• 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!

Does Op-amp Rolling Work?

Rate this article on opamp rolling:

  • 1. Terrible. Didn't learn anything

    Votes: 9 3.4%
  • 2. Kind of useful but I am still not convinced

    Votes: 18 6.9%
  • 3. I learned some and agree with conclusions

    Votes: 53 20.3%
  • 4. Wonderful to have data and proof that such "upgrades" don't work

    Votes: 181 69.3%

  • Total voters
    261
Actually it's like saying there are no better DACs because no expensive DAC measures better than the best $100-$300 "mass-engineered" options.

Which is true.*


*There may be features/aesthetics/etc