Does Op-amp Rolling Work?

[Buckeye Amps]

Some people can hear the difference, some can't.

When those people can hear a difference in a properly setup and controllered comparison environment, let me know.

There were Best Buy employees for years that said Audio Quest HDMI cables improved picture quality and they could clearly see the difference.
So. Yea. Someone saying they can hear or see a difference in things that have been objectively disproven over and over means nothing to me.

 

[tmtomh]

Sure. To a point.

Take two violins. Both are the same size, using brand new strings so that the only difference is the age. This means differences in the wood and varnish.
They should measure the same. (Or close enough). Yet you may hear the difference. (Professional musicians will. At least those who play string instruments.)

The violin analogy is fatally flawed.

 

[Multicore]

And this also begs another question... if you believe that op amp rolling is a waste.. how do you feel about the statements that speakers and pairing amps to get a better/different sound?
E.g Speaker Brand A paired w Brand B amp versus Brand C or Brand D amp...

Some amps are not compatible with some speakers. But assuming we eliminate those cases, which means that the amps will be operating linearly, and assuming also that none of the amps involved are really bad (i.e. measurably colored or noisy to a degree we know is significant) then I don't think you'll be able to show the effect in controlled experiments.

They way that choosing the exact right amp to pair with a given pair of speakers can make sense is when we imagine that the amp's frequency response compensates for issues in the speaker/room system (which are usually significant). But it's really hard to do this way and much more convenient to use measurements of the overall system and programmable equalizer. (For example, I use Dirac Live.) I mean, even that's hard to do for non-technical people but it's way easier than trying out a whole string of amps and hoping one might help a bit.

 

[oleg87]

If you send a 1K Hz to the opamp with THDN of 0.001 or better and then to the power stage with THD of 0.03 you will get final THD of 0.03 or worse. I also disagree that swapping opamps is pointless. Yes, there is plenty of the IC and discrete opamps that perform worse then NE5532. There are some that perform better. Some people can hear the difference, some can't.

There are better opamps than the 5532 but I suspect aimless swapping (especially with the ludicrous ways I often see discrete op amps wired into the design, dangling off some bundle of hookup wire) is more likely to create an RF oscillator than a more performant circuit

 

[Multicore]

The violin analogy is fatally flawed.

It sure is. Each piece of wood is different. Even the manufacturing of epoxy-composites is subject to tolerances, although I believe they are smaller than with wood and joinery.

 

[im_gumby]

That, in general, is the science of heterophenomenology. Here in our specific audio circles we can call it psycho-acoustics.

First you have to agree and write down a definition of what people mean when they say that X is more "clear" than Y or has a different "soundstage" or whatever. Then set up a test environment and get a panel of listeners and double-blind statistically measure the differences they report. It's the same basic procedure that we use to say that lossy compression codec X is better than Y. It's a well established process because we've been doing it for ages as it's important in telecommunications, sound recording and so on.

However, I feel that the onus is on those who claim that Amir's test methods fail to reveal important audible differences in the signals involved to do the work.

The fact is that we've been in this same situation with all sorts of audiophile claims for decades and seldom do those who insist they can hear something that we cannot measure take up the challenge to demonstrate the phenomenon in a controlled manner.

You bring us back to the point of objective to subjective.
Subjective meaning I know I can hear a difference but I can't quantify it.

He was going fast... but what was meant by fast?

 

[amirm]

But when we hear the audiophiles and even non-audiophiles talk about depth of sound stage, tone, clarity, imagery (sound placement), control, etc ...
How do you measure that? Or rather how do your measurements reflect perceived changes in that?

If any of those happened, it would impact the measurements I make. If "depth" is lower noise floor as is routinely claimed, the SNR/DNR tests would clearly show that. If clarity is improved, it means fundamental signal is different which would translate into THD numbers and certainly FFT spectrum changing. If imaging is changed, then the levels between channels or timing between them is changed. While I don't measure the latter, it is trivially done.

In other words, in the context of what we are investigating here, it is trivial to show that none of these things happened.

We are "lucky" that way in that vast majority of these tweaks have no foundation in reality so measurements confirm the same. If there were real signal changes, we would absolutely see it impact measurements as we make today. We may not be able to explain them in these measurements, but existence of the same would absolutely be there.

