two unrelated things. When you say “content size” you’re talking about data compression. Paul’s talking about dynamic compression, the amount of difference between soft and loud.
Every record and the vast majority of CD's, SACD's and streamed music have dynamic range of 80db or less. More like 60db for LP's due to compression required to "cut" the vinyl without distortion.
@Douglas Blake The software used to measure those numbers is the DROffline MkII Analyzer and it’s giving a number of ITU and DRi metrics aimed at music production and actual dBs between the tiniest details in music vs the peaks can still be high numbers like 80+dB even those average numbers look horribly low.
@Douglas Blake But how exactly do you define dynamic range? For how long time do you analyze it by which math? Of course music is not compressed such that every harmonic is aligned to fit in the same envelope or it would completely ruin everything. You can still have >80dB of actual deltas between the peaks and finest harmonics even with those few dB of DRi values.
@Douglas Blake Of course I’m intimately aware of this “loudness war” thing and it annoys me when music is manipulated to reach a certain state of compression with the intent of maximizing the Spotify revenue of the track playing on cheap earbuds.
Would like to know where you got your specs for the "Headroom". Amplifiers are measured in power and Signal to Noise ratio. If maximum power is 500 Watts and you play it at 125 Watts, you have 6DB of headroom. If you play it at 250 Watts, you have 3DB of headroom. Using commercial PA, the sound board typically is run with about 17 DB of headroom. Nominal output is +4 dBu. Maximum is at +21dBu. Headroom is the space between the nominal level and clipping. The headroom is fully controlled by the volume levels set by the operator. Professionally, I prefer many of the QSC amplifiers, some Peavy, some Mackie, Myer Sound, and other professional gear. When I buy an amp, I'm interested in the maximum P-P voltage it will output into it's designated load. From there I can set the headroom to protect the amp and speakers. Yes this often involves compression and limiters, to limit power within the rage of the equipment.
North Holland and South Holland are 2 of the 12 provinces of the Netherlands, it is getting rarer and rarer to refer to the country as Holland these days, but Paul is a rarity. Now to rewatch that video and audionerd instead of linguanerding...
as i recall (and i'm just a roadie, not a sound engineer) the widespread over-compression of most music goes back to the 80s at least, and has to do with volume wars in radio and tv land..... i wasn't paying much attention to media then, but i sure noticed that records were getting more and more compressed... my solution was to get a dbx 3bxIII, with the "impact restoration"..... guess i've been using it 30-35 yrs now, usually moderately, and you'd have a tough time getting it away from me... think "cold, dead hands".. several much more eminent sound guys than i have, in fact, called it"integral" to a serious system, welll, one that allows for control (eq,, xover, range control).. so yeah, i just deal with it on the analog side..... but, having said that, "roadie attitude" isn't everyone's cup of tea
Small muscles in the ear are compressing the sound too (Tensor tympani and Stapedius muscle) A test: If you listen to a short powerful drum solo, it will sound loudest the first time you hear it. When you know what to expect it will not sound as loud anymore. (volume set at the same level)
I really don't want my ears to ring when someone shoots a pistol in a movie and I still want to be able to hear a conversation. Dynamic compression done right isn't a bad thing and most AV receivers have added Dynamic compression in the form of night mode or similar. In music would need to be able to turn down the horns etc when they approach 100 db and turn up quiet passages to increase space from the noise floor. In a quiet dedicated listening room 80 db would be about the limit to anything you would want to try and would approximate I believe a good seat at a live performance. Compression isn't bad but should be limited.
In pro audio . Where amps have seriius voltage abiluty ( wattage into any load..dynamics become your enimy.. who has played a movie and the loyd parrs are too loud if you turn up rnough to hear the quiet parts..?
Oh my god people. Put the music on and jam. This ain't the FCKn starship Enterprise. The more I get into audio the crazier it gets. Guess I'm just a simple man with my Carver gear and old kappa speakers but damn it's music not FCKn rockets. Hahahahaha
I think that there is the aspect of dynamic range, where you can hear the quietest sounds along with the loudest sounds (ie. the quiet sounds don't get lost). Some might look at the range as being the difference between the loudest sound and the noise floor. e.g. Digital can have a very low noise floor. Vinyl has a much higher noise floor, so may not have as large a dynamic range. However, digital is sampled both in the amplitude and time domains, whereas vinyl is continuous, which means that between the max and min amplitudes, you can have more resolution with vinyl. One hears of talk about great dynamic range with digital, but if the smearing from the rest of the system hides all the details, the effective dynamic range of the system may be quite low. ie. The dynamic range bottleneck is probably in the speakers.
One can look at a digital audio signal as a map of cone displacement in the speaker (as a percentage of the maximum possible.) Louder passages require greater cone excursions. The inertia of the voice coil and cone limit frequency response which is why we need speakers of different sizes to cover the complete audio frequency range at comparable listening volume.. You are right that speakers are the limiting factor.
@@machintelligence Yeh. I have a pair of Legacy Audio Valors. They use 14 inch drivers for (1) lower midrange and (2) woofers, but 12 inch drivers for the (2) subwoofers. The philosophy was the lower the frequency, the more surface area. What the designer also wanted was the fastest drivers, in order to be able to follow the input signal more closely.
I would love Hans to hop across the North Sea and come and listen to Stravinsky's Firebird in my music room. Visitors to my house have actually jumped out of their seats when the infernal dance of Kashchey's subjects starts up. It moves from almost total silence to a resounding crash within the space of a few seconds and, although I have no idea what the dynamic range is, it must be greater than the 15dB maximum that Hans mentioned in his question. Of course decibels is a logarithmic rather than a linear unit so it is quite dificult to present a meaningful illustration but if breathing is measured as 10dB (equivalent to the quiet, almost inaudible moments of the music) and let's assume the crash is around 110db then I make that a difference of around 100dB. I certainly would not want my listening to be constrained by a hifi system that was able to deliver anything less than this.
