A Rant about Speaker Measurements - They are Mostly BS
Vložit
- čas přidán 19. 07. 2023
- PLEASE READ THIS! IT'S DEEP!
Don't misunderstand what I'm saying - measurements ARE important, but it's my opinion that they are of limited value when used to judge how a speaker sounds. They can tell you what the speaker is doing under ideal conditions, but they don't include the other major components of the listening experience: the room and the person listening.
Speakers are a complex system and you can't arrive at easy, simple answers when you try to objectively assess the performance of a pair. There are some widely accepted "rule of thumb" guidelines that can give you insights as to how they may sound, but there are no written in stone rules.
Flat response and good off-axis response are two of those objective guidelines, but you have to remember that musical enjoyment is a purely subjective experience. What one person loves another may not care for, where sound "quality" is concerned. There's the expression that "beauty is in the eye of the beholder" and that certainly applies for audio when you swap "eye" for "ear".
Whatever music you are listening to sounds the way it does to you due to three major factors:
- the speakers
- the room the speakers are playing in
- your ear / brain combo, where the brain interprets what the ears are receiving.
"Interpret" is a word that allows a lot of variability. Your hearing system isn't perfect - far from it - and it is subject to a long list of factors that can influence it.
All three factors are interdependent and together they determine what the listening experience will be. If you move the speakers to another room, they will sound different. If you are stressed and unable to relax, they will sound different. Not different in reality, but in your perception of how they sound through your ear / brain interface.
On top of this there is the fact that there isn't an exact representation of how a piece of music is supposed to sound. Music isn't like an engineering sample kept at a constant temperature in a sealed vault and taken out when needed to compare with another sample for accuracy. It was recorded and produced and nowhere in that chain did anyone actually hear the finished work until it was played back on a system in a room and into ears that is different from yours. The best you can do is get close to how it's supposed to sound, because there's no exact reference to ensure exact reproduction. So that throws the concept of accuracy out the window, since there isn't a precise point of comparison.
In other words, music, and this is probably the reason why we can listen to the same piece over and over, is a singular experience every time you play it. Like a snowflake, no two listening sessions are exactly the same. And none will be exactly what the music is, because what the music "is" doesn't really exist.
Where measurements are needed is during the design process. The speaker designer can use them to determine whether there are fixable problems. For example if he listens and hears something isn't quite right, he can use measurements to pinpoint where the problem may be.
On the other hand it's entirely possible to design speakers without measurements. Will they be "accurate"? Probably not, but as I said above, accuracy isn't a possibility anyway, since there isn't a precise reference.
Think of it this way: While the musicians were playing were they measuring how accurate their performance was? Did the engineer measure how accurate his mix is while producing the music?
So if objective measurement isn't of much use, are subjective reviews better? Well, it depends on what the reviewer likes to listen to - it's called subjective for a reason. It also depends on the other factors listed above, so there really isn't any way to know for certain based on either objective or subjective review whether you will like a pair of speakers. The reviewer isn't listening to them in YOUR room or with YOUR ears and he isn't YOU.
But chances are you WILL like them, if they were competently designed and you aren't going in looking for problems to "discover". And as I said in the video, even if you don't love them to begin with, you will get used to how they sound and possibly grow to love them.
Although I didn't mention it in the video, another example of this is creating "upgrade" kits for retail speakers, where the speaker is measured and the upgrade makes the response flatter. Does it sound better? Why was the speaker designed that way to begin with? Could it be that the designer WANTED it to sound like that? I know I'd be pissed if someone undid the work I did to justify selling overpriced crossover parts. - Jak na to + styl
PLEASE READ THIS! IT'S DEEP!
Don't misunderstand what I'm saying - measurements ARE important, but it's my opinion that they are of limited value when used to judge how a speaker sounds. They can tell you what the speaker is doing under ideal conditions, but they don't include the other major components of the listening experience: the room and the person listening.
Speakers are a complex system and you can't arrive at easy, simple answers when you try to objectively assess the performance of a pair. There are some widely accepted "rule of thumb" guidelines that can give you insights as to how they may sound, but there are no written in stone rules.
Flat response and good off-axis response are two of those objective guidelines, but you have to remember that musical enjoyment is a purely subjective experience. What one person loves another may not care for, where sound "quality" is concerned. There's the expression that "beauty is in the eye of the beholder" and that certainly applies for audio when you swap "eye" for "ear".
Whatever music you are listening to sounds the way it does to you due to three major factors:
- the speakers
- the room the speakers are playing in
- your ear / brain combo, where the brain interprets what the ears are receiving.
"Interpret" is a word that allows a lot of variability. Your hearing system isn't perfect - far from it - and it is subject to a long list of factors that can influence it.
All three factors are interdependent and together they determine what the listening experience will be. If you move the speakers to another room, they will sound different. If you are stressed and unable to relax, they will sound different. Not different in reality, but in your perception of how they sound through your ear / brain interface.
On top of this there is the fact that there isn't an exact representation of how a piece of music is supposed to sound. Music isn't like an engineering sample kept at a constant temperature in a sealed vault and taken out when needed to compare with another sample for accuracy. It was recorded and produced and nowhere in that chain did anyone actually hear the finished work until it was played back on a system in a room and into ears that is different from yours. The best you can do is get close to how it's supposed to sound, because there's no exact reference to ensure exact reproduction. So that throws the concept of accuracy out the window, since there isn't a precise point of comparison.
In other words, music, and this is probably the reason why we can listen to the same piece over and over, is a singular experience every time you play it. Like a snowflake, no two listening sessions are exactly the same. And none will be exactly what the music is, because what the music "is" doesn't really exist.
Where measurements are needed is during the design process. The speaker designer can use them to determine whether there are fixable problems. For example if he listens and hears something isn't quite right, he can use measurements to pinpoint where the problem may be.
