Amplifier Op-amp Rolling Part 2 [Video]

[rdenney]

...
Distortion has to be surprisingly high to be audible. The only time I've encountered that is from a tiny market niche, certain SET amplifiers. And even then, it only manifested at higher volumes where the amp is clearly struggling or from frequency response errors which aren't intrinsic to the amp but a result of the loudspeaker's impedance interacting with a high source impedance.

"Voicing" does NOT happen (or has never been demonstrated to happen) from the choice of capacitors or resistors or wire or whatever, which is usually where these evidence-free claims originate.

On the second quoted point, I think Bob Carver was voicing his amplifier in that hotel room when he was responding to Stereophile's amplifier challenge. It might have been a mix of distortion and frequency response. A lot of second-order harmonic distortion can change the timbre of instruments, it seems to me. The upper harmonics of tuba sound, for example, are noticeable in their absence but are maybe 15 dB down from the loudest harmonics and 25 dB down from the combined loudness of the signal. Tuba sound isn't as fundamental-rich as some instruments, but most people think it to be. A lot of second-order harmonic distortion--maybe 10%--would be easily noticeable. I don't know how much it would take to be subtly detectable in a blind test of a substantial population. For me, I think I'd need probably 2-5% harmonic distortion to make a difference I could reliably hear, if the recorded sound is only tuba.

Of course, that is a LOT of low-order harmonic distortion in an amplifier--far more than anyone would do on purpose even if they were trying to achieve something worthwhile. (I make an exception for distortion in instrument amplifiers, which is done to an extreme level for artistic effect. We are talking about amplification used for playback.)

But for the instruments in the bass region, it's not particularly a lot of distortion for even good speakers. 10% distortion below 100 Hz is just not uncommon when speakers are pushed. I have trialed speakers at stores where some, when pushed, made tubas sound in my ears like euphoniums (which are pitched an octave higher).

We should point out that a harmonic distortion product that is 40 dB down from the fundamental (1%) might only be at an SPL of 35 or 40 dB in the room, even when the overall music is fairly loudly amplified. Even if that product was the only signal being played, it would be hard to hear. But 40 dB down out of a preamp might become noticeably loud when amplified, though the masking effect of the (very) loud fundamental would still make it hard to hear. The farther we are in the gain structure from the speakers, the more distortion products may be amplified.

I would also like to make the distinction between harmonic distortion and the enharmonic mess that results from clipping. Clipping may produce loud distortion products all the way into the peaky parts of the Fletcher-Munson curve even for bass instruments. This can make a tuba sound like a bass trombone, or a French horn sound like a trombone. A given percentage of distortion resulting from clipping seems to me more deleterious than from harmonic distortion, because of which harmonics are most emphasized. Clipping is brightly colored distortion, whereas low-order harmonic distortion may fit in the overall sound a lot more pleasantly (and therefore unnoticeably).

None of this changes the conclusions that Amir came do: Changing op-amps doesn't cause enough difference in distortion or frequency response to be anything like reliably detectable, let alone desirable.

Contrast this with the reliably detectable changes in frequency response of a dB or two when broadbanded (as with a spectral tilt). Changing the frequency response is what we do with tone controls and equalizers, of course. That SET amp in your first point might have been clipping (a lot of SET amps put out single digits of watts before clipping, it seems to me), and it might have become non-linear in terms of frequency response due to being unable to drive the speakers. But even that is a form of clipping.

Rick "stops chasing distortion at about -80 dB" Denney

 

[SIY]

It might have been a mix of distortion and frequency response.

Note that they never tested whether the amps sounded different in the first place.

A lot of second-order harmonic distortion can change the timbre of instruments, it seems to me.

About two orders of magnitude higher than the CJ amp they used. So... unlikely in the extreme.

For me, I think I'd need probably 2-5% harmonic distortion to make a difference I could reliably hear, if the recorded sound is only tuba.

For 2nd, maybe higher. I knew that it's not very audible, but I was still surprised when I used Distort of how much it took before I could hear it.

But even that is a form of clipping.