@Douglas Blake Thank you for your explanatiion. I have no way of measuring the volume produced by my speakers so you are probably right and the dynamic range is far less than I estimated. My estimates were based on some examples I found - the sound of someone breathing was rated at 10dB and 110dB was an estimate of the sound produced by a bass drum. However, the original question was asking if we needed a powerful amplifier in order to produce the sound we expect and I think the answer is yes, but it really depends what type of music you listen to. My daughter listens to pop and I have noticed that the music plays at pretty much the same volume throughout. I listen mostly to classical and the opposite is true - there are some very loud moments followed by quiet, almost inaudible passages. The hifi system I choose must be able to reproduce this - retaining all the clarity and detail in the quiet parts but still having plenty in reserve for the louder passages.
The last time I played a "high dynamic" range CD I blew a speaker cone. I should have never played a CD loud in the first place. Music should only be analog from start to finish.
I´ve been listening to cd´s for over 20 years. And i like it loud. I mean, really loud. Never had that problem. If that would happen to me ever, my conclusion would be "what a crappy speakers, i need to get better ones".
16bit of depth is good enough..when used wisely.. btw all Kpop girl groups' dance songs have nearly no dynamic range by maximized out the gain.... I think the gain maxing out is kinda core technology in the mix industry.
Bob would like to remind you that origami folding is not compression.. in fact it's better than the original! You just have to believe! I have this weird feeling that at some point meridian audio is going to join forces with the Church of Scientology don't ask me why...
Dynamic compression I believe is for people who listen on crummy little radio sets that obviously doesn't have any ability to cope with loud and soft and of course people who use headphones who hear everything else that little radio can't produce. Think I've got this right...not sure
i dont like too much compression but not too much dinamic range either. i hardly hear anything and then booooomm. some operas and classical pieces are impossible.
I get very annoyed with compression artifacts. It sounds like a warble to me. I've had people sitting next to me tell me they can't hear it but it drives me nuts.
It's funny ... although it took me a long time to get to fully understand how the amplitude ( hence dynamic range) is represented by a series of 0's and 1's ..... no one on here has successful described just how this is achieved ... plenty has been mentioned regarding the methods ( PCM PWM and DSD ) but never the audio level ... bit depth is only part of it as it's only part of sampling ...
@Douglas Blake ... Thank you 😊.... you've done a great job of explaining PCM 101 ..! I misled you .. sorry .. I fully understand all modes of digital recording.... I was only saying that Paul has often talked about the pros and cons of the various formats but he's never delved into how amplitude ( Dynamic headroom ) is handled ... Your explanation of PCM was more or less spot on .... I hope others who read this are enlightened now ... thank you 😃
Hey Paul, interestingly the European Union has a law against compression in place for a few years now. It is basically for TV and how you have to broadcast stuff, but it is a start.
The EU has a law that defines the curvature of a banana. That corrupt institution is about the worse example you can use to convince somebody that something is a good idea.
@@hugobloemers4425 that's true and I don't want to glorify the EU, but this is a good law. The EU has thousands of laws in place, some are bad and only reflect the power of lobbyists, but some are pretty good, like in every nation on the planet.
@@geddylee501 well... the EU calls music "noise" and the law is about noise emissions and not about increasing music quality, but still, it does the right thing ;-)
@@endrizo People are funny about audio reproduction. This and not that, that and not this..etc etc. LOL. I just read in the Google news yesterday Paul has a new mono amp out. When I read the pricing for two of those mono amps.😳
I cannot stand listening to highly compressed music. It works in the car with high background noise. Despite having a great system in my car, I find I'd rather listen to the sound track of the car rowing through the gears. In the future when only silent cars are around, I'll let it drive itself and I'll put on my headphones and listen to dynamic music. But for now, music is for quiet times in a dedicated audio room and cars are for driving.
@@GTRxMan what do you consider high dynamic range? How loud do you listen? Say that you listen at 85 db, which is fairly loud, and a typical listening room has around 55 db background noise. Even if the crescendo of the piece is 18 db louder, which would be a whole lot, we're still well within the range of even a modest lp playback system. Tldr version you don't listen loud enough for a 120 db dynamic range. Let alone 96...
Do speakers grills/dust covers add to a speakers life span? I know most guys want to show off but is it worth it? My quess is no In my opinionleave the grills on. In the 1970s speaker grills were not remarkable and yet the Zenith or Sony Speakers sounded excellent. Then someone like me curious tore the speakers apart only to find nothing impressive, The speaker cones were ugly to say the least. No Foarm or Rubber Surrounds around the drivers, just tested paper. Yet if one finds one of these 40 0r 50 year old speakers in a thifft store they work 100%. Find a same age speaker with foam or rubber and it's garbage.
I imagine there is some sonic benefit to having the cone made of rigid material and the surround of a different much-softer material, so that the cone can move without changing its shape. The problem with the foam or rubber surrounds is that they were (at least at one time) made of a short-lifetime material. I still (naïvely?) hold out hope that a surround material could be found that has a reasonable service life.
@@kc9scott🤗 100 % AGREE that was the case in the past … which is why I switched to B&W speakers and I haven’t had a problem since 🤗 And IMHHO and many of my friends and family agree …they sound better than many popular new speakers 🤗💚💚💚
"We are controlling transmission. If we wish to make it louder, we will bring up the volume. If we wish to make it softer, we will tune it to a whisper."
-- from The Outer Limits
One of the greatest advantages of a high bit depth (eg, 24bit; 32bit, 64bit etc.) is the for any post-processing, the rounding errors become insignificant. Even digital faders become totally transparent when you use very high bit depths. In contrast, with a 16 bit fader you would easily hear the output going grainy as you lower it. With a 24bit, 32bit or 64bit fader any such grainyness disappears.
Likewise, with AV processors, when you an apply EQ etc, it works much better when the input source music is using a very high bit depth.
Technically, the name Holland refers to just two of the 12 provinces that comprise The Netherlands, but people often use it colloquially to refer to the entire country. About a third of the population of the Netherlands lives in either North or South Holland.
True, but most Dutch should not be too bothered if people use Holland to refer to the entire country.
Next, you will tell us some Hollanders don't wear wooden shoes.
@@BTom16 🤣
@gotham61 thanks! I have a goal to learn something new everyday and that's taught me something I didn't know and will hopefully remember!
So who are the Dutch?