On the other hand it's entirely possible to design speakers without measurements. Will they be "accurate"? Probably not, but as I said above, accuracy isn't a possibility anyway, since there isn't a precise reference.
Think of it this way: While the musicians were playing were they measuring how accurate their performance was? Did the engineer measure how accurate his mix is while producing the music?
So if objective measurement isn't of much use, are subjective reviews better? Well, it depends on what the reviewer likes to listen to - it's called subjective for a reason. It also depends on the other factors listed above, so there really isn't any way to know for certain based on either objective or subjective review whether you will like a pair of speakers. The reviewer isn't listening to them in YOUR room or with YOUR ears and he isn't YOU.
But chances are you WILL like them, if they were competently designed and you aren't going in looking for problems to "discover". And as I said in the video, even if you don't love them to begin with, you will get used to how they sound and possibly grow to love them.
Although I didn't mention it in the video, another example of this is creating "upgrade" kits for retail speakers, where the speaker is measured and the upgrade makes the response flatter. Does it sound better? Why was the speaker designed that way to begin with? Could it be that the designer WANTED it to sound like that? I know I'd be pissed if someone undid the work I did to justify selling overpriced crossover parts.
The loudness of everything around you doesn't change to fit your hearing so neither should the speakers. Speakers should sound flat in an anechoic chamber. And any irregularities in your room below 500Hz are to be fixed with EQ, and above 500Hz with sound treatment. There are scientific papers about this stuff.
Well Stated John. I’m a musician (trumpet). A speaker is somewhat similar to a musical instrument. For instance trumpet players will have different ideas and perceptions as to how any particular trumpet sounds. Many parameters are similar. There are differences but with speakers like a musical instrument individuals can have different interpretations and perceptions as to how an instrument plays and sounds. So do speakers. I had a friend go to a very large music merchant and he wanted to buy a trumpet. He asked to bring him every trumpet they had in stock. He played over 50. In the end the first one he played was the one he thought played the best and he admitted that as he got past the first 10 horns he no longer could tell any differences. So he bought the first one he played because it made the best impression in playability and sound. Sometimes the same thing could happen in an audio store’s listening room especially if you start with a high quality speaker to begin with and then everything after that will be compared to the first one. Ok this may be an oversimplification but it’s an example. Despite all the test data a prospective buyer might digest over a particular speaker model it’s all about listening and how it’s going to sound in your listening room with your recordings, electronics and sound source, acoustics etc. At this point speaker test results can be tossed because it’s all about your ears, brain, and expressions of music that count more than published/printed data.
@@Aswaguespack Objectively not true. Speakers are meant to reproduce sounds exactly as they are told to. You dont want them to colour or change the artist's intentention.
@@buka9330 that’s very true but there are many speakers not designed this way. Many have “enhanced” or “exaggerated” high end response or a strong prominent bass, some have unnatural midrange coloration. So far too many speakers are not designed to be neutral or accurate in playing exactly what the source puts into the speaker and the output could be much different from the input with many speakers. Not even some very high priced speakers have a perfect in/out response. In a perfect situation 100% accuracy should be the goal but it hardly ever occurs.
@@Aswaguespack I agree, there is a lot of snake oil and bullshit in audio.
You are absolutely right that the room dominates. So that is why research shows listeners prefer speakers with smoothly changing off-axis dispersion because that defines how it will interact with the room. This is why Klippel is important because it captures that 360 degree dispersion data accurately.
I certainly think the more data the buyer have access to the better.
I totally agree. I see so many manufactures that don't even release anechoic amplitude response. One I remember e.g only released the two -10db points which is utterly useless
I'd honestly love to listen to a discussion between yourself and Erin (Erin's Audio corner), provided it stays friendly. I think you will probably both agree on a whole lot and would make some interesting points.
I scrolled down to the comments section to see if anyone else was thinking this 😅
While an individual’s opinion of how a speaker sounds is without a doubt subjective, Dr. Floyd Toole proved that under repeatable situations, “most” people prefer a speaker that sounds a particular way and with training can identify accurately and repeatably which speakers sound better and which had problems. Taking this information and anechoic measurements a speaker manufacturer CAN improve their product. With all that said, probably, once you reach a certain level of fidelity, and eliminate obvious super bargain speaker issues, spending money on room treatments is where the most bang for your buck is--AND--speaker location, toeing end, etc. Yes, in real life the room dominates but I would never assume that you might get lucky and put a crummy speaker in a bad room and have good sound. Ain’t gonna happen.
Most people looking at the Klippel reviewers are aware of in-room VS anechoic response and that's why both measurements are provided. Anechoic is how you know flaws purely with the speaker. Why would you want consumers to have less data on the products they buy? Your room can't really fix a huge dip at the crossover for example - like was discovered with the Klipsch 600m by the Klippel.
Room response is like a fixed mechanical EQ.
@@pablohrrg8677 Kind of but not really. It's extremely difficult to target EQ a speaker with a room. Also, in relation to bass that room gain is basically modes and nulls which have severe consequences in addition to the boost that they give.
If it took the Klippel scanner to discover the huge dip, how much of an issue was it in actuality? Why didn't a listener discover it?
Nothing wrong with consumers having more data as long as that data is actually relevant and they know how to interpret it. My point was (and it pains me to have to explain things twice or three times...) that the data is missing key components that determine what the sound quality will be.
The Klippel system gives a rough approximation of what the developers of that system THINK the speaker will sound like in a typical room. That's like buying a suit off the rack vs going to a tailor. Sure that suit off the rack will fit pretty close, but nowhere good as the one that's made to fit perfectly. It's mediocrity vs excellence, which would be okay, but the mediocrity is being touted as excellence, which is a lie.
The measurement data shows just one aspect of the big picture, but it's being pushed as the be-all-end-all, and all you need to know when you want to buy a pair of speakers. Very good sounding speakers are being maligned because they don't fit the narrow range of "acceptable" (presented as exceptional) flatness and directivity. It's the human (lazy, low rez thinking) desire for easy, simple answers to very, very complex questions, and pandering to people who are more interested in specs to prattle on about than the actual sound quality.