 

[DMB90]

You bring us back to the point of objective to subjective.
Subjective meaning I know I can hear a difference but I can't quantify it.

He was going fast... but what was meant by fast?

This may have to do with different slew rates.

 

[morillon]

If any of those happened, it would impact the measurements I make. If "depth" is lower noise floor as is routinely claimed, the SNR/DNR tests would clearly show that. If clarity is improved, it means fundamental signal is different which would translate into THD numbers and certainly FFT spectrum changing. If imaging is changed, then the levels between channels or timing between them is changed. While I don't measure the latter, it is trivially done.

In other words, in the context of what we are investigating here, it is trivial to show that none of these things happened.

We are "lucky" that way in that vast majority of these tweaks have no foundation in reality so measurements confirm the same. If there were real signal changes, we would absolutely see it impact measurements as we make today. We may not be able to explain them in these measurements, but existence of the same would absolutely be there.

ccif?
smpt?
din?
ap32?
nid?
etc
the music is very very very essentially intermodulation, no? ;-)

 

[Multicore]

You bring us back to the point of objective to subjective.
Subjective meaning I know I can hear a difference but I can't quantify it.

He was going fast... but what was meant by fast?

We accommodate that with appropriate explanations and questions to the subject about their subjective experience.

I'm not sure what you had in mind with you example but we could have a question with two possible answers: "Was it fast or slow?" together with an explanation of that, say a video that shows a fast one and a slow one. Then with repeated exposures and questions and multiple test subjects, we can obtain a measurement of that perception.

Heterophenomenology isn't just a cool word. It's how we objectively inquire about people's subjective perception. You can externalize the internal. If you can experience it then you can, with suitable questioning, tell us about it.

 

[DMB90]

There are better opamps than the 5532 but I suspect aimless swapping (especially with the ludicrous ways I often see discrete op amps wired into the design, dangling off some bundle of hookup wire) is more likely to create an RF oscillator than a more performant circuit

I didn't say all discrete opamps are good and I only measured them with Hypex NCX500. I am using audio analyzer so when there is RF oscillation you can see it. In my experience, once you start measuring THDN 0.001 or better you start seeing the importance of clean and well secured binding posts, shorter cables, grounded cables etc.

 

[im_gumby]

Some amps are not compatible with some speakers. But assuming we eliminate those cases, which means that the amps will be operating linearly, and assuming also that none of the amps involved are really bad (i.e. measurably colored or noisy to a degree we know is significant) then I don't think you'll be able to show the effect in controlled experiments.

They way that choosing the exact right amp to pair with a given pair of speakers can make sense is when we imagine that the amp's frequency response compensates for issues in the speaker/room system (which are usually significant). But it's really hard to do this way and much more convenient to use measurements of the overall system and programmable equalizer. (For example, I use Dirac Live.) I mean, even that's hard to do for non-technical people but it's way easier than trying out a whole string of amps and hoping one might help a bit.

Ok...

Lets focus on those cases where the amps are not 'compatible' with those speakers?
How or why? I mean if we look at the amp measurements, any differences are usually below our hearing.
So using your logic... if the differences are in audible, what makes the pairing not compatible?

Look, hopefully you're starting to see the gaps in relying solely on measurements.

 

[DMB90]

Speaking at least in terms of Hypex and Purifi designers....those who use sockets to allow OP changing are doing so almost exclusively to raise their profit margins.

I am sure some will disagree. It is good to have options. Yes, NCX500 OEM opamp sounds and measures very good. But they allow bypassing it for a reason.

 

[Buckeye Amps]

Ok...

Lets focus on those cases where the amps are not 'compatible' with those speakers?
How or why? I mean if we look at the amp measurements, any differences are usually below our hearing.
So using your logic... if the differences are in audible, what makes the pairing not compatible?

Look, hopefully you're starting to see the gaps in relying solely on measurements.

Huh? There are plenty of amps (and DACs) that have very measurable differences that are in the audible range/able to be heard.

No one said there are not products out there that have audible differences that also correspond to measureable differences in the audible range. You seem confused.

 

[DMB90]

When those people can hear a difference in a properly setup and controllered comparison environment, let me know.