From what I have observed over the years, many people forget about transients and peaks and think more about average level. When I compare the two, I’m not talking about seconds verses minutes, I’m talking about microseconds verses milliseconds. Whether it is your amplifier that can’t reproduce it or the recording chain that has compressed or limited those transients and peaks, once they are gone, they are gone. While a violin, may not have massive peaks, a drum kit sure does. To get the true nature and nuance of an instrument, you need those transients. That requires dynamic range. In most cases the ratio of average to peak is of several magnitude.
In a professional recording studio (proper acoustics and sound isolation from the outside world) using world-class mics, properly set to record a drum kit, I can guarantee you will use every bit of dynamic range the mic pres and recording chain can muster to faithfully record what the player may be playing on that kit. During “tracking” of all instruments, getting a faithful recording of all nuances is important. You need dynamic range (among other things to do that). Keep in mind , when setting levels, you should be thinking about average max levels well below zero, or there will be transients well over zero. Keep the levels on the low side.
During the mixing stage (depending on the music genre) compression will most likely be used on a track by track basis and again maybe on groups or stems. Hopefully, some of the transients and peaks are almost intact. If you think about the difference between a jazz quartet, a rock band and modern dance music, you get the idea. We still need dynamic range, but maybe not as much, as decisions have already been made.
After mixing, the project will many times head to mastering. More eq, more compression and more peak limiting. After this stage it may be streamed, released for download, converted to MP3, burned to CD or possibly even pressed to vinyl. Hopefully they did multiple master sessions.
As a side note, most broadcast FM radio stations have spent anywhere from $6000 - $20,000 for audio signal processors that do nothing more than alter the eq and dynamic range of their audio to “fit” their program material, target audience and make every song “sound the same” over the air. They have a government regulation that says they cannot exceed 100% modulation, peaks are their enemy. They limit or clip the peaks then compress everything else to raise the average level. This is done in a very sophisticated way where the audio is split, frequency- wise into 6-10 bands and each band receives different treatment (programable).
Restoring dynamic range after it has already had the life squeezed out of it, is almost impossible (without coloring the sound).
Having worked in live sound for many years, I have “Pro tools“ sessions that were recorded live that range from 20 something tracks up to about 60 tracks (per song) that have full dynamic range that I can mix, pan, eq and process as I see fit and listen to in my listening room. You have no idea how much I wish I had that for every artist I was interested in. Maybe someday we all could have that.
Of course, If you were trying to listen to that late at night, in an apartment or with your kids trying to sleep in the next room……..you would be right back to FM radio……over headphones……..with a dynamic range of 6db.
Seriously, don’t think about perceived loudness (average level) think about the nuances and peaks that you may not think you hear, but make a huge difference.
As a lover of classical music, we definitely need the abiility to portray a wide dynamic range. I was listening to a recent release of the Shostakovich 5th symphony and the magnificent slow movement Largo amply demonstrates such a need, with string tremolos barely audible one moment and triple forte outbursts the next.
I’m really glad Paul has been doing this channel. While there are times I don’t agree with him, I appreciate his passion and real world experience. It’s very rare that the top brass of a company is willing to get in front of the camera. Really proud of this guy… even bought his weight loss book.
From another Hans from the Netherlands: I remember having once sold a live record by Herman van Veen from the seventies because of the dynamic range being too big... I considered it unplayable in the living room with other people present. Most of the time one could not hear the soft parts and when the volume had been turned up sufficiently, people got scared by the loud parts
Yes, with music there’s a certain amount of dynamic range that sounds either “natural” (without obvious compression) or “kind of large”. If you go much beyond that, it just becomes bad music.
Thanks for this one Paul. Great break down of the process. Love your videos and watch often. Greatly appreciated. LLAP
An extremely quite listening room has a background noise level of at least 30db. Usual noise level is 40db plus. Listening even at the dynamic range a CD theoretically supports and still being able to hear the quietest of passages would blow your ear drums during the loud ones. That's all Paul should have answered instead of circumvavigating the question. You can't reasonably have music DELIVERED with a DR of more than 30-35db and even that would be oftentimes too much even for in house listening. How you record, mix and master is another matter altogether, but in the end you have to limit the DR. Just not too much. 10-15db of DR in the music content is enough.
My hat off to you Sir , exactly right !
If you hadn't noticed this was a gateway to a shameless plug for his new "dynamics" recordings.
@@JonAnderhub I am sure the recordings delivered by Octave Records adhere to the above principles, no matter what Paul says to his audience. I would be extremely surprised if there is more than 20db DR in them, even 15db would be a welcome surprise. There's no need for more.
To add to this: using the full 16 bit DR would result in at least 126dB in a quiet listening room…. Sorry neighbours (and ears!).
@@rw80 I actually made that same calculation a few years ago and came to the same conclusion . 24 bit , whatsyagonnauseitfor ?
This is a poor explanation that conflates bit depth used in the recording process with bit depth in the distribution format. High bit depth is used in the recording process to give headroom for editing. 16 bits gives a dynamic range of 96dB and 24 bits gives 144dB. The decibel is a logarithmic scale, +10 dB is a multiplier of 10. The quietest studio microphone has a s/n ratio of about 92dB due to the Browning Effect, the random movement of air molecules hitting the diaphragm. A full symphony orchestra has a dynamic range around 60dB (10 bits). No home audio equipment can handle 144dB.
Ignoring all the above let us imagine you have obtained a true 24 bit track that has captured a dynamic range of 144dB. A quiet listening room may have 50dB background noise. You play the track and set volume so the quietest sound can just be heard above background. Then there is a sound at 144dB. If this does not instantly destroy your speakers then your hearing will be destroyed.50dB+144dB=194dB which is in the lethal range (not joking). It is fortunate 144dB dynamic range is impossible for any home system. Even a track with 96dB dynamics can potentially cause ear pain and damage. Increasing bit depth has no effect on audio fidelity other than lowering the noise floor which is already silent with 16 bits.Most music recordings require a sensible level of dynamic compression to be playable on a home system.
Sometimes there can be overkill. A perfect example is Christoper Herrick's "Organ Dreams 2". Some soft notes you can almost not hear and suddenly the loud notes blow the roof off. I could be a nightmare for people living in apartment buildings.
Instead throw out the money for useless gear and speakers invest it in the room
Love the channel I usually get a cup of coffee and start my day with one of your videos.
