That's what I personally find/believe as well, but it's certainly not accepted in the audiophile community. Why do people keep spending big $$ on amps/DACs? Why does Amir keep reviewing them without putting them into perspective? Is a $2000 amp actually going to sound better than a $200 one? It's not comprehensively answered here.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Master Thread: Are measurements Everything or Nothing?
- Thread starter amirm
- Start date
That's what I personally find/believe as well, but it's certainly not accepted in the audiophile community. Why do people keep spending big $$ on amps/DACs? Why does Amir keep reviewing them without putting them into perspective? Is a $2000 amp actually going to sound better than a $200 one? It's not comprehensively answered here.
Because that is not the answer we’re looking for. The question most of these reviews answers is: is the device properly engineered (for a limited set of testable parameters). Sure with speakers you get the occasional listening impression thrown in for fun, but that’s about it.
There are plenty other topics going into the audibility of certain aspects in sound.
The problem is that if I could pass a blind test and hear a difference (with sufficient statistical reliability to suggest that I wasn’t guessing) but 2 other people sitting with me couldn’t hear the difference, would you accept that there was an audible difference, or would you assume that I just got lucky?
It doesn’t really matter, as I can’t prove anything on an Internet forum, but, in case you’re interested, there’s some good and fun ‘blind tests’ on the following site with which you can test your own skills.
Free Online Audio Tests, Test Tones and Tone Generators
Tests your audio equipment, speakers, room acoustics and hearing. Audio signals and test tones playable online. High resolution sound test files available for free download.www.audiocheck.net
It isn't about proving it to anyone else. It is more about setting up conditions for yourself that will let you see behind the curtain. The Wizard is actually naked...or however the story goes.
It isn't the nearly impossible task that many try to make it out to be (much easier to dismiss as 'too hard' that way after all).
I mixed up cables going into my preamp, matched levels with a multimeter and switched between DACs with a remote. Not nearly good enough to pass critical muster, but enough to teach me much more than I expected.
Goodbye multi-kilobuck DAC, hello 9xJBL home theater.
I disagree, at least, that's the question I'm trying to have answered, and I think should take a higher priority in the reviews. Yes whether the device is well engineered is good to know, but in the end it's the audio that matters - I mean, it's the first word in the title of this website. Electrical characteristics are far removed from audio, but mic response is much closer. In the end, audio is a subjective concept - it only exists in your mind, and isn't an inherent property of oscillating vibrations in air or fluctuations of electrons in some piece of metal.
Cool, so what do we use to determine what sounds better to someone else? An MRI machine? May cost a little more than an AP analyzer. Plus those wishing Amir to do the measurements will need to ship themselves with their audio equipment. Doable, you think?
The reality is that it is just as difficult to prove components sound the same as it is to prove they sound different. First there is the listener variability. My wife is happy with her $79 Amazon Echo. So everything is going to sound the same for her. Second is the quality of the rest of the components. Speakers and room will add a huge amount of variability to the question and most may lack the transparency necessary to ascertain subtle differences. Then there is the source material, if it's say 14 bit resolution on a not so good recording, then will it matter if one DAC has 17 bits and the other 19 bits of linearity?
I go back to the point that many many times, we realized the measurements we thought were complete, really were not. Very often this was when a new technology came about that cured the ills of the previous technology, but had it's own set of problems. This happened when we went from tubes to transistors and then again from LPs to CDs. So what makes you think that this time it's different and we now know exactly what to measure? History would say the odds are against you. But I will not buy equipment that has measurements that don't show obvious flaws.
All I'm saying is that it's quite possible, if not likely, that there are many components that measure well (not necessarily identically), and that they may not all sound the same in a high resolution system for a discerning person using great source material. Even using the same DAC, it's not hard to hear the effects of oversampling (assuming the DAC is not doing this internally). Classic Nyquist theory would say we cannot hear it, but that ignores all the practical limitations of filtering in the analogue domain. And there are many more filtering variations beyond oversampling that one can hear.