@@IBuildIt Klippel has tons of technical data to back up their claims about the systems accuracy. Both they and many other people with engineering backgrounds seem to agree that outside of the most advanced anechoic chambers in the world, of which only a couple exist, the Klippel is by far the most accurate measurements you can get - especially into the low frequencies. I appreciate the contributions of both you and the Klippel reviewers, but to convince me you would have to provide technical data equal to Klippel to disprove their system - and you have not done that. Subjectively, I have owned tons of HiFi speakers and when I started buying based off good Klippel measurements I noticed a huge improvement. The speakers that measure good are not too bright, not too dull, and sound full with no details missing.
And once again I can see that I've wasted my time. And not because I "failed to convince you that the Klippel data in inaccurate", but because that's not even close to what I was saying in this video.
I'm an "advanced" amateur musician/recording engineer. I've got a few albums under my belt, but I'm not a pro by any stretch. I've recently started doing some "real" work on a pair of Adam Audio near field monitors set up in my newish home office . They weren't wrought from sacred materials by Tibetan Monks, but they sound pretty damn good. Yesterday, I was mixing a track containing drums, piano and upright bass. I was having an unusually difficult time getting the bass right. I rather quickly traced the problem and assumed that it must be the speakers and so I switched to my supposedly dead flat in-ear monitors and finished up the offending section with no more issues.
Once that was done, I decided to give the entire track a once over on the speakers to listen for stereo imaging and sure enough the bass was terrible again.
But as I pushed back my chair and stood up in disgust, the mix was suddenly perfect!
Long story short, the song has a section in the key of C and it just so happens that the second lowest C of that particular recording on that particular track sets up a standing wave in my mixing position. You would think that I would have noticed this before but alas, I did not. It took me focusing on listening to those frequencies for me to catch it.
Needless to say I am now in the market for a measurement mic lol.
REW, UMIK or the PartsExpress equivalent. And start to see how forgiving your ears are at some frequencies out side of the 800 to 4 kilohertz range, and 60 to 500 range. It can be an interesting thing to measure up what you are listening to. Remember sharp peaks or dips are not what our ears listen to. So the response that John showed for his ELAC speakers is actually quite stellar.
John, I worked as a commissioned stereo salesman for most of my life. When I started in the 70s, the speakers we sold sounded very different from each other. It was obvious then which manufacturers used science/measurement/knowledge/quality components (like crossovers) and which were just throwing drivers in a box and adding attractive trim rings and a pretty wood finish. Many brands did not even have crossovers. Most customers back then did not know the difference with a 5 minute demo in the show room. Independent tests, the magazines that measured, were extremely useful to weed out which manufacturers used science, and which were hokum. Measurements are still useful for that. If a speaker measures horribly, with no obvious design goals, then the manufacturer should be looked at with skepticism. There are some high end products that still seem to be designed that way. Determining which speakers would sound best in a customer's room was part of the value I added. I could suggest placement, and speakers that would likely sound best in the room. As you point out, now we have DSP, which is just measurement taken a step further, including room effects. Heck, even Amazon Echo speakers use self-measurement DSP to optimize their sound based on location. Measurements are not B.S. You have to understand what they are telling you.
IF the complete Frequency response, which is a Complex Variable, with the Real Component, the Amplitude response and the Imaginary Component, the Phase response are Hilbert Transforms of each other, that is an indicator that the loudspeaker is true Minimum Phase system. Which means it will deliver a quite reasonable facsimile of the recorded performance. Richard Heyser wrote multiple papers published in the JAES during the 70's. That said, I've been building loudspeakers since the late 70's while earning my first undergrad in Applied Math/CS. The analytical tool I've used since then is a pair of Stax MK III/ RD 7 earspeakers, to contrast with in room, the systems I've built. Number crunching, was my profession, playing guitar and listening to others do it so much better, a very important part of my existence.
Erin's Audio Corner has the good stuff
- audio reproduction is not subjective, it is varying degrees of inaccurate / distorted
- we all have ears and we all hear the same things
So many opinions. I work at this for a living. Designing drivers and finished systems. I do an enormous amount of measuring. And a microphone has it's place. Understanding how our ears listen is also paramount. And knowing how to listen is also paramount. There is a reason why a studio engineer listens in the nearfield. You have little to no dominance from your listening room. It is all primary sound from the speakers. If they have been carefully designed and placed to have an even response you will have a great rendition of what was recorded. The every speaker has a sound issue is a sad truth. We all have microphones. And there are well established best curves known and available. How our ears process sound is also reasonably understood. The sad part is that this audio industry has fought tooth and nail to keep standards out of the industry. There are AES measurement standards, and IEC measurement standards and CES measurement standards. I design to them. And I am not the only person to do so. I can tell you that your loved one will sound like your loved one on speakers that I design. Your car door thud will be realistic. But this means that a great many recordings will sound excrable. ( Tone controls? EQ) Because they are not designed for truly neutral speakers. John basically designs reasonably neutral speakers. And has spent some serious time making a great listening room. Hats off to you sir! Spend a little more time on the off axis versus on axis response and the driver placements to get a truly even set of lobing at the listening position and you will have speakers that are state of the art.
Mark
P.S. Those ELAC speakers actually sound awesome. Our ears filter out those sharp peaks and dips. And I know that you understand this John.
Second thought. Klippel is not the be all and end all of measurements. I have done true anechoic measurements and compared them to multiple mic position processed REW measurements and they are indeed comparable. So as with any tool. In the right hands you can do something useful. A screw driver can be used as a chisel or a hammer used as a screw driver right? It works! But it is not as effective as is possible with application knowledge and understanding as to what can be accomplished.
i can agree with this video in general.
i would like to add something: is the measures output of a speaker has a peak or dip at some point, if that frequency is equally augmented in your room, then things get worse. that is the benefit from anechoic measurement.
the room could also fix dip and peaks, but it is a crapshoot if it works. so in general a flat response works in most cases.
i measure with my ears and found that microphones do not give any indication how my ears hear things when it comes to loudness.