As a former recording engineer if you want to blow listeners minds, Recording and Close-miking a rock or jazz band without any compression at all is amazingly dynamic from the softest passage to an explosive drum solo. Most people have never heard anything like it and it will sound alien to some people.
Tom Jung from DMP used to do some of this with his recordings in the 80’s/90’s that were recorded on the Mitsubishi X80 digital recorders but even he used some compression. The Flim & the BB’s albums are some examples of this.
There seems to be some confusion on the term Dynamic Range. There are two basic camps, one that considers dynamic range to be the difference between Peak and Average levels. In this case, many rock recordings are in the 6 to 10dB range, 10dB being relatively great. The other camp is that of dynamic range being the difference between maximum and minimum recorded levels. This can be very large, but still is only about 8 bits worth (48dB). Aside from those two groups, something that gets lost is resolution. For resolution, I consider 16 bits to be adequate, but just barely. Personally, I think that 20 bits resolution is good and provides some headroom for most recording streams. Of course, in our computing world, this rounds up to 24 bits that can readily be manipulated by our systems. I don't think that there is ANY piece of equipment in our hobby that can resolve the full 144dB range of 24 bits, so 24 is a bit of overkill, but reasonable.
I have not seen any equipment with over 130 db dr especially from start to finish.
To appreciate dynamic range you need a quiet listening environment. One of my pet peeves about car stereo is that inside a vehicle at 70 mph is anything but a quiet place. When I mention this to car stereo salesmen the invariable answer is to "turn up the volume." But what about the quiet parts? Perhaps this doesn't matter to the average driver.
That is why so many people use Kilmat everywhere they can.
Dynamic range in a recording can itself be overkill at times, believe it or not. If the listener has to adjust the volume mid-track because something is too quiet or too loud, it ruins overall immersion. Simply more difference between the quietest and loudest parts is not always better. There's a golden sweet spot between the extremes of dynamic range and compression.
Having asked this question that Paul addressed, I have read many of the comments below.
It's been very helpful for me and thanks for that !
I referred to DSOTM for that's an album many of us know. I have it on CD's and DSD. According to Roon and Jriver the most dynamic track on that album is the track " the great gig in the sky " and it's about 19dB on both CD and DSD ( R128) . I'm not really interested in numbers but it just raised the question why botter to have 100 dB or more if we only have 15 dB or less recordings ?
What I read from most of you, is that indeed recordings are often below 20 dB Dynamic range and hence we don't need more for it will be just to much for us humans to absorb.
Is dynamic range than nonsense ? No it's not. more bits or DSD will provides you much more granularity ( provides space, generates more nuances, more details ) and these little details will pull you more into the music. If I'm taking a wrong turn here, don't hesitate to correct me 🙂
@Douglas Blake the second plane is something I started to think about to see if I can make the correlation between Dynamic range and distortion. Is it fair to say that when you have an amplifier with a high dynamic range the likeliness that when playing at medium volume the distortion of that amp is so low that it will not influence the added little nuances you will get when playing high res music? (High bit rate ) An amplifier with less dynamic range will likely have a higher distortion level at that same volume and therefore these nuances are lost ? Meaning if you want to have the benefit of listening to the little details you just need to have a amp with a high dynamic range ?
This is why I love this channel & I like Paul the most! Thus comments below also help us educate, think & focus on the topic.
I really like high dynamic range so the tracks Paul has in mind can be lots of fun. However, if you want to listen to it, you need to play it always loud otherwise you can not hear the low passages, and that is not always possible. I have a CD with the Bolero on it. Living in an apartment I can not play it start to end at one volume, so I hear the music at the start and not annoy the neighbors at the end of the CD.
Bolero is one of the best demo tracks in existence, but few audiophiles will sit through a 15 minute piece.
I think in rock/pop,etc., there has been a move to make everything sound near field, particularly vocals where for some reason they want you to hear the vocalist's tongue leaving the roof of their mouth.
dynamic range is important, and more than we need of it is great for master recordings and the future.
however it can be pretty brutal for casual listening to music, depending on the system and person some recordings just are to intense and kick like a mule :)
Definitions vary, but "dynamic range" of a track generally means something like the difference in dB between the volume of the loudest passage and the volume of the softest passage in that track. But by "volume" we mean the sum of all the sounds going on at that moment. Our ears can totally hear a faint violin solo thats many dB below the noisefloor. The woodwinds may be tens of dB below the brass in the same passage but be very audible. Overtones that tell us timbre may be dozens of dB below the fundamental. The tiny cues we use to perceive the room/space can be _scores_ of dB down from the main signal. Automated "dynamic range" measurements miss all of these things. But you do need the dynamics to actually reproduce them, and so you do need the bit depth to store them. 24-bit might still be overkill (as it exceeds the pain threshold of the human ear), but as other commenters have mentioned it helps with things like rounding errors.
As the inquirer points out, a dynamic range of 15 dB is a common score for a well-mastered song. If that were truly the full story on dynamics, you would need only 1.5/log10(2) = 5 bits to reproduce it - 8 bits would be overkill. But search for 8-bit music - we millennials grew up listening to it in video games - and it doesn't take an audiophile to notice it is limited.
But what you not realize is that in the digital domain 16 is not just double 8
THANKS PAUL 🤗 WE AWAIT THE FUTURE 👍😎💚💚💚
Well said Paul. Dynamic range gives music emotion and passion! Personally, I'm not interested in listening to a loud transistor radio from 1973.
I used to own a DBX 3BX dynamic range expander. Yes I could crank up the dynamic range and it made the music sound artificial but dynamic. When used reasonably it would lower the noise floor of my cassettes and VCR tapes. You have to remember that not only does dynamic range affect the louder portions of music but can also lower the softer sounds like noise or certain instruments making it harder to hear the more subtle sounds. Probably the reason why I like vinyl.
While I do agree that modern music is unnecessarily dynamic compressed, I do not think there is anything lacking with the dynamic range currently offered by CDs.
At 16bit CDs offer excellent dynamics. It's not the fault of the pcm format that the dynamics are not utilised properly and that the music is compressed in the mastering.