That's why I caveat "within their power limits."

stops chasing distortion at about -80 dB

From a sonic standpoint, I don't worry below about -60dB. From an esthetic and pride-in-my-work standpoint, -80dB is a more typical thing for me, too.

 

[solderdude]

Plus ... most forget that not only Harmonic distortion is added but also the more audible IM distortion. That may well have been more 'sound degrading' than the harmonics and the harmonics of the harmonics that were already present.

But yes... it would have to be bad to become audible. For this reason I have severe doubts if people indeed 'prefer' sound where only H2 in levels that would be masked anyway would 'improve' the sound.

 

[DrZingo]

and of coarse he gets to use the familiar words like 'tighter bass', better soundstage, smoother presentation, more 3D and whatnot. People all want that and as the guy is clearly sponsored by Sparkos he will say stuff like that and his followers will eat those words and rush to Sparkos to 'upgrade' their cheap amps.
They too (like Danny) will hear the vast improvements because .. well... placebo works like a charm.

The output signal of the amp provably does not change but of course the sound does because ... well ... measurements say nothing about the sound.

The sad part is for every review/video with objective measurements showing actual performance (differences) one can find 10x (or more) YT videos and 'reviews' spouting nonsense and 'this sounds better' stuff.

I would say ... proper info is out there but hard to find and 'dry' ... people want 'better sound' and flashy videos with 'flowery talk' and nonsensical 'semi-technical' explanations confirming what they hope confirms what they think (actually t... know) in their heart to be true.

Yes. But I don't think it's only about misinformation, I also think it's about psychology. People want to believe, because they want to interact with their equipment. They don't want it to just sit there, they want it to be a project, to be more "alive". That's what is behind all the cable swapping nonsense as well.

 

[rdenney]

Yes. But I don't think it's only about misinformation, I also think it's about psychology. People want to believe, because they want to interact with their equipment. They don't want it to just sit there, they want it to be a project, to be more "alive". That's what is behind all the cable swapping nonsense as well.

Yup. They are into equipment before music.

There is actually nothing at all wrong with that. I trade equipment out a lot, but I have zero expectation that doing so makes a difference I could detect. It's just fun to play with different stuff.

And the problem isn't when they delude themselves into thinking it sounds better when it can't possibly sound different. The problem is when they believe themselves qualified to delude others.

It's almost like an ignorance cult, or at least hostile to the notion that technical people know how to do things that non-technical people don't know how to do. This isn't revenge of the nerds, this is normies trying to create a smarties tribe without actually learning and understanding the technical stuff. This forum is tribal, too, but at least has the advantage of knowing the technical stuff.

Rick "amazed at the reverence the cult has for some early equipment without realizing that it has some downright primitive op-amps" Denney

 

[solderdude]

People want to believe

Yes, this is true for just about anything in life though.
The majority of the audiophile (and regular audio) world depends on this aspect.
The only thing that 'helps' is education and the willingness to be educated and do well performed comparisons.

 

[DualTriode]

Generally. And that is giving the device a sound that deviates from neutral in a specifically designed manner.

Distortion has to be surprisingly high to be audible. The only time I've encountered that is from a tiny market niche, certain SET amplifiers. And even then, it only manifested at higher volumes where the amp is clearly struggling or from frequency response errors which aren't intrinsic to the amp but a result of the loudspeaker's impedance interacting with a high source impedance.

"Voicing" does NOT happen (or has never been demonstrated to happen) from the choice of capacitors or resistors or wire or whatever, which is usually where these evidence-free claims originate.

First off, I am not buying lunch for a room full of blind people to do listening tests. The large corporate names do just that, Think the development of the Harman Curve as an example.

SIY, I am 90% in agreement with you, "Distortion has to be surprisingly high to be audible." Earl Gettes PhD published experiential results that show low level Harmonics are not audible until they are well into double digit percentages. It seems funny as in odd that we cling to high SINAD as being important. SINAD is just a tool.

@amirm, I am 70% in disagreement with you. Yes people do use tuning the type and level of distortion in an attempt to voice their electronics. You claim that it is just bad design yet they have many followers including the DIY community and even many commercial customers.