Very well said. In addition, many people don't look at ALL the specifications for a speaker. Sensitivity being the most overlooked spec. Same thing with slew rate for amplifiers. I'd bet that 90% of the audiophiles out there have no idea what "slew rate" means or how it affects anything. Audio systems are like legos, you put them together one way and see how it sounds, if you don't like it, dink around with it until you do like it. Doesn't matter what the spec sheet says, if you are happy with it, that's all that matters.
Brave and spot on statements. Measurements gives the users an idea so they can compare. Nothing more. Real life is another story, as expected.
Many "objectivists" believe they can take conclusions from Single frequency or random anechoic measurements. "This engine can provide 500 HP to your car because that's what the dyno measured"
I'd like to add to the list:
- Speaker setup like toe in/out, spacing between speakers and listener etc, time of arrival (especially with multichannel audio or stereo over quad speakers).
- Room treatment
- Recording and mastering.
I agree listening to music is mostly subjective. Though there are attempts at standardizing listening to audio like dolby dts/atmos. It tries to calibrate the room/system to a reference target. It's for movies of course but it's still insteresting for music.
As for components, I've been experimenting with this the last few years a lot with my DIY quad 3 way full sized speakers. First I used off the shelf crossovers with electrolytic caps. The phase was all wrong as you might expect. I made some L-pad adjustments to get it mostly right SPL wise and corrected the rest with umike1 and rew. I had all sorts of issues with holes in the sound stage, wandering vocals/instruments when playing at different frequencies/notes the phase would shift and the stereo balance changed. Also due to a large dip in the base region I could clearly hear a base progression from low to high that the amplitude changed while on headphones it sounded perfectly equal in volume.
When I've built my own crossover using measurements and xsim (followed Audio Judgement 'How NOT to design a crossover') I've particularly focused on correct phase and impedance, and of course a flat response as much as possible. I used air cored inductors, polycaps (yep that cost me an arm and a leg, but it was worth it) and cheap resistors for now until I've settled on the L pads. It took a week to get used to the difference in sound, like you said you get used to a bad setup and you might reject a 'correctly' sounding system due to what you're used to. But I've settled in and it sounds amazing every time. I had so much trouble getting the stereo image centered before but not anymore.
I think that's the importance of good speakers, a good flat SPL and phase response. A good speaker setup and room treatment does the rest. My speakers now follow the room response curve naturally without any equalization and it sounds so much more natural. Equalization can introduce phase errors, ringing, clipping etc. just making it sound 'off'. I prefer no EQ at all.
To me get this all right and then adjust to taste like base boost or 'smiley curve' etc to taste is the way to go.
As for commercial speakers, sure some designers put in the effort but sales could enforce cheap speaker parts. After changing to good quality parts I was convinced, subjectively or otherwise. It's what GR-Research said, smearing of vocals and instruments is much improved. I was sceptical, and I still don't believe in the snake oil power cables and interconnects etc. But removing steel parts from the audio path (tube connectors, caps using steel in the leads etc) and polycaps instead of electrolytics did make an audible difference. YMMV
You're wrong about EQ. Modern DSP eq's can compensate for and eliminate phase shift, and have zero clipping or issues with ringing or smearing. EQs are now perfect, and there is zero reason (and imo excuse) not to use them. The main issues with older EQs wasn't even the phase shift but the analog circuitry - noise and nonlinearities. That doesn't exist in DSP. No speaker sounds good flat, unless you're using a sub, to me subjectively because I prefer a low bass boost. Use a mini DSP 2x4hd or a good VST EQ with a computer/dac as the source and there are no downsides. You can also test minimum phase (classic) EQs and linear phase EQs.
Also phase shift isn't even an issue in all cases... Look at it like this. If you put a high q boost or cut in the sensitive range of our hearing, a standard minimum phase EQ (old school), then yeah that could cause an issue and be less than ideal, and potentially noticeable. That's because you go from one phase, to a significant shift out of phase, and then back.
But if you were to use a low frequency shelf boost that starts at say 120 and maxes out at 30 or 40hz, that phase shift is meaningless, not least of which because we're not sensitive to timing information down there, and each channel is in phase with each other - it's not like a sub out of phase with the mains, there's no cancellations or constructive interference. A gradual phase shift on either end of the spectrum isn't harmful subjectively. I'm open to being critical of rapid shifting phase relationships within the middle of the spectrum.
I can prove this to you objectively so you are absolutely without doubt:
All your music is already EQd. All the "negative" effects of EQs you're talking about, if true, are already in the recordings, because virtually every single album ever recorded mixed and mastered has already had a variety of EQ effects applied, to each track, and to the whole mix. And you don't know what type they used. If it's older music, it was almost certainly linear phase eq, which altered the phase. Modern producers have access to limitless EQ options, and can choose minimum or linear phase. You wouldn't know.
So the very thing you're avoiding has already been done to your precious music.
@@Artcore103 I'm aware of that and I'm no expert on audio processing. I've done some experiments with eq-ing speakers with minimum phase etc it still sounded like cr@p to me.
And I had the eq on average surround receivers in mind, no audiophile equipment by any stretch of the imagination. It allows you to eq each speaker individually, a recipe to get phase alignment issues whatever precious music you play. Agreed a sudden phase change, usually in bumps and dips in spl can have noticeable effect. I don't like a flat response from my listening position either, just a flat ish respone from 1m away from the speaker (with some IR windowing to remove reflections).