The best example for me, of an album with great dynamics is the Moody Blues' "Days Of Future Passed"...............the opening gong crescendo is amazing, as is the rest of the recording! Not bad for 1969!
Paul, The number of bits does not increase the amount the sound can go loud as any 8 bit, 12 bit, 16 bit, or 24 bit signal can be amplified to 0DB FS. (DB Full Scale) What changes is the resolution. It's like taking a battery voltage measurement with a 2 digit multimeter and seeing 12 volts, but a 4 digit multi-meter on the same scale can read up to 99 volts, one reads 12 volts, the other reads 12.57 Volts, a much more precise number.
Bitrate on the other hand, limits the maximum frequency that is able to be recorded without aliasing. For example, CD quality can have a nyquist frequency up to 22 KHZ, but that would result in generation of many beat frequencies in much lower frequencies, so a low pass filter is used to cut the high frequencies that create aliasing.
Experiment for the audiophiles. Use Audacity to make a pure digital recording of the following frequencies at -18DB. 18 KHZ, 19 KHZ, 19.5 KHZ, and 19.8 KHZ. Play it back on your favorite high quality DAC and sound system. Use a spectrum anlyzer to see the pure sine wave played back while reproduced at the 44.1 KHZ stereo sample rate. Now you know the reason Pink Floyd sounds so different between CD and an LP on the Dark Side of the Moon album. The coins in Money and the clocks in Time don't have a chance to sound correct on a CD.
Most of the aliasing would be removed if recorded at 96KHZ with a 40KHZ Low Pass Filter.
Paul, if you have the time, post a video with digitally generated test tones sampled at CD quality. Not sure what CZcams would do to the recording in addition to the original aliasing. A perfect DAC can't fix this problem as the aliasing is well into the audiable range of a few KHZ. This can only be cured by filtering out the high frequencies in the recording, so they are not recorded to produce aliasing products.
Thanks Sir for the info
Love those chairs!
All music to some degree goes through compression. modern music much more then classical. personally too much dynamic range makes me crank up the volume for the small subtleties and then when a big ensemble hits it knocks me off the couch. I can't under my normal listening deal with that much dynamic range. some compression is needed so you can hear the subtleties.
Their is some compression on Dark Side of the Moon, as the TG console series 4 (16 track) had it's own compressors built into the board per channel
Hello Paul. Just a quick feedback message to say that I genuinely enjoy these videos that you generate. I realize that you spend quite a bit of time creating these videos and it appears that you are honestly dedicated to what you are doing here. I want to tell you in clear terms that I appreciate your efforts and your dedication towards creating said videos, and I really enjoy listening to you, as you work your way through the various topics which you speak on. While I do not agree with every, single, one of your positions (what a boring world it would be if we were all exactly the same), I do very much respect you and your efforts in putting yourself out here, in CZcams land, spending a portion of your valuable time with us through these videos. You are appreciated, no matter what, and thank you very much for sharing some of your passion of audio with us! I genuinely look forward to your next educational video Paul! Thank You! Fred
That database seems reliant on voluntary contributions so not a good start in my opinion.
Dark Side of the Moon, recorded at Abby Road Studio 3 would have used 30ips TBR 2 tape machines with a (for that time) clean 40db of dynamic range. Later it was digitally remastered to PCM. Old masters can be tweaked quite well with a competent engineer.
There is also a large difference in the dynamic range of an LP depending on the equipment used for playback. Some analogue techniques used back then are almost impossible to replicate in the digital domain such as the tape phase effect on The Small Faces - Itchycoo Park that if you wear headphones (Sigh-I know) will contract and expand to the point you think your head is getting crushed. Never heard that effect work to that extent on a digitally sourced recording to date.
While I agree modern recordings are much cleaner all round with better dynamic range, the actual useful amount of loudness is around the threshold of human hearing (About 96db). As others here have also explained there are problems associated with going loud. Even momentarily as could be needed for classical music.
In the 70s we were more concerned about crosstalk from one magnetic track being picked up by its neighbouring track than overall dynamic range which was beyond the engineers control.
(There were other problems as well such as the monitors were not much better than the tape machines and few studios had mics as good as the a Neumann U87).
Transferring that to an LP was more about groove depth vs time duration vs bass depth (frequency) to stop styluses from bouncing out of the groove. Hence the introduction of the 12" single where both bass and volume could be increased.
Looking at the site I can see one other fundamental flaw and that is placing their meter into a daw using at least a 32 bit floating point audio engine and placing a recording which at most will be 24 bit and measuring the dynamic output.
To say I'm sceptical about the results on that site would be an understatement.
Average dynamic range is different than the total (or absolute) dynamic range of a program. While "Dark Side" may have a DR of 14, that's the average "peak to trough" range.... but the total range of the recording is measured from the noise floor of the silence to the absolute peak, which may be around 70 dB. But don't go thinking you only need 12 bits for 70dB of dynamic range. Do you really want your noise floor and quiet parts to be encoded by only 2-3 bits of resolution? That sounds pretty grainy.
Rock / Pop has always been compressed due to stage amplifier/speakers limitations, genre dependant, from Peak-to-Average of 20dB for Queen to 8dB for Metallica. The typical P2A for [mastered] acoustic recordings is ~25dB. Live [acoustic] recordings often have P2A around 40dB and beyond (Verdi operas), with an average of around ~80dBSPL. To record and reproduce them faithfully, we need 110+dB of dynamic range and we are nowhere near that point. Rare top-notch loudspeakers have distortions > 0.1%(60dB) on 90dBSPL, much worse on > 110 dBSPL. As of today, only a few premium headphones are capable of 80dB (0.01%) S2D, and people rave about them.
Does increased bit depth lead to an overall wider dynamic range or finer gradations within a more or less fixed overall dynamic range? One of Paul's early comments seemed to imply the latter. I can see the value of finer gradations that should lead to more accurate coding of the signal. A wider overall dynamic range does me little good. There is a limit to how loud I can COMFORTABLY listen to music (mainly classical). If the overall dynamic range is too great I have to turn up the volume on the quieter passages to hear them well enough to enjoy them. Judiciously applied compression has its place.
