• Welcome to ASR. 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!

Amplifiers can you hear a diffrence?

Good morning everyone. At least this side of the pont. After reading and reading and asking recommendations, and again reading all technical answers and questions and answers on answers from very respected people and become in doubt again ... cause, when I am almost sure there is no audible difference, someone respected, highly manufacturer comes with explanation why there could be may be a difference.

"Could." A plausibility argument is different than solid evidence. If I might beg a favor from @March Audio , could you make two recordings, one from each amp type (taken electrically, of course, maybe with a speaker load) and post them? That way we can do a solid ABX experiment.
 
"Could." A plausibility argument is different than solid evidence. If I might beg a favor from @March Audio , could you make two recordings, one from each amp type (taken electrically, of course, maybe with a speaker load) and post them? That way we can do a solid ABX experiment.
Sure I can do that.

However you have to bear in mind this isn't just "plausible", it is backed by research and I don't think it's contentious.

Toole
The simplest deviation from flat is probably a spectral tilt. There is evidence that we can detect slopes of about 0.1 to 0.2 dB/ octave

I have tried to find the research that Toole refers to. It's from a Japanese acoustical society paper which I have a relevant extract. Unfortunately my Japanese isn't so good ;), but the tables give us relevant information.

Response profile D

Screenshot_20200430_222526_com.google.android.apps.docs.jpg



Table 2

Screenshot_20200430_222550_com.google.android.apps.docs.jpg


I can hear the filtered section of the file posted above quite easily. Have you given it a listen?

There may of course be other factors that cause the perceived difference. Whilst good class d is low output impedance and has minimal effect from speaker load, it's not entirely impervious. Also Bruno talks about the effect of the output inductor which the new design has taken mitigations for.

As an aside from this, with dac filters I don't hear differences, except with ones that roll off early into the audible band.
 
Last edited:
Sure I can do that.

However you have to bear in mind this isn't just "plausible", it is backed by research and I don't think it's contentious.

Toole
The simplest deviation from flat is probably a spectral tilt. There is evidence that we can detect slopes of about 0.1 to 0.2 dB/ octave

I have tried to find the research that Toole refers to. It's from a Japanese acoustical society paper which I have a relevant extract. Unfortunately my Japanese isn't so good ;), but the tables give us relevant information.

Response profile D

View attachment 61131


Table 2

View attachment 61132

I can hear the filtered section of the file posted above quite easily. Have you given it a listen?

There may of course be other factors that cause the perceived difference. Whilst good class d is low output impedance and has minimal effect from speaker load, it's not entirely impervious.

As an aside from this, with dac filters I don't hear differences, except with ones that roll off early into the audible band.

What I want to try to validate is the specific claim of audibility differences between those specific amplifiers. I’d do it myself except that I only have the Purifi on hand. So I’m enacting your labor.
 
I think another way to look at these tilts is in loudness levels perceived. If you had a continuous tilt of .2 db/octave from 20 hz to 20 khz whether up or down, if you use pink noise at an RMS level of -25 db, then the tilt will be more than .4 db different in overall level if the inflection point is at 1 khz. The idea being average music levels are something like pink noise and run around -20 or -25 db from max level. So they'll be perceptible as different on that basis alone. As well as it changing the perceived timbre just a bit. I'd think the change in timbre would only increase how noticeable it is vs a simple level difference.

If the tilt is all on the low frequency side of 1 khz then the difference is large. If it is only on the high frequency side of 1 khz it is lower though still just over .1 db loudness difference.
 
??? In human Language. Audible yes or no?
 
Dear Alan, thx for your answer, just I hoped the answer would be different, i respect you a lot. I ve been reading your conversation with Alan Shaw from Harbeth.
This means indeed that people have to go and buy, or try to listen to lots of different amplifiers at home, with their music with their speakers cables Dac etc. Which consequence does this have for this forum?
 
Dear Alan, thx for your answer, just I hoped the answer would be different, i respect you a lot. I ve been reading your conversation with Alan Shaw from Harbeth.
This means indeed that people have to go and buy, or try to listen to lots of different amplifiers at home, with their music with their speakers cables Dac etc. Which consequence does this have for this forum?
I don't think this extends to Dac and cable. What it means is we need more testing of amps with speaker loads. Nor would it mean listening to lots of gear at their home.

With room eq being increasingly common most difference could be fixed there if we have data from testing with the correct speaker loads.

The consequences for the forum are to develop and use those synthetic speaker loads. None of this is outside a normal rational approach to good sounds.
 
I don't think this extends to Dac and cable. What it means is we need more testing of amps with speaker loads. Nor would it mean listening to lots of gear at their home.