I must be a little confused or not. I spend dollars on test equipment yet I still do DIY with amplifiers that measure at more than a little distortion. It is just for fun. The highest SINAD does not matter so much. Can you hear the difference between 120 dB's SINAD and 100dB's SINAD?

Also sitting on my bench is a Benchmark Pre / Headphone Amplifier and a pair of Stealth headphones. I like the low noise, remote control and Display.

If it was just high SINAD it would just be $100 class D amplifiers on ebay.

Thanks DT

I have a couple of repaired, by me, NAD and Crown amplifiers that measure better than 100 dB's SINAD into 8Ohms.

 

[PristineSound]

I am wondering, are there perceptible distinctions above highly measuring noise & distortion measurement thresholds (i.e., 95+ sinad measurements) for amplifiers and DACs, or are these truly inaudible and the comments of "listening perspective" component reviewers like "clinical", "warmth", "sound stage width & depth", "instrument placement" ...... a function of the source recording material, not the component? Are comments such as these about the character of a components sound due to the addition of color or distortion to the audio signal output by the component designer who "voices" the amplifier or DAC? It makes me wonder what paying thousands of dollars buys other than reliability perhaps versus much lower cost well measuring components, or are there audible differences that are not captured by the measurements you make? Thanks for your insight!

People hear a difference because they want to hear a difference. If I have a $1000 amplifier and then I upgrade to a $10k amplifier. My brain is going to look for a difference despite if both amplifiers measures identical or close to.

You can do the experiment yourself, just make sure they are level matched, you can use a multimeter to measure voltage of a tone to level match. Assuming both amps are load and frequency invariant and the distortion and noise profile aren't outrageous.

 

[amirm]

@amirm, I am 70% in disagreement with you. Yes people do use tuning the type and level of distortion in an attempt to voice their electronics.

They absolutely don't. Proof is that they never make that claim, nor show any measurements to prove the same. All they do is cater to nonsense from audiophiles such as SET, no feedback, high output impedance, etc.

Their claim of "voicing" is always related to component choices, or some circuit design for which, they have not verified a thing as far as voicing. This phrase has become a marketing claim to win over people who don't know anything about electronics. It gives an empty explanation to why "this device sounds different."

The only voicing I know of in electronics that had some merit was Carver's Digital Lens. He took L-R in the signal and amplified it and merged it back in. This sharply changed staging. From what I recall though, it also magnified buzzing so it never took off.

 

[amirm]

I must be a little confused or not. I spend dollars on test equipment yet I still do DIY with amplifiers that measure at more than a little distortion. It is just for fun. The highest SINAD does not matter so much. Can you hear the difference between 120 dB's SINAD and 100dB's SINAD?

This is off-topic. But yes, you can hear the noise floor difference if that is how you get from 100 dB to 120 dB. It is all matter of amplification and gain. You can see this in my headphone amp tests with 50 mv output. You best have an ultra quiet amplifier or you will hear that noise with sensitive IEMs. Using equalization robs you of headroom, making it more important to have a higher SINAD amplifier.

 

[daniboun]

He has an ideal "job." You can say anything you want and there will be a number of people who believe you. No need to prove anything either.

This is how urban legends are born)

 

[daniboun]

People hear a difference because they want to hear a difference. If I have a $1000 amplifier and then I upgrade to a $10k amplifier. My brain is going to look for a difference despite if both amplifiers measures identical or close to.

Indeed, it is a form of cognitive dissonance supported by confirmation bias....

 

[DualTriode]

Plus ... most forget that not only Harmonic distortion is added but also the more audible IM distortion. That may well have been more 'sound degrading' than the harmonics and the harmonics of the harmonics that were already present.

But yes... it would have to be bad to become audible. For this reason I have severe doubts if people indeed 'prefer' sound where only H2 in levels that would be masked anyway would 'improve' the sound.

After seeing the measurements it is evident to me that the electronics, including the Op-Amps, are innocent. That does not mean that those that roll op-amps should be shunned. They will learn or they won't, that is fine with me.

If you want to measure and hear real differences look at the speakers.

In an effort not to offend anyone's darling driver I made some measurements on a Scanspeak P17WJ-00-08. For demonstration purposes the input voltage to the driver was 2.83 VAC. The driver was installed in a 0.55 Ft enclosure.