I agree with the compression comment, it indeed can kill the dynamic life....Side chain compression is however very valuable and has its place in high fidelity sound engineering
Compression Law? 🤔 I still struggle with Ohm's Law but I do like the idea.💡👍
There should be a law against compression!
and another against too much dinamic range
There should be a law against uneducated, clueless and foolish comments
There is something that bothers me when I hear discussions about amount of bits and dynamic. I've always thought that if we know that, for example, the output voltage of a standard CD player is usually between 0 and 2V, that means that whatever digital word we choose for the sampling amplitude, the maximum (ie. all 1s) will always be 2V. So, the dynamic should always be the same. For example, with 16 bits, the maximum possible value is "65535", which is then mapped to 2V output. With 24 bits, the maximum value is now "16777215", which I assume is again mapped to 2V output (if it would be proportional, it would be 256V which I don't think is the case). And so on with 32 bits, etc...
To me it seems that what varies with the number of bits is not the dynamic range, but the amount of levels/nuances that the player can output within the same dynamic range, which is always between 0 and 2V. The lower and the higher possible levels of a signal are always the same regardless of the amount of bits.
What am I missing here?
@Douglas Blake thanks for the explanation! Now it is more clear.
Am I right then if I say that the same volume level when put into a bigger digital quantization word is placed in an less coarser area in terms of available step levels?
If I remember well, the lower is the digital number, the coarser are the number of steps (thus definition) available.
That's the reason why usually in digital photography there is the concept of "exposure to the right", to make maximum usage of the finer definition present in the hight numbers.
I think you are missing one part: resolution. Those 50dB of dynamic range must be represented by a number, so if we were using only one bit at reproduction only square waves will come out. Like the pixels of a digital image.
A long time ago, over 50 years, another {at the time!} young sailor described how digital works with music. His father was doing very well in the computing industry. At, say. 44.1 kHz, the digital converter, looks at the analog signal, say, a sine wave, samples, `Looks,' at the signal position / values, & says, "Ah Ha! The value of the signal is, HERE!" And so, it is! THEN, the Digital sampling device, GOES STRAIGHT ACROSS, TO THE RIGHT, for a fraction of a second, & then looks for the next value. If the sine wave value is rising, the digital reader value goes straight up, to the new value, & says, "O. K., now we're HERE!" This sounds great, but what we have in the digital conversion, IS A SERIES OF SQUARE WAVES! These last can be, `Rounded off, smoothed off,' with filters, chips - but this last provides a better - sounding approximation! And consider: While 44.1kHz may sound fast, if you have good hearing, & can hear, say, 20,000 cycles, a cymbal crash, other high frequency instruments, electronic instruments, 20 k Hz is, `Only,' being sampled, slightly over twice, two times, during its complete cycle! You hear some highs, but the resulting signal sounds, mushy, indistinct. Studio reel - to - reel, while having its own distortion issues, at 15 to 30 ips, can sound much more realistic, precise, & onto a record, especially with a fine turntable, tone arm, cartridge. At the very least, where Paul's DSD 256, at 11.2 Mega Cycles, helps a lot, is a much higher Sampling Rate, & improved over PCM {at, perhaps, `Only,' 352 k Hz!} Also, inertia in speaker components, I feel, also helps to, `Round - off,' the, any, `Square Waves.' Listen, & see what you think! `One - Bit Technology, seems to help ~
I have heard 78 RPM records played properly with mono needles having tons of dynamic range.
Thats right we don't need that amount of compression anymore. Movie and game industries have been increasing their content size for over a decade now to deliver higher quality visuals. Music industry seem to be still lacking...
Data compression and dynamic compression are two unrelated things. When you say “content size” you’re talking about data compression. Paul’s talking about dynamic compression, the amount of difference between soft and loud.
You don't know what compression means in different contexts
You’re right, Paul.
Every record and the vast majority of CD's, SACD's and streamed music have dynamic range of 80db or less. More like 60db for LP's due to compression required to "cut" the vinyl without distortion.
@Douglas Blake The software used to measure those numbers is the DROffline MkII Analyzer and it’s giving a number of ITU and DRi metrics aimed at music production and actual dBs between the tiniest details in music vs the peaks can still be high numbers like 80+dB even those average numbers look horribly low.
@Douglas Blake But how exactly do you define dynamic range? For how long time do you analyze it by which math? Of course music is not compressed such that every harmonic is aligned to fit in the same envelope or it would completely ruin everything. You can still have >80dB of actual deltas between the peaks and finest harmonics even with those few dB of DRi values.
@Douglas Blake Of course I’m intimately aware of this “loudness war” thing and it annoys me when music is manipulated to reach a certain state of compression with the intent of maximizing the Spotify revenue of the track playing on cheap earbuds.
When it comes to dynamic range have a listen to a NAD amplifier. +3db headroom and huge dynamic power. Got a fair few nad units.
Would like to know where you got your specs for the "Headroom". Amplifiers are measured in power and Signal to Noise ratio. If maximum power is 500 Watts and you play it at 125 Watts, you have 6DB of headroom. If you play it at 250 Watts, you have 3DB of headroom.
Using commercial PA, the sound board typically is run with about 17 DB of headroom. Nominal output is +4 dBu. Maximum is at +21dBu. Headroom is the space between the nominal level and clipping. The headroom is fully controlled by the volume levels set by the operator.
Professionally, I prefer many of the QSC amplifiers, some Peavy, some Mackie, Myer Sound, and other professional gear.
When I buy an amp, I'm interested in the maximum P-P voltage it will output into it's designated load. From there I can set the headroom to protect the amp and speakers. Yes this often involves compression and limiters, to limit power within the rage of the equipment.
North Holland and South Holland are 2 of the 12 provinces of the Netherlands, it is getting rarer and rarer to refer to the country as Holland these days, but Paul is a rarity. Now to rewatch that video and audionerd instead of linguanerding...
There should be a law that music should be only available in LOSSLESS.
The days of Dial-Up are over and we now have Terrabyte hard drives.
Where are you enforced to lossy compression? Looking at round about 100k FLAC files here I don't get your problem
And lossless or not don't change anything relevant to the topic
@@Harald_Reindl
I’m not enforced but it’s still going around and should be eliminated.
@@Harald_Reindl
It doesn’t matter. I’m just giving an opinion which is a sub topic.