With room eq being increasingly common most difference could be fixed there if we have data from testing with the correct speaker loads.

The consequences for the forum are to develop and use those synthetic speaker loads. None of this is outside a normal rational approach to good sounds.

I really need to get my act together and participate in some amplifier testing with Swedish AES (am member) using this device ...

1588323961531.jpeg
 
For anyone not familiar with the "Richard Clark $10,000 Amplifier Challenge" I highly recommend reading about it. While generally one must be skeptical of big claims, in this case it has the "ring of truth."

https://www.stevemeadedesigns.com/board/topic/193850-richard-clark-10000-amplifier-challenge/

His challenge demonstrates that competently designed amps, in good condition, operating under clipping, are sonically indistinguishable. There are some important conditions for his test. One of them is that the frequency response has to be matched. So if the two amps have a different frequency response, an EQ is used to make them identical. The other is that the levels must be carefully matched.

"Ah hah" you might think! Amps do sound different. Yes they do. But the context is important here.

The testee/listener is allowed to choose which amp the EQ is put on.

They are also allowed to bring any amps within the mentioned criteria to test, and any cone based speakers to take the test.

The frequency response thing is interesting, because according to "golden ears" thinking, a mere EQ shouldn't be sufficient to remedy the "flaws" of an inferior amp in comparison to a mega-buck "audiophile" amp.

What the relevance is, to me, that a well designed amp with a flat frequency response and sufficient power will do the job, and I don't have to worry about some minute differences.

The problem with placebo effects and sighted listening is profound. In the "real world" type of comparisons that any of us do, trying this amp, then that amp, the ability to make a discerning perception of any differences in sound is rendered impossible simply by the difficulty in matching the playback level precisely.

The audio "placebo" effect is a real phenomena, it really makes things sound different.

The bold claims of "golden ears" are outside the level of this type of minute differences, and I feel confident in saying that they are largely the effect of subjective bias.

Richard Clark's $10k prize was never claimed, and he ran his test hundreds of times.

Another famous test along these lines was the Bob "Carver Amp Challenge" hosted by Stereophile. Here's their writeup.

https://www.stereophile.com/content/carver-challenge

In this test, Carver matched the sound of a much more expensive "audiophile" amp using null testing, basically adjust parts of his circuit to null the output of his cheap amp against the fancy amp. In the Stereophile description, they describe him actually having to add distortion into his circuit, which implies that the fancy amp must not have been "accurate" in the sense we talk about here. I'm not sure what to make of that.
 
I really need to get my act together and participate in some amplifier testing with Swedish AES (am member) using this device ...

View attachment 61287
Well now. Since you are a member, can you tell us more about the testing they did. Like is there a list of amps they've tested over the years? Do they still do this testing currently?
 
Well now. Since you are a member, can you tell us more about the testing they did. Like is there a list of amps they've tested over the years? Do they still do this testing currently?

Sorry, purely passive member, only read their regular papermagazine. Website does not contain much useful info. Only remedy is if I change my status to active, maybe after summer/corona.
 
Well, that fact certainly speakes for itself.

There was a stretch of time on this site where a $$$ challenge was brought out to the more...ummm...insistent contributors. It tended to go pretty badly, so I haven't seen a lot of it lately. Tends to just make people angry.

I wouldn't be against a 'crowdfunded' money pool...somewhere to refer people who are all talk and no controls willing to put their money where their mouths are...but probably a bit too in your face to be considered, and honestly a bit tricky to actually do for anyone interested, but still...
 
Just out of curiosity, I have my amp now, in human language, is it fair to say, inaudible is maybe a big word, but let say the differences between amplifiers, maybe DACs, streamers if there are differences, and for people with very good hearing, the differences will be minimal?
 
Just out of curiosity, I have my amp now, in human language, is it fair to say, inaudible is maybe a big word, but let say the differences between amplifiers, maybe DACs, streamers if there are differences, and for people with very good hearing, the differences will be minimal?

I think that is fair to say.

Anecdotally I am aware of listening tests comparing DACs (from pro audio world) and when done carefully, people usually fail to discern any audible difference between converters.

I participated in an online test involving comparing an original audio file to a “loopback” version, where it was run out of a DAC and back into a ADC. The resulting files were made available for people to test using ABX software.

Listening in the studio I was working in at the time, which did not have great monitoring, I completely failed to discern original audio from re-recorded. That means the converter was “transparent” at that level.