Getting back to the topic of the thread, I compared the bridged mono AHB2s with my Constellation Centaur mono blocks. I was not trying to prove anything, just having some fun to see which one was better. The engineer in me was telling me that the AHB2s should be pretty good, the audiophile was saying the $50,000 Constellation had to be better, after all The Absolute Sound loved them. My TAD R1s are very resolving, but they can sound harsh with the beryllium coax. So after many years I had ended up with the Constellation amps as they are very musical. But now subjectively for me the Benchmarks were even smoother than the Constellations. Certain recordings that were difficult to listen to were less so, great recordings were even better. So I sold the Constellations.
Another important point is that the AHB2s don't simply make everything smoother. Crappy recordings still sound bad, and I don't sense any loss of resolution on really great recordings. So I don't fear they are simply masking things to make everything warmer sounding.
I ended up buying the Benchmark interconnects and cables. Sure it's possible Amir found something on Amazon that works just as well, but the Benchmark cables are really well made, and the last pair of cables I bought on Amazon were defective. The SpeakOn connectors are great, polarity will be perfect every time, just make sure if you buy on Amazon that you get a real Speakon, the knock offs are not reliable.
The real game changer for me was when with the help of Andrew Jones, who designed the TADs, I was able to vertically bi-amp the speakers using a Pass XVR1. Andrew still had all his anechoic measurements on the TADs, and the details of the passive crossovers in the speakers, and we were able to come up with a hybrid bi-amp such that the slopes and phase angles played well with each other. The low pass in the speakers was bypassed such that a pair of AHB2 is directly connected to the 10" woofers with no thing more than a cable in between. The low pass is done by the XVR1, so there is no loss of damping factor and no over saturating of the iron core inductors in the stock low pass. The speakers are also tucked into corners, with RPG damping to avoid mid to higher frequency reflections. An Audiolense convolution filter EQs the bass with 10-15db of attenuation, and a -3db point at 17 hz. So for the most part the woofers are not moving much and are operating in a very linear manner. There is another pair of AHB2s connected to the coax drivers using the internal high pass. Andrew also routed this via the high pass on the XVR1 to increase the slope of the overall high pass. We were lucky that this happened to work well with the low pass on the woofers.
FYI, Andrew Jones's twin brother was one of the engineers that worked on the design of the AHB2s. This is a brilliant design that measures spectacularly and sounds great too. I suspect much of that has to do with the lack of crossover distortion and the fully regulated power supply. As long as the amps are not overdriven, they are essentially performing with near zero distortion from the first watt to the last. And if you clip them, they let you know and turn off. That happened once with a super loud bass passage before we implemented the convolution filter. The bass was likely 10-15 db louder than it should have been.
All in all, I went through three iterations of AHB2s. First was a pair in bridged mono. Then second was two pairs in a passive bi amp. And finally two pairs in the active bi amp. So far it's the most satisfying system I've had. No limits to dynamic range across the frequency spectrum. Totally quiet, the music just pops out of a silent background.
I go back to the point that many many times, we realized the measurements we thought were complete, really were not. Very often this was when a new technology came about that cured the ills of the previous technology, but had it's own set of problems. This happened when we went from tubes to transistors and then again from LPs to CDs. So what makes you think that this time it's different and we now know exactly what to measure? History would say the odds are against you. But I will not buy equipment that has measurements that don't show obvious flaws.
All I'm saying is that it's quite possible, if not likely, that there are many components that measure well (not necessarily identically), and that they may not all sound the same in a high resolution system for a discerning person using great source material. Even using the same DAC, it's not hard to hear the effects of oversampling (assuming the DAC is not doing this internally). Classic Nyquist theory would say we cannot hear it, but that ignores all the practical limitations of filtering in the analogue domain. And there are many more filtering variations beyond oversampling that one can hear.