@@Bassotronics your nonsense isn't even a sub-topic because you are not smart enough to distinct between dynamic range and data compression
as i recall (and i'm just a roadie, not a sound engineer) the widespread over-compression of most music goes back to the 80s at least, and has to do with volume wars in radio and tv land..... i wasn't paying much attention to media then, but i sure noticed that records were getting more and more compressed... my solution was to get a dbx 3bxIII, with the "impact restoration"..... guess i've been using it 30-35 yrs now, usually moderately, and you'd have a tough time getting it away from me... think "cold, dead hands".. several much more eminent sound guys than i have, in fact, called it"integral" to a serious system, welll, one that allows for control (eq,, xover, range control).. so yeah, i just deal with it on the analog side..... but, having said that, "roadie attitude" isn't everyone's cup of tea
Small muscles in the ear are compressing the sound too (Tensor tympani and Stapedius muscle)
A test: If you listen to a short powerful drum solo, it will sound loudest the first time you hear it.
When you know what to expect it will not sound as loud anymore. (volume set at the same level)
Lol I hope that isn't most people
@Douglas Blake Wow probably a defense mechanism
@@asadhafeez7713 after about 80 db your ears quit hearing quiet sounds (low Db) and it takes a few minutes of quiet to hear low db sounds again.
@@joegreen4089 NEVER experienced this in my HiFi theater, but that sounds like adaptation to changes in sound.
@@asadhafeez7713 yup.
I agree compression should be illegal
Compression is bad, yes, but what do you mean by "illegal"?
I really don't want my ears to ring when someone shoots a pistol in a movie and I still want to be able to hear a conversation. Dynamic compression done right isn't a bad thing and most AV receivers have added Dynamic compression in the form of night mode or similar. In music would need to be able to turn down the horns etc when they approach 100 db and turn up quiet passages to increase space from the noise floor. In a quiet dedicated listening room 80 db would be about the limit to anything you would want to try and would approximate I believe a good seat at a live performance. Compression isn't bad but should be limited.
@@mindtraveller100 the overuse of it should be banned in my opinion
Dumbness should be illegal and still you speak
“…like a crack a lightnin’ “ 😀
I'd guess Dark Side could well be 60dB range since it was mixed for vinyl.
In pro audio . Where amps have seriius voltage abiluty ( wattage into any load..dynamics become your enimy.. who has played a movie and the loyd parrs are too loud if you turn up rnough to hear the quiet parts..?
Oh my god people. Put the music on and jam. This ain't the FCKn starship Enterprise. The more I get into audio the crazier it gets. Guess I'm just a simple man with my Carver gear and old kappa speakers but damn it's music not FCKn rockets. Hahahahaha
yeah but it's the speed of dynamic range you never going to get to the speed of sound its not just amplitude to noise floor
Hans from The Hans Beekhuyzen Channel?
i was thinking the same :-)
I know right? A Hans from a tiny country called The Netherlands, asking a technical question.. could be 'our' Hans!
when im on yt, i want my sounds very compressed. but when i build my system, i want high dynamic range.
I think that there is the aspect of dynamic range, where you can hear the quietest sounds along with the loudest sounds (ie. the quiet sounds don't get lost). Some might look at the range as being the difference between the loudest sound and the noise floor. e.g. Digital can have a very low noise floor. Vinyl has a much higher noise floor, so may not have as large a dynamic range. However, digital is sampled both in the amplitude and time domains, whereas vinyl is continuous, which means that between the max and min amplitudes, you can have more resolution with vinyl. One hears of talk about great dynamic range with digital, but if the smearing from the rest of the system hides all the details, the effective dynamic range of the system may be quite low. ie. The dynamic range bottleneck is probably in the speakers.
One can look at a digital audio signal as a map of cone displacement in the speaker (as a percentage of the maximum possible.) Louder passages require greater cone excursions. The inertia of the voice coil and cone limit frequency response which is why we need speakers of different sizes to cover the complete audio frequency range at comparable listening volume.. You are right that speakers are the limiting factor.
@@machintelligence Yeh. I have a pair of Legacy Audio Valors. They use 14 inch drivers for (1) lower midrange and (2) woofers, but 12 inch drivers for the (2) subwoofers. The philosophy was the lower the frequency, the more surface area. What the designer also wanted was the fastest drivers, in order to be able to follow the input signal more closely.
Hans wouldn't notice a nonmusical addition behind Dark Side standing at just 15db below the peak?
I would love Hans to hop across the North Sea and come and listen to Stravinsky's Firebird in my music room. Visitors to my house have actually jumped out of their seats when the infernal dance of Kashchey's subjects starts up. It moves from almost total silence to a resounding crash within the space of a few seconds and, although I have no idea what the dynamic range is, it must be greater than the 15dB maximum that Hans mentioned in his question. Of course decibels is a logarithmic rather than a linear unit so it is quite dificult to present a meaningful illustration but if breathing is measured as 10dB (equivalent to the quiet, almost inaudible moments of the music) and let's assume the crash is around 110db then I make that a difference of around 100dB. I certainly would not want my listening to be constrained by a hifi system that was able to deliver anything less than this.
@Douglas Blake Thank you for your explanatiion. I have no way of measuring the volume produced by my speakers so you are probably right and the dynamic range is far less than I estimated. My estimates were based on some examples I found - the sound of someone breathing was rated at 10dB and 110dB was an estimate of the sound produced by a bass drum. However, the original question was asking if we needed a powerful amplifier in order to produce the sound we expect and I think the answer is yes, but it really depends what type of music you listen to. My daughter listens to pop and I have noticed that the music plays at pretty much the same volume throughout. I listen mostly to classical and the opposite is true - there are some very loud moments followed by quiet, almost inaudible passages. The hifi system I choose must be able to reproduce this - retaining all the clarity and detail in the quiet parts but still having plenty in reserve for the louder passages.
The only genre that has a very satisfying dynamic range is Classical music. Some Jazz also but for the rest it’s all compressed.
The last time I played a "high dynamic" range CD I blew a speaker cone. I should have never played a CD loud in the first place. Music should only be analog from start to finish.