I agree, just the other day I had a mate around listening to my system which I thought lacked a little bit of Mid/treble, but after listening to it with him, I thought there was way too much treble, I believe drawing the curtains back may have propagated the effect more. Thanks.
start with clean signal than have fun with placement than jam out love this channel
I totally agree with you that our mood matters to how our systems will sound. If I have a day of noise pollution I won't even bother listening, because clearly the audio center of my brain is done and wants a rest to recover and reset. It's kinda like watching a screen too long and your eyes (or visual center of the brain) want a rest. My high end system sounds great most of the time, but sometimes I just have to turn that sh!t off. If the first song or two doesn't sound good, you might as well turn it off and wait for a better day. I understand the reasoning behind the drastic differences and I am okay with it. We are all human.
John, I think that many of the issues you bring up evaporate with controlled/constant directivity designs. A speaker with bad directivity performance that targets flat anechoic will almost always sound bad. Opposite is true for CD. And CD designs IME are just "relaxing" and sound much more consistently good to me. I think this is because our brain doesn't have to work so hard to process and compensate for the mismatch between the reflected sound and the direct sound. They are also less affected by room acoustics and dramatically so in the case of 20hz-20khz dipole/cardioid.
Hi John. I agree here. As someone who regularly measures speakers I know the importance of all of them. On axis, phase relationship, horizontal and vertical off axis and spectral decay etc. But I also know that speakers are for listening to and not viewing there perimeters on a screen. Some of the poorest measuring speakers can be the most enjoyable. But from that you learn what to look for and also when something doesn't sound right to see why. Once there in a room things change again but the measurements do point you in the right direction. I often after the measurements are done tweak a crossover by ear following hours of listening. After all it's about the listening and the music. Love your videos. Cheers. Matt
Thanks Matt. I've been watching your restoration videos - good stuff!
Most viewers didn't fully grasp what I was saying, instead half listened and reacted to the title.
It's this incredibly banal need for people to have stats to point to that are supposed to represent sound quality. They miss the point of listening to music - it's art, not a computer printout, or a set of rules to follow.
@@IBuildIt I completely agree. Also very well put. It's like judging a car purely on its 0 to 60 time and buying it for that reason rather than driving it!
Anyway 🤣. Keep up the amazing work. Your woodworking skills continue to amaze me. All the best. Cheers
Another example is food - judging the taste based on the precision of the ingredients. Whether they are in the "correct" ratios, etc.
How boring would life be if these guys got their way? Establishing a standard for speaker performance, like it's a toaster or the store bought sliced white bread you'd put in it.
And on top of that, the absurdity that most of them would actually hear the difference!
@@IBuildItyou are so close to wisdom and truth on this topic and then you go off the rails with comments like this. You confuse sound production ,the art of music and audio recording and pre-production. Ruminate on the difference between the two and they're different roles and I think you'll have more cohesive and consistent thinking on the topica. Dr Toole's book would be a recommended read.
@@IBuildItattack in this manner seeing that most of the standards you refer to are result of researching what people like normally. It's descriptive not prescriptive.
It is refreshing to read comments that have as much knowledge and experience as the presenter. I am a rank amateur in audio. The assumption that everyone hears the same way the testing sound mic does is definitely not the case. What may be flat to the test equipment can be a complete mess to a person's hearing as they hear different frequencies at different levels. Have a look at your latest hearing test and you will see how "flat" it is. From the listener's perspective sound quality is subjective and they will get greater benefits by moving the speakers around, or adding a little room treatment than changing equipment. That being said, not everything is snake oil but different does not always mean better, and a change is not always bad.
I wonder if the frequency change you are experiencing is due to room temperature. I have seen it in my system. I have no AC, so the temperature in the morning or evening is usually different throughout the day and by season. I can see it when doing frequency sweeps. The response seems to be higher when it is warmer. Or is it the temperature of the actual components that make the difference, like the speaker crossovers or amp components? It would be interesting to hear if others see the same thing with their systems.
I agree fully with your assessment of the quality of sound, John. I have played acoustic guitar for 12 years, and been into audio for most of my 76 years. Not once have I heard acoustic guitar reproduced accurately, by myself or others, so that it sounds like it does when I quietly play by myself at home. Sadly, I find reproduced sound lacking.
Exactly: there's no reference point. There's videos on youtube about how the ideal "kilogram" is manufactured and stored. If you need actual precise measurements, you need a reference point. There's no such beast in this endeavor.
Well I'd rather have a perfectly flat speaker with perfect dispersion and low distortion and put that in my less than ideal living room and then just have to deal with that with a few absorbers and some EQ, instead of having a wobbly speaker
So yeah, a good measuring speaker is a better starting point, so why not? Wouldn't spend audiophile money on a maybe ideal speaker (unfortunately sometimes the money goes into other things than sound quality thought), but up to maybe 1000 euro is fine for me. My 600 euro Ino Audio piP does everything I need though with some absorbers and EQ, and that EQ I got to using measurements ;)
You're right, loudspeaker measurements are a minefield.
Fortunately we dont listen and experience playback anything like the way mics capture the very same material.
That said, ... there's been a quest for several decades to establish measurable characteristics, and how each correlates to certain subjective qualities of a listener in a room.
Dr Toole, JBL Northridge, NRC in Canada, Wolfgang Klippel, there's many devoted to establishing common benchmarks from which to move forward.
Collectively their work is extraordinary.
However its imperative the relevant data is gathered competently, interpreted correctly, and implemented wisely.
There is recognized minimum standards for professional systems in a room.
Actually moreso with film and less so with music.
I believe Toole coined that the "circle of confusion".
Always thought provoking John ... good stuff.
The measurements themselves aren't the minefield, but they way they are presented and received by people who want simple answers certainly is.
If you listen to interviews with Toole, you'll hear him saying pretty much the same things I said in this video. For instance, that there's no truly accurate representation of a piece of music to judge the so called precision on. Also how the measurements are extremely helpful for the designer, but it comes down to actually listening to determine sound quality.