There were Best Buy employees for years that said Audio Quest HDMI cables improved picture quality and they could clearly see the difference.
So. Yea. Someone saying they can hear or see a difference in things that have been objectively disproven over and over means nothing to me.

When Hypex was only making NC500 modules which required external buffer which Opamp was your favorite?

 

[Buckeye Amps]

I am sure some will disagree. It is good to have options. Yes, NCX500 OEM opamp sounds and measures very good. But they allow bypassing it for a reason.

They can disagree but it is true.

The bypassing allows for a more broad integration with existing input board setups manufacturers already had in place, so as to not force manufacturers to redesign something from the ground up. I can guarantee Hypex did not put bypass into their module because they felt their design could be audibly improved by a Sparkos or Weiss or Sonic Imagery.

 

[JustJones]

Our AI overlords claim no two violins have the same frequency response.

• Individual Violin Characteristics:
  Each violin, even those seemingly identical, has unique characteristics due to variations in wood type, thickness, age, varnish, and construction techniques. These variations influence how the instrument vibrates and resonates, leading to subtle differences in its frequency response.

 

[Multicore]

Ok...

Lets focus on those cases where the amps are not 'compatible' with those speakers?
How or why? I mean if we look at the amp measurements, any differences are usually below our hearing.

No. I gave my definition of compatibility of the amp with the load: linearity. This is easily determined with technical measurements. For example, some amps can't drive a very low impedance speaker and will misbehave (nonlinearity). Or some aren't powerful enough and clip (nonlinearity).

So using your logic... if the differences are in audible, what makes the pairing not compatible?

That's not using my logic.

 

[im_gumby]

If any of those happened, it would impact the measurements I make. If "depth" is lower noise floor as is routinely claimed, the SNR/DNR tests would clearly show that. If clarity is improved, it means fundamental signal is different which would translate into THD numbers and certainly FFT spectrum changing. If imaging is changed, then the levels between channels or timing between them is changed. While I don't measure the latter, it is trivially done.

In other words, in the context of what we are investigating here, it is trivial to show that none of these things happened.

We are "lucky" that way in that vast majority of these tweaks have no foundation in reality so measurements confirm the same. If there were real signal changes, we would absolutely see it impact measurements as we make today. We may not be able to explain them in these measurements, but existence of the same would absolutely be there.

There are a couple of issues raised...

First, consider that what you measure may not capture the differences in these op amps. Or rather the perceived sound differences and people's preferences.
( I mean some prefer ESS Dac chips over Rohm or a different brand. )

Looking at your noise floor... I've seen reviewers talk about having to switch to more sensitive headphones to see if they can hear the noise floor and how 'black' it is, while if it were me with my cheaper headphones, I couldn't hear the difference.
Along this issue... I know that I've listen to some YT videos where they playback different units playing the same song... and I couldn't hear a difference. I switch to my IEMs which are better quality... and I can hear the subtle changes. (I went back to my cheaper headphones and couldn't hear the differences, so it wasn't psychosomatic )

Then looking at the 'incompatibility' between some amps and speakers.
How would explain that because your measurements would most likely not capture that. Would you argue that the incompatibility is subjective and not a measurable thing?

And then finally the issue of distortion. I seem to recall several 'experts' talking about the perception of distortion and that its not necessarily a bad thing.


Note: Again, I'm not suggesting that there is anything wrong with your measurements. I for one, thank you for taking the time to make them and do your reviews.
I'm just trying to reconcile this issue when its more than a handful of people from all walks of life that claim to hear a difference.
I have not tried it, so I can't say one way or another.

I am just a bit skeptical of both sides and believe that the truth is somewhere in the middle.

 

[im_gumby]

No. I gave my definition of compatibility of the amp with the load: linearity. This is easily determined with technical measurements. For example, some amps can't drive a very low impedance speaker and will misbehave (nonlinearity). Or some aren't powerful enough and clip (nonlinearity).


That's not using my logic.

To be clear, we're not talking about the impedance issue. It was the subjective ... I thought these speakers sounded better when played using Amp A versus Amp B.

Or that the type of music you listen to will sound better or worse on these speakers, IEMs, amps, whatever.

And I do have to apologize this is getting into the weeds where we are talking about all of the subjective comments that get made regarding products.