Getting back to the topic of the thread, I compared the bridged mono AHB2s with my Constellation Centaur mono blocks. I was not trying to prove anything, just having some fun to see which one was better. The engineer in me was telling me that the AHB2s should be pretty good, the audiophile was saying the $50,000 Constellation had to be better, after all The Absolute Sound loved them. My TAD R1s are very resolving, but they can sound harsh with the beryllium coax. So after many years I had ended up with the Constellation amps as they are very musical. But now subjectively for me the Benchmarks were even smoother than the Constellations. Certain recordings that were difficult to listen to were less so, great recordings were even better. So I sold the Constellations.
Another important point is that the AHB2s don't simply make everything smoother. Crappy recordings still sound bad, and I don't sense any loss of resolution on really great recordings. So I don't fear they are simply masking things to make everything warmer sounding.
I ended up buying the Benchmark interconnects and cables. Sure it's possible Amir found something on Amazon that works just as well, but the Benchmark cables are really well made, and the last pair of cables I bought on Amazon were defective. The SpeakOn connectors are great, polarity will be perfect every time, just make sure if you buy on Amazon that you get a real Speakon, the knock offs are not reliable.
The real game changer for me was when with the help of Andrew Jones, who designed the TADs, I was able to vertically bi-amp the speakers using a Pass XVR1. Andrew still had all his anechoic measurements on the TADs, and the details of the passive crossovers in the speakers, and we were able to come up with a hybrid bi-amp such that the slopes and phase angles played well with each other. The low pass in the speakers was bypassed such that a pair of AHB2 is directly connected to the 10" woofers with no thing more than a cable in between. The low pass is done by the XVR1, so there is no loss of damping factor and no over saturating of the iron core inductors in the stock low pass. The speakers are also tucked into corners, with RPG damping to avoid mid to higher frequency reflections. An Audiolense convolution filter EQs the bass with 10-15db of attenuation, and a -3db point at 17 hz. So for the most part the woofers are not moving much and are operating in a very linear manner. There is another pair of AHB2s connected to the coax drivers using the internal high pass. Andrew also routed this via the high pass on the XVR1 to increase the slope of the overall high pass. We were lucky that this happened to work well with the low pass on the woofers.
FYI, Andrew Jones's twin brother was one of the engineers that worked on the design of the AHB2s. This is a brilliant design that measures spectacularly and sounds great too. I suspect much of that has to do with the lack of crossover distortion and the fully regulated power supply. As long as the amps are not overdriven, they are essentially performing with near zero distortion from the first watt to the last. And if you clip them, they let you know and turn off. That happened once with a super loud bass passage before we implemented the convolution filter. The bass was likely 10-15 db louder than it should have been.
All in all, I went through three iterations of AHB2s. First was a pair in bridged mono. Then second was two pairs in a passive bi amp. And finally two pairs in the active bi amp. So far it's the most satisfying system I've had. No limits to dynamic range across the frequency spectrum. Totally quiet, the music just pops out of a silent background.
Last edited:
Cool, so what do we use to determine what sounds better to someone else? An MRI machine? May cost a little more than an AP analyzer. Plus those wishing Amir to do the measurements will need to ship themselves with their audio equipment. Doable, you think?
How's about just some mic measurements to start? Or are you telling me there is no way to determine objectively if certain electrical components have audible impact, and we have to rely on people's opinions of how things sound in highly uncontrolled tests? The "science" is extremely poor on this topic.
I mean, the fact that a simple question like "does this $2000 amp sound better than this $200 amp" does not have a clear answer, is an indication that the understanding on this subject is poor.
Last edited:
Newman
Major Contributor
- Joined
- Jan 6, 2017
- Messages
- 3,556
- Likes
- 4,411
Beat me to it and better and more succinct than I would have been. Good job.
@Mike F and hiding behind the above words are some realities of human perception, that are so counter-intuitive that we need some time and convincing to believe that they could be true. Namely, that if you or I connect cable 1 to our hifi, knowing what cable 1 is, and attentively listen to music, then swap to cable 2, knowing what it is too, and listen again, logic says we should be able to conclude that any differences in our sonic perception of the two cables are due to differences in the sound waves reaching our ears, ie we can Trust Our Ears on this matter….BUT, that logic is wrong! And in fact our minds are creating impressions that are not only independent of the sound waves themselves, but overwhelm the sound waves. And just to cap it off and confuse it even more, our minds are assigning these non-sonic impressions to the sound waves, and we are utterly convinced that the sound waves are creating the impression, when they are NOT.