This is ironic, since I have so many comments from guys invoking the name of Toole to critique me, but not even aware that what I'm saying really isn't any different from what he says.
Says this: guys only hear what they want to hear to back up their point of view and just ignore the rest. There's no nuance, no subtly, no deeper thinking, just quick, easy answers.
@@IBuildIt
Indeed
The better sounding "flat response" speakers have inert cabinets and owners of whom know to decouple their speakers from anything that will resonate, treat room problems and find the best placement for performance, not where they look pretty. Too many so called "audiophiles" care more about "pretty" than performance. Flat speakers tend to sound the best in a bad rooms and have the potential to sound great if the room problems are addressed. Everything is relative. Most rectangle rooms are pretty much the same and if people put their systems in really bad rooms, then they will never sound very good. Flat measurements mean a lot.
People have personal preferences (like bloated boomy bass), but if we want to talk about anything we need to agree on a standard (natural and realistic sound, of which we all heard our entire lives, no matter our personal biases). Thus a flat response is a great starting point. Crappy speakers are all over the map to start with and can never be tweaked to sound great. Even DSP can only correct so much without causing more problems. I have see it many times when people build speaker and want so badly for them to sound good, but it's difficult for most people to not be influenced by their personal cognitive biases. We all enjoy taking pride in what we create, but not all of us can evaluate honestly against our biases. If any speaker (no matter if I built it) doesn't sound right and hurts my brain (listening fatigue) it's crap and I can't listen to it.
Why do you even measure your speakers if they are, "mostly BS". Just use your ears (really your mind, because that is where sound is processed). Barring hearing loss, people's ears are pretty much the same, but minds very greatly. I would argue that listening fatigue comes from our minds, not our ears, simply because it can occur at volumes too low to ever harm your ears. It's all heard in our heads.
It goes even beyone the room, the used cone materials, the filter components, the tuning and in first place, the ability and taste of the listener to hear, identify and proces, basted on subjective preferences. That's why i made my own blend (speaker untis), design and tuning in my room to get what i prefer to hear. It's that subjective...
While I agree that changes in equipment driven by what you call wanting is very true with audiophiles, but sometimes it's not just a change in sound, sometimes it's obviously just better. and that's what I find drives audiophiles, not just a change but a change for the better.
You should have ended your speaker and room listening rant with a …Mic Drop . Great job explaining
Frequency Sweeps dont tell the full story.. because speakers tend to reveal most of their Distortions, when multiple sounds are playing at the same exact time. I had speakers that sounded good... but when compared to a much better Audiophile speaker... I noticed that for the first time, I could CLEARLY understand the lyrics that a singer was singing. With all of my previous speakers... this mans vocals in this particular point in the music... were getting distorted amongst the other instrumentation and effects. But this particular audiophile speaker, was so "distortion free" ...that it was able to accurately deliver the sound... in a way that I could easily differentiate his vocals, and all of the other instruments as well.
Of course, these speakers were the EPI 100v speakers. They are sealed, and have very powerful magnets on the woofers. The extra magnetic power, allows for FAR more precision, control, and much faster acceleration potentials... compared to drivers that have much smaller and weaker magnets + weaker coils. They also use a thicker gauge wire winding on the woofers, which automatically filter out the High Frequencies... thus eliminating the need for an additional crossover component (which arguably could also cause distortions).
As for the Measurement system... Its still impressive, in that its able to get readings, in 360 degrees of the speaker. Im not so sure that the room matters in those readings, because the mic is facing the speakers, getting the soundwaves before they hit the walls. It might depend more on how the capture program is set up. That said, they probably should pad the walls, just in case.
But like Ive said before... These sweeps have virtually no bearing in a speakers true potentials. As once you crank your speakers up... and depending on the dynamic layered complexity of the actual music... you are going to get distortions.. and the amount of distortions, will depend of the drivers capabilities. Since Distortions are rarely tested on anything but high end headphones... and even then might be flawed in the way that they test... most of the presented data becomes flawed.. almost meaningless.
Also, the EPIs are one of the rare speakers... that I can place virtually anywhere in a room... and they will sound fantastic. This isnt the case with most speakers. Heck, Ive had them under my desk, and partially blocked... and they still sounded flipping fantastic (meaning, you couldnt tell that they were partially blocked, by merely listening).
Thanks for sharing.
Are using MiniDSP 2x4HD for one channel ?
It never ceases to amaze me how TVs have hard numbers for picture quality (and having a propercroom for it is an user issue), yet audio guys still act like a vibrating material is some arcane unmeasurable magic that has to be experienced. Yeah... no.
I ain't watched yet.. but that's a clickable and provocative title!! ❤
Too true John!
Elephant in the room is non-linear distortion,
Power and dynamic compression. call
And especially IMD.
FR ( linear distortion), is not of much value everything else considered. It's just the fact it's easy to test
Listening is very much a process that is internal to each person and no two are the same. Nor is one individual the same from moment to moment. Hence why being able to make adjustments to your system, while listening, is extremely important.
*GR Research beware!!!!* 🤣😂🤣😂
You nailed it. So much is subjective and I hate when people ask me what the "best" speaker is.
For me, it is someone that overspent on his system and then asks me what I think about his sound! What do you say, it sounds like crap?
Your brain will make adjustments to the sound coming out of your speakers, provided you don't outright hate the sound, in a way that new speakers might not sound as preferred over speakers you have had for awhile until you listen to the new ones exclusively for a sufficient time. Then swap the older speakers back in and you may find you now prefer the new ones. Generally speaking.
I agree and disagree with your theory. I believe it's best to start with a flat FR then adjust (tuning) for your room and ear. (Subjective) Wait until you start rolling caps. Hahaha.
I think Totem stopped running ads in Stereophile because of JA's test results and comments.