There was an individual though who was consistently able to distinguish the files. He was obviously a very skilled listener, but one comment he made was that he was in a mastering studio with $40k monitors, and he thought that was a big factor. (The monitoring environment in a good mastering studio is generally superior to a regular recording studio.)

I was blown away by this result.

For anyone that’s never tried ABX testing, it’s much harder than you might think. Every playback sounds different, even with exactly the same audio. Our perception of audio is affected by both conscious and unconscious processes.

However, the bold claims many make about audio gear X sounding way better than gear Y should be easily discernible in ABX testing.

One criticism of ABX testing is that some people claim there are subtle differences in audio playback quality which are hard to perceive in the immediate comparison, but become apparent over extended (days or weeks) listening periods. This has been demonstrated to be false.

(ABX testing works like this: there are two sources A and B. The subject has a switch, and they can select A or B to listen to, as often as they like. Once they feel comfortable with being able to, hypothetically, identify the sound characteristics of the two audio sources, they press a switch for X and are presented with a randomly selected A or B source. They challenge is for them to identify which it is A or B. A result around 50% is a fail. Usually the test is run multiple times to get a lot of results, eliminating apparent successful results based on chance alone. This was the method Richard Clark used in his amp challenge, he built custom hardware to run it. There are software based versions that are easier to run if you have recorded material to compare.

For the test to be relevant it is paramount that the levels between the two sources be matched as closely as possible. It is trivial to discern the difference between signals at different volumes.)

One interesting aspect of ABX testing is that it does not involve preference testing at all. The listener does not need to choose which source sounds “better”. It is simply a test of audible difference. Failing an ABX test renders claims of audio superiority moot.)
 
Last edited:
Dear Alan, thx for your answer, just I hoped the answer would be different, i respect you a lot. I ve been reading your conversation with Alan Shaw from Harbeth.
This means indeed that people have to go and buy, or try to listen to lots of different amplifiers at home, with their music with their speakers cables Dac etc. Which consequence does this have for this forum?
No it doesn't mean this. :). Having detectable differences doesn't necessarily mean a difference is significant. We need to put this into context.

Whilst anyone who reads this forum regularly knows my views on people's ability to hear differences, or the accuracy of their conclusions (lack of) , it doesn't mean I am of the view that everything sounds the same. I have said this in the past.

However where I do hear differences in equipment it is often so subtle it's of absolutely no consequence. Usually there is a technically attributable reason to explain it.

Not all amplifiers sound the same. The extreme example of this are tube amps where their high output impedance creates an interaction with the speaker load causing variable frequency response. All easily technically explainable. Any difference in solid state amps is generally much much smaller, often to the extent of being difficult or impossible to detect. What this means in practical terms is that a difference is so small often its not worth bothering about, it does in practical terms "sound the same".

The message of this forum isn't "everything sounds the same", so I don't think it has any consequence for it.

The forum provides information about products. We know generally what is technically state of the art for amps and dacs at the moment. You can compare performance and value of other products against that reference. What differentiates this forum is that it doesn't get involved with the totally unreliable subjective only views on product performance proclaiming big differences that just don't exist. It does show that many components technical performance is so similar that differences in sound will be difficult or impossible to detect.

So in the context of the Purifi amp we were discussing, if I had a choice of it or a Hypex, I would choose the Purifi. I know its technically slightly better and when concentrating hard I can hear a subtle difference. However would I be bothered about having a Hypex instead? Absolutely not, it's just fine and in most practical definitions the difference isn't significant.
 
Last edited:
The forum provides information about products. We know generally what is technically state of the art for amps and dacs at the moment. You can compare performance and value of other products against that reference. What differentiates this forum is that it doesn't get involved with the totally unreliable subjective only views on product performance proclaiming big differences that just don't exist. It does show that many components technical performance is so similar that differences in sound will be difficult or impossible to detect.

This is true, but there is some irony involved with the reviews of this site, as they can have a tendency to have their own biasing influence on peoples perception of equipment. I've seen people obsessing in here about small differences in performance in converters that I am certain are not distinguishable in listening tests. This is an inversion of the traditional idea behind "objective evaluation" of audio gear.

The origins of the "subjectivist vs objectivist" debate are the result of claims by subjectivists that some, usually very expensive, piece of audio gear sounds amazing and much better than the cheaper competition. Even when measurements show their is not a significant difference, the "golden ears" audiophile will claim that there are aspects of sound reproduction (the electrical versions of the signal) that cannot be measured!

I've been humbled many times in audio, so I am wary of absolutes, but I am confident that this is nonsense.