And get this: these non-sonic impressions that I describe above are so strong that they are not just about cables and DACs: they even apply to loudspeakers that are quite easily detectable as sounding very different (in a proper controlled blind study). So strong that they can reverse one’s preference for speaker A over B blind tested (ie based on sound waves alone), to B over A sighted (based on things other than sound waves, but our brains fool us into thinking it was really the sound waves).
And that is why it is so wrong to claim that people into the Science of Audio are into measurements with instruments and not into listening. The science of audio includes learning about what kinds of sound waves we prefer when listening (and no, we are not all unique individuals in terms of preferences, not at all). ASR can teach interested audiophiles a lot about listening, and what sound waves are generally preferred.
It would be really seriously ignorant for a fresh newbie to the science of audio to register here on ASR and launch into finger-pointing lectures filled with name-calling about Objectivists, Measurement Addicts, Lab Coated Cloth-Ears, and Listening to Gear Not Music. Such newbies here, and there are a few, have a lot to learn…if they are genuinely interested. Some are, some are not.
cheers
Acoustic measurements are much less precise (additional distortion and frequency response errors from microphones and speakers) and very noisy (background noise is easily some 10,000 times higher than @amirm's APX) compared to electronic measurements. So you can safely say that distortion and noise below the background noise will not show up in acoustic measurements, but can be shown precisely by the APX, and in fact almost all decent audio ADCs.
Real audibillity and psychoacoustics are not such an uncharted field as you seem to think it is. A lot has been studied and known since almost a century ago, when some of the early audibility experiments were conducted to improve telephone service. Distortion levels, timing differences, frequency deviations were all studied and results published. Those who rely on uncontrolled listening evaluations often hear all kinds of things that are not there. Then, they claim that science can't explain or measure these non-existent differences. No sh*t, how do you measure something that only exists in one's imagination? An MRI machine was my solution
I'm often interested in the question whether a particular component appears to be well engineered - a good thing to know before you buy. I'm never interested in categorical assertions that the engineering of a particular component determines how it sounds. The same component will often sound different in a different configuration of components. We only hear configurations of components. The variables across configurations are too various to measure in the real world.
Newman
Major Contributor
- Joined
- Jan 6, 2017
- Messages
- 3,556
- Likes
- 4,411
Well there is this one...
Message to golden-eared audiophiles posting at ASR for the first time...
Hello friend. Hey, listen...we know how it is. Believe me, most of us have been there too. You've spent years toiling in the muck of audiophilia. You read ALL the reviews. You watched ALL the youtube videos. You visited ALL the other forums where everything always makes a difference. You...
www.audiosciencereview.com
…which is a sticky. But the subsequent 118 pages are hardly succinct…and testament to the difficulty of the message to the intended recipients.
Here are some examples ...
You remember when solid state first came out, it was all about power and THD. It got so bad the FTC had to regulate how power was specified. McIntosh had clinics that would measure your amplifier's noise, THD, power and frequency response. To out spec each other, designers kept piling on more negative feedback. Very few were looking at IMD at the time, or crossover distortion. The power supplies were tiny, and you could even see the lights on your amps dim on loud bass passages. These solid state amps we all thought measured great, but they sounded horrible. A 25w tube amp sounded better than a 100w transistor. The great US companies mostly were sold and the brands were used to sell cheap mass produced stuff. Then slowly we started realizing we needed to measure IMD, look at the power supply, that huge negative feedback caused other problems with stability ... In short, there was a learning curve.