Unfortunate comments like this, including the more unfortunate one in writing under the comment section, originate from absolute lack of knowledge about the subject and I am saying this with the most positive and compassionate intention. The conditions which are referred as “room” are nothing but a mix of anechoic and reverberation chamber conditions. The name brands can hardly afford an anechoic chamber, the ones that sell speakers less than 100k$ don’t even bother and depend on gated measurements, maybe reaching down to 100Hz, but even more rare manufacturers, maybe none left as of today, have a reverberation chamber to check the power response of their filter setup. So, if you look at the frequency response transfer function, it is tip of the iceberg. What about THD, what about IMD, what about the impedance loading, what about the power response which will define the behavior of the system in a listening environment, or even off-axis response if the reverberation chamber is busy for the day? Such measurements are not mentioned in any of the universities which have courses on this topic, and none of the text books have enough explanation. The people who share information are commercially charged and their information is solely aims to sell their product, which further confuses the end user. Some even have a misleading results like Klippel (just because in some comment somebody dropped the name). What is misleading, you may ask. Have you ever sent an email to their support and asked why they are sharing the Le(x) curve with reference to 1kHz while your sub-woofer will never go beyond 150Hz? They will avoid saying that they can not do any other frequency other than the 1kHz because of their system identification model. So, does this qualify as a “measurement device” if you are going to be measuring only what you are told to measure, rather than what you need to measure? So, get the pair of speakers you like, throw those rubbish review magazines that have hidden commercials like Google, and enjoy your audio adventure as you don’t have to prove anything to anybody if you are enjoying what you hear. Nobody can judge you anyhow because no two people’s hearing is the same. Don’t believe me, go read about Nura Headphones and how the hearing can be equalized according to a persons ear channel impedance.
For the most part, T/S specs are useful for enclosure planning and some basics of performance limits. Beyond that, every single speaker sounds vastly different just based on their construction. I have found no good way to determine a good sounding speaker versus a bad one other than actually listening to them. Every single driver, every combination in a system, it all varies.
You hit a lot of good points on the entirely of the system and every single step along the way. Every step matters. Every step effects.
I don't know why you gotta harp on the other guy. I like watching his stuff too. Can't we all just get along?
To be fair, key things he brings up are important, but like you, I've played with raw drivers and played with all the specifics. We learn that so much of this stuff is adjustable, correctable. But...if you take a premade at face value and only face value, then it will have the one output. That is the constraint of premades and simply using them as-is. And for that, yes there are oddities you live with, have to compromise, and start to worry about peaks, dips, off axis behavior, toe in/out, and whatever else. In terms of buy and use, that's all that's left. You don't afford yourself the luxury of all the other customizations.
If you play the music loud enough, frequency response doesn't matter....lol.
It's either he doesn't understand the measurement or doesn't know what they actually mean. So this is basically a ten minute long apology for this sorry
recordings are so far out of whack from each other that a good graphic equalizer and L Pads are still king imo. The Doors LA Woman v Dire Straits Brothers In Arms v Black Sabbath 13 are great examples of three extremes, being in order, flat, crisp and full powerful sounding. No matter what amp, speakers and preamp I use the three recordings have the same reaction; one is ok, one is great and the last one is AWESOME!
Well once you get to a certain age inevitable despite of the measurements you will not be able to hear it anyways.
If we could hear music "as it was recorded" we probably wouldn't think too much of it. It's because it's recorded in controlled conditions, and then remixed and mastered, that it sounds as cohesive as it does.
I watch many of your videos and respect your opinion John, but I disagree with you in this case. Every component in the music reproduction process needs to be designed for flatness, except for the mixing.
Why? The engineer mixes it to sound the way he wants you to hear it, hopefully using flat response amps, monitors, etc.. If the recording is not flat, it will change the response. Also applicable to the CD player output signal (or whatever source you use): non-flat response will change what you hear. Same with the amp: if it is not flat, it will change the response more. Ditto with the speakers: another stage of changing response. All these changes compound!! If you lost a little bass and treble, and added peaks & dips every step of the way, by the end of the line it will be even worse.
Without flat response in every stage of the audio equipment chain, the resulting playback could be very, very different from what the engineer wanted you to hear. You might like this distorted version better, but that's you. Another listener might feel very differently. If systems are aimed to be flat response, at least it is going to sound similar on every music system you play it back on.
Is perfect sound reproduction obtainable in practice? Not with today's technology, but it's get better and has sure progressed a long way from the old phonograph and tube amp days. Some people like the coloration that imperfect response adds, and that's fine. But if no system comes close to flat reproduction, you might hate the way your favourite song sounds on your system, but love the way it sounds on your buddy's system. For a different song, it might be the other way around.
Does the room further change the response? Of course it does, and that's the hardest one to control, but as you know there are things you can do to fix that too.
I agree, see my comment please. Studio recordings determine how a song will be heard no matter what the playback system is.
Measurements tell you if there are any glaring problems with the speakers. If your room is crap then get it fixed. Ignorance is bliss.
The title should be "Frequency response measurements are BS".
HD and IMD measurements are absolutely indispensable. If a speaker has high HD or IMD, it is going to sound bad no matter what room or DSP config you use.
Yeah theirs so many equations with measurements theirs no set standard
I haven't read all the comments but there's one more variable not mentioned in the video, and that's *signal source* . Are you listening to a DVD, CD, CZcams, Netflix, Amazon Prime? How much compression is applied. There's so many things to consider. To be fair, John did brush on this idea by mentioning the desire to hear the signal as it was done right in the studio. Great video, John. I enjoy your content!
Yep ... I guess that is why great headphones are "interesting", but, even beyond the listening environment, referring to your concept of a "box with the music in it", in most cases, anything recorded since to the 1960's has been multi-tracked, mixed and remixed, so that there is no "standard" to go buy, unless you could make an absolutely pristine recording of the sound that the producer heard when they said "that's it!"