In addition to the measurements, careful AB or ABX testing has proved that "Golden Ears" cannot distinguish the gear they showered with superlatives in sighted listening. (This then leads to the argument that the kind of sonic differences they are perceiving are not well detectable in AB/ABX type testing. This is a weird claim.)

Amazingly, this argument still rages.

The traditional objectivist view is that anything perceptible in audio reproduction is measurable. A practical corollary is that once a piece of gear reaches a certain level of performance, no audible improvements can be made. This means the consumer can shop with confidence, for an amp lets say, based on criteria like: power, price, visual appearance, features, ergonomic aspects, warranty, reputation and the like, as long as the amp is from a reputable company and performance hits a certain objective point.

On ASR the speaker reviews are in a different category than the measurements of electrical performance. The research these tests are based on are grounded in subjective preference testing of speakers. Without this foundation, the measurements have little value. It is not axiomatic that accurate reproduction of sound is always the best.

This holds because speakers are easily distinguished, even in blind tests.

The reviews of other gear on ASR have no corresponding foundation in careful evaluation in listening tests.

I think most people here would agree that the audible performance of gear that measures differently is often not perceivable in listening tests.

This puts the drive for the "best performing" amp into a weird category. It reminds me of the drive to own a fast sports car, even though the driver never has the opportunity to test the limits.

I find this to be strange, but I appreciate the impulse to find the absolute best performance.

One thing it does allow the listener is, at the least, rule out that particular piece of equipment if they are unsatisfied with the sound they are experiencing with their system.

A fascinating test would be to see at what point on the ASR master DAC list an audible difference could be detected in a reasonably well done listening test. Presumably some of the worst performing converters would be noticeable, but my guess is that it would get pretty hard even down in the 3rd section of "mediocre" devices, like the phone dongles.
 
This is true, but there is some irony involved with the reviews of this site, as they can have a tendency to have their own biasing influence on peoples perception of equipment. I've seen people obsessing in here about small differences in performance in converters that I am certain are not distinguishable in listening tests. This is an inversion of the traditional idea behind "objective evaluation" of audio gear.

The origins of the "subjectivist vs objectivist" debate are the result of claims by subjectivists that some, usually very expensive, piece of audio gear sounds amazing and much better than the cheaper competition. Even when measurements show their is not a significant difference, the "golden ears" audiophile will claim that there are aspects of sound reproduction (the electrical versions of the signal) that cannot be measured!

I've been humbled many times in audio, so I am wary of absolutes, but I am confident that this is nonsense.

In addition to the measurements, careful AB or ABX testing has proved that "Golden Ears" cannot distinguish the gear they showered with superlatives in sighted listening. (This then leads to the argument that the kind of sonic differences they are perceiving are not well detectable in AB/ABX type testing. This is a weird claim.)

Amazingly, this argument still rages.

The traditional objectivist view is that anything perceptible in audio reproduction is measurable. A practical corollary is that once a piece of gear reaches a certain level of performance, no audible improvements can be made. This means the consumer can shop with confidence, for an amp lets say, based on criteria like: power, price, visual appearance, features, ergonomic aspects, warranty, reputation and the like, as long as the amp is from a reputable company and performance hits a certain objective point.

On ASR the speaker reviews are in a different category than the measurements of electrical performance. The research these tests are based on are grounded in subjective preference testing of speakers. Without this foundation, the measurements have little value. It is not axiomatic that accurate reproduction of sound is always the best.

This holds because speakers are easily distinguished, even in blind tests.

The reviews of other gear on ASR have no corresponding foundation in careful evaluation in listening tests.

I think most people here would agree that the audible performance of gear that measures differently is often not perceivable in listening tests.

This puts the drive for the "best performing" amp into a weird category. It reminds me of the drive to own a fast sports car, even though the driver never has the opportunity to test the limits.

I find this to be strange, but I appreciate the impulse to find the absolute best performance.

One thing it does allow the listener is, at the least, rule out that particular piece of equipment if they are unsatisfied with the sound they are experiencing with their system.

A fascinating test would be to see at what point on the ASR master DAC list an audible difference could be detected in a reasonably well done listening test. Presumably some of the worst performing converters would be noticeable, but my guess is that it would get pretty hard even down in the 3rd section of "mediocre" devices, like the phone dongles.
Quite. Objective extremism is no more helpful than subjective extremism.

Re speaker tests, isn't it interesting that whilst our current view on "what's best" (Toole) is based on subjective evaluation, it still correlates with intuitively technically good performance. Ie smoothest flattest anechoic response with smooth off axis response.
 
On the issues of amplifier sound and amplifier measurement, this Q&A interview with Nelson Pass from a year ago is pretty interesting.
 
Back
Top Bottom