I´ve been listening to cd´s for over 20 years. And i like it loud. I mean, really loud. Never had that problem. If that would happen to me ever, my conclusion would be "what a crappy speakers, i need to get better ones".
@@mindtraveller100 I had bought Polk SDA's. They were never rock speakers anyway. But the were warranty repaired by Polk.
16bit of depth is good enough..when used wisely..
btw all Kpop girl groups' dance songs have nearly no dynamic range by maximized out the gain.... I think the gain maxing out is kinda core technology in the mix industry.
Bob would like to remind you that origami folding is not compression.. in fact it's better than the original! You just have to believe!
I have this weird feeling that at some point meridian audio is going to join forces with the Church of Scientology don't ask me why...
Dynamic compression I believe is for people who listen on crummy little radio sets that obviously doesn't have any ability to cope with loud and soft and of course people who use headphones who hear everything else that little radio can't produce. Think I've got this right...not sure
24/48 is more than enough... SQ depends on what is all done to the music before it is sent on to us the consumer
Play "Jurassic Lunch" and you'll learn about dynamic range.
i dont like too much compression but not too much dinamic range either.
i hardly hear anything and then booooomm. some operas and classical pieces are impossible.
I get very annoyed with compression artifacts. It sounds like a warble to me. I've had people sitting next to me tell me they can't hear it but it drives me nuts.
The Netherlands is a country on the North Sea bordering Germany and Belgium.
It's funny ... although it took me a long time to get to fully understand how the amplitude ( hence dynamic range) is represented by a series of 0's and 1's ..... no one on here has successful described just how this is achieved ... plenty has been mentioned regarding the methods ( PCM PWM and DSD ) but never the audio level ... bit depth is only part of it as it's only part of sampling ...
@Douglas Blake ... Thank you 😊.... you've done a great job of explaining PCM 101 ..! I misled you .. sorry .. I fully understand all modes of digital recording.... I was only saying that Paul has often talked about the pros and cons of the various formats but he's never delved into how amplitude ( Dynamic headroom ) is handled ...
Your explanation of PCM was more or less spot on .... I hope others who read this are enlightened now ... thank you 😃
Hey Paul, interestingly the European Union has a law against compression in place for a few years now. It is basically for TV and how you have to broadcast stuff, but it is a start.
The EU has a law that defines the curvature of a banana. That corrupt institution is about the worse example you can use to convince somebody that something is a good idea.
@@hugobloemers4425 that's true and I don't want to glorify the EU, but this is a good law. The EU has thousands of laws in place, some are bad and only reflect the power of lobbyists, but some are pretty good, like in every nation on the planet.
-23dB re FS sine wave?
@@geddylee501 well... the EU calls music "noise" and the law is about noise emissions and not about increasing music quality, but still, it does the right thing ;-)
If I remember correctly this was about commercials sounding much louder than the regular broadcast signal, as a deliberate result of compression
or there was no mouse with clogs on the stair
I want the best I can get... but my classic rock ears aren't that good. Just crank it to 11 and call it good.
Cover all your walls and ceiling with egg carton’s. That will solve your sound issues.
too dry walls is no good either. just treat the main reflection zones
@@endrizo
People are funny about audio reproduction. This and not that, that and not this..etc etc. LOL. I just read in the Google news yesterday Paul has a new mono amp out. When I read the pricing for two of those mono amps.😳
I cannot stand listening to highly compressed music. It works in the car with high background noise. Despite having a great system in my car, I find I'd rather listen to the sound track of the car rowing through the gears. In the future when only silent cars are around, I'll let it drive itself and I'll put on my headphones and listen to dynamic music. But for now, music is for quiet times in a dedicated audio room and cars are for driving.
Anyone in favour of compressed music should have to eat a loaf of white bread the grocer packed beneath a 20# turkey in the grocery bag.
Anyone too dumb to understand audio should have to shut up - you couldn't enjoy music completely without compression at home
Holand is within the Netherlands, it is not a country.
It’s all about this: (skip to 10:40)
czcams.com/video/Y0xwE2fF1Bs/video.html
@@geddylee501 well, I think to satisfy measurement-boys.
99% of all music is grade A loudness crapola. So the answer to Hans question is no.
It truly isn't that important.
You've obviously never experienced music with a high dynamic range.
@@GTRxMan what do you consider high dynamic range? How loud do you listen?
Say that you listen at 85 db, which is fairly loud, and a typical listening room has around 55 db background noise. Even if the crescendo of the piece is 18 db louder, which would be a whole lot, we're still well within the range of even a modest lp playback system.
Tldr version you don't listen loud enough for a 120 db dynamic range. Let alone 96...
@@jamesrobinson9176 dynamic range has NOTHING to do with the volume on your gear clown
@@Harald_Reindl spoken just like someone with nothing to say, just spewing hate.
Not much innovation in home audio
Do speakers grills/dust covers add to a speakers life span? I know most guys want to show off but is it worth it? My quess is no In my opinionleave the grills on. In the 1970s speaker grills were not remarkable and yet the Zenith or Sony Speakers sounded excellent. Then someone like me curious tore the speakers apart only to find nothing impressive, The speaker cones were ugly to say the least. No Foarm or Rubber Surrounds around the drivers, just tested paper. Yet if one finds one of these 40 0r 50 year old speakers in a thifft store they work 100%. Find a same age speaker with foam or rubber and it's garbage.
🤔Sometimes,especially if you have small children 🧒 or pets 🐱 😅🥹
I imagine there is some sonic benefit to having the cone made of rigid material and the surround of a different much-softer material, so that the cone can move without changing its shape. The problem with the foam or rubber surrounds is that they were (at least at one time) made of a short-lifetime material. I still (naïvely?) hold out hope that a surround material could be found that has a reasonable service life.
@@kc9scott🤗 100 % AGREE
that was the case in the past … which is why I switched to B&W speakers and I haven’t had a problem since 🤗 And IMHHO and many of my friends and family agree …they sound better than many popular new speakers 🤗💚💚💚
@Douglas Blake My guess is well over 95% of the time?
@Douglas Blake 🤗. THANKS FOR SHARING … antidotal evidence … but interesting still 😅💚💚💚
dr dot loudness dash war dot info has they dynamic range of 164K recordings