The FFT Plot is a two-tone test 50 Hz and 425 Hz. Note that you can see HD's of the test tones plus IMD side-bands on either side of the 425 Hz test tone.

The IMD side bands are very much audible, particularly as the SPL is turned up. What you hear of IMD is amplitude modulation to your ears. Kind of fuzzy and irritating to me.

 

[rdenney]

First off, I am not buying lunch for a room full of blind people to do listening tests. The large corporate names do just that, Think the development of the Harman Curve as an example.

SIY, I am 90% in agreement with you, "Distortion has to be surprisingly high to be audible." Earl Gettes PhD published experiential results that show low level Harmonics are not audible until they are well into double digit percentages. It seems funny as in odd that we cling to high SINAD as being important. SINAD is just a tool.

@amirm, I am 70% in disagreement with you. Yes people do use tuning the type and level of distortion in an attempt to voice their electronics. You claim that it is just bad design yet they have many followers including the DIY community and even many commercial customers.

I must be a little confused or not. I spend dollars on test equipment yet I still do DIY with amplifiers that measure at more than a little distortion. It is just for fun. The highest SINAD does not matter so much. Can you hear the difference between 120 dB's SINAD and 100dB's SINAD?

Also sitting on my bench is a Benchmark Pre / Headphone Amplifier and a pair of Stealth headphones. I like the low noise, remote control and Display.

If it was just high SINAD it would just be $100 class D amplifiers on ebay.

Thanks DT

I have a couple of repaired, by me, NAD and Crown amplifiers that measure better than 100 dB's SINAD into 8Ohms.

Hmmm. If “tuning” results in a change in SINAD for a speaker amp from 120 to 100, which we agree is an inaudible difference, then what is the point of tuning?

If I was building an speaker amp for myself, I’d be ecstatic if the SINAD was as 100, which would probably be as good as I could measure anyway. I’d be very happy if it was 80, which is probably also inaudible. I might enjoy a project that achieved 40, but I don’t think I’d claim to the world that it’s “tuned to sound good” versus “the best I can do with this design I ripped off from an old tube-type console amplifier, but that I’m happy makes music-like sounds at all.”

Of course, with op-amps, which work before amplification above line level, whatever is wrong with them might get amplified by 20 or 30 dB. That might make a difference in the quiet bits, if noise is the limiting defect, and depending on the nature of the noise.

The question is what satisfies us versus what we think can be recommended to others.

Rick “doesn’t want to hear the amp” Denney

 

[PristineSound]

$80 spent in the name of scientific truth and to dispell fairy tales. . .and fairytale story tellers

 

[375HP2482]

I'm still looking for a solid example of "voicing" in home audio electronics. Over twenty years now since I started asking and still not a one. Just handwaves and anecdotes. So I think it's safe to give up that belief.

Bob Carver. Drat! RDenney beat me to it.

 

[SIY]

Bob Carver.

Except he (Carver) likely didn't.

 

[DualTriode]

In a previous post I posted a link.

Here is a more direct illustration of people tuning the distortion profile to change the sound.

Yes it is audible, especially as you turn up the gain. We are not speaking of flat top clipping, just honest stair stepping harmonic distortion.

This thing is called H2 generator. Looking at the distortion profile on the analyzer it is H2, H3, H4 and stair stepping in the harmonic sequence.

This is an educational toy to me.

https://www.firstwatt.com/wp-content/uploads/2024/01/art_h2_v1.pdf

Thanks DT

 

[SIY]

Here is a more direct illustration of people tuning the distortion profile to change the sound.

Exactly zero actual listening tests. Why would you say something that blatantly false?

 

[DualTriode]

Exactly zero actual listening tests. Why would you say something that blatantly false?

As I have said already a couple of times; I am not buying lunch for a room of blind listeners. Plus I do this for fun.

I have built and used the "H2 HARMONIC GENERATOR". I have measured the output and can hear the difference particularly with the gain turned up. That was only a half blind test, with one eye closed.

"This is an educational toy to me."

While we are at it, ASR has also done "Exactly zero actual listening tests".

Tell us more about "blatantly false".