Same thing happened with early CD players. They spec'd much better than LP, cassette or reel to reel in terms of things like noise, wow/flutter, flat frequency response, etc. But early digital sounded harsh, even though the frequency response, noise and channel separation was infinitely better than LP. Then we learned about jitter, the importance of dithering signals, the problems with digital and analogue filters, linearity errors .... It took years before jitter was understood or measured. Same for filters both analogue and digital.
Most of the problems with tubes were at high power levels. But with class transistor B amps the problems were at low power levels with crossover distortion. Same with LPs and CDs. the early DACs had big linearity errors at low levels. The delta sigma DACs cured this, but created other problems. And it seems measurements are best suited to evaluate legacy tech, not the latest tech that solves the old problems but creates new ones.
I'm not saying we cannot measure things that are important, I'm saying it takes time to know every measurement that is important. So the lesson for me was first look at equipment that measures well on the parameters that are known at the time, but then listen.
Last edited:
Newman
Major Contributor
- Joined
- Jan 6, 2017
- Messages
- 3,556
- Likes
- 4,411
Entirely rubbish. Sorry for blunt.
I wasn’t saying tube or ss was better or worse, instead that the specs that were useful judging tube gear missed a lot when ss came out, and yes it is a business, so expect marketing may be unscrupulous, so yes, they will obscure things to make a profit
Also was not referring to early digital recordings, but to early CD players, and yes they sounded like shit but measured wonderfully on the measurements that were prevalent then, remember the Sony Phillips line perfect sound forever
anyway, differences in current electronics are mostly insignificant compared to what happens converting the electric signal to sound at the speakers and then how that is affected by the room, fortunately sota dsp can help a lot, don’t need ABX to hear this, it’s not subtle. I spent years and thousands trying to get my system right by changing the gear, speakers made some difference, electronics almost no difference, room treatments about as much as speakers, dsp made huge difference
Last edited:
I was a young guy working as box boy, sales and car audio installer around the time the first car CD player a Sony CDX-5 was released. About the same story for the Sony CDP-101 the first home audio CD player. I was able to compare the CD players to turntables on many speakers and amps. The CD players absolutely sounded better than a turntable.
Than “a” turntable, sure, but than a Linn LP 12 with a great cartridge set up and a spectral preamp? Probably not. I kept my Linn until I got my Wadia DAC. CD was much more convenient thoughI was a young guy working as box boy, sales and car audio installer around the time the first car CD player a Sony CDX-5 was released. About the same story for the Sony CDP-101 the first home audio CD player. I was able to compare the CD players to turntables on many speakers and amps. The CD players absolutely sounded better than a turntable.
I've had turntables up to ~SL-1200 grade from Technics with a decent MC cartridge and still CD players sounded better. The diminishing returns on a Linn LP12 and the Spectral would make it a no deal for me.
Newman
Major Contributor
- Joined
- Jan 6, 2017
- Messages
- 3,556
- Likes
- 4,411
Yep, that's the entirely rubbish bit. Sorry for blunt again, but you are not picking up on the clues.
You have fallen for a myth.
Here are the top 3/6/10/whatever reasons why people might have reported a perception of harshness (or other negatives) from early CD player output:-
- bias x10. This would cover 9 out of 10 reported instances. Late adopters (even middle adopters) would be biased against the very idea that a first-generation product could possibly demolish a refined, late-stage-of-development turntable that they have been tweaking into superlativeness over a period of years. Their brains won't let the truth in, instead, build a negative perception wall.
- speakers that aren't flat in response, but instead have been 'hand-picked' to 'organically synergise' with their LP output, and that just happens to mean boosted treble from the speakers, or even recessed mids.
- use of 'certain' CDs that have been badly transferred to CD by negligent labels, especially in the popular genres.
Actually all the opposite. Early reviews of CDP 101 were great. People were very impressed with low noise, deep bass, dynamic range. It took a while and then the shortcomings became more obvious. At the time a Linn with a cartridge and tone arm was about 1700. A spectral pre was 1000. But even a humble Rega or Thorens outperformed the CDP 101. It wasn’t until Meridian that CD players became listenable.
Similar threads