I admit that I did try to get my "homemade" speakers pretty flat, measured at the listening "point", and using a parametric equalizer to get it reasonably "flat"... Measured a metre to the left...hah! And, in a small home, no room large enough to carry very deep bass!
Tell that to an audio engineer.. 😂
I wonder how your favourite albums would sound if they were mixed on radio shack speakers.
The title is clickbait BS, everything else is 100% correct.
When you built speakers you need to measure them regardless of the fact that the room f..ks it up again. It would otherwise be too much of a guessing game
It doesn't matter where the music was recorded. It matters where it was mastered, and that is in the mastering studio. Your setup should aim to be what a mastering setup is. Also most people hear things the same way, I think it was some harman study with the only real preference in the amount of subbass. In-room response should be slightly tilted down. You eliminate reflections by having sound absorbers on the ceiling, carpet on the floor and some bigger absorbers behind your speakers and behind you. Also speakers should ideally be coaxials in a waveguide. Everything weird going on in a room below ~500Hz can be fixed using EQ (unless you have a bass cancellation). Above that you need to fix it with sound absorbers (that is ofc if your speakers measure flat in the ideal scenario). And at around 105Hz subwoofers should come in and be at least flat, or better for most people around +6dB. Also amplifiers and cables stuff is mostly a scam. Just get or make a decent amp (like -80dB thd) and you'll be fine. In the end music is very much about having a good frequency response. First you start wtih flat (on graph) speakers and then you make them flat in the room. All of this is objective. That's it.
This is what I was thinking but articulated in a much more technically competent manner! 😅
You da man... 😸
Oh, I read the your deep write-up.
And I am a little bit of upset 😅 and thought that you somehow understood this technical measurement stick..
I didn't see this coming 😂
If someone bought a pair of speakers from you. And own their property they undone your work. And forget if you believe or not that more expensive components or not have any impact or not.
By undoing your work they on the speakers they own they have somehow arrived at a conclusion that if the value (mH, Ohm or uF) if the inductor, resistor or capacitor should ve of another value or the topology of it they needed to change.
Why should you be upset?
As you know the owner taste and preferences changes by time and he is maybe adjusting to that changes. And even your taste and reference are moving you are not that special in a vault that we can fetch when we want a opinion.. if you "voice" the same speaker today then you will most likely not arrive to the same (mH, Ohm or uF) if the inductor, resistor or capacitor values! (You are a floating variable)
So you have nothing to be upset about. 😊
Only one thing to be upset about, if the owner of your built speaker change/modify them and dont disclose that and say that it is your spekers.. the listener is fooled that you might have "voiced" them like that.. and that is not cool! 😅
I think you are throwing all research and development in the audio arena under the bus from the last 70 years.
And that is not something I myself thinks is cool. 😢😊
Yes, your preferences and mine is not the same. Yes where the speakers is placed, angeled or where the sweet spot is placed in the same room will effect the sound and also if we change the room entirely.. as you say.
ALL that that plus all the issues you brought up and difficulties is just bla bla bla.
I meen that if we just decade after decades just complaining that this and that makes all of HOW a speaker will sound in your room just pointless.
Then we will not progress ANYWHERE and being stuck to listening to subjective reviewers withe their ever changing preconditions. When the goal post is moving around like a boat in a storm..😮🎉
What I meen is research has moved us to where we can see on the measurements how wide the dispersion, how deep down goes the extension and so on and on. We also get predicted in room response. Those things is as static that we can get in 2023. With all standardizing work tht has been undergone.
All that is what we have come as a example to so far from research and development.
All THAT is what you throw under the bus. 😅
I would never thought that you would prefer a subjective reviewer that as you said your self his mood changes and would for example could without a problem say:
"the base extend low when I listening to them" and saying the exact same sentence 6 months later for another brand/model of speakers.. or maybe changed listening location, room and whatnot.😢
(and those two speakers could most likely have a very different bass extension or levels.. ..but the reviewer had no distinction between them and the poor consumer cant and will never know..)
Measuring is knowing that is a little part of why you probably do your own measurements.
You learn and optimize fom those things you learn. But not all can invest time and money in all of the things you have been doing.
Kudos to you and all of your lovely sharing!❤👍🎼
And sorry for my rambling and that were just from reading your video comment, I probably are to scared to watch the video now.😂🤣👍
B I N G O...! You've got it right on the money...!!!
When I read the title I was like WTF!!! But you are right that speaker measurement is mainly interesting for builders/manufacturers so that they can create a good base for a system. It can be useful for room EQ and finding "flaws" in your room acoustics but in the end it all comes down to one thing : do you like what you hear.
A great point you make is that people get used to what they hear. One example is when some "audiophile" replaces all the caps in his speaker x-overs and starts breaking in those caps over a period of several days (maybe weeks) and claiming how it sounds so much better after a while ... well, I am pretty sure that the main thing that has been "breaking in" over that period is the guy getting used to what he is hearing (especially wanting to hear the difference).
My advice to people is to maybe start with decent speaker specs (doesn't have to be expensive speakers) and then start listening to them doing lots of A/B comparison. If you are one of those that prefer those cheaper speakers ... good for you, as long as you like it.
PS : I have a set of very flat studio monitors and it's something you have to like hearing, I personally like the sound but most people are not into "uncoloured" very flat response. Depending on the type of music I listen to I sometimes up the LF a little because flat sounds very "cold".
When a music is recorded in a studio, an environment is absolutely artificial. And only an experience of a sound director makes it kind of alive. So, it’s not possible to reproduce something that didn’t exist. You are free to build your sound )
Adding that the "final mix" is a collage of cutouts of different recordings recorded separately.
It's not DSP... DIGITAL... if analog is going to the DSP unit. DSP sucks the soul out of music, unless under 200hz compared to nice passive caps and inductors and resistors. However, eqing analog like crazy is the absolute worst sounding. At least someone who does DSP can say is it's in the digital domain.
I like the title😂😂😂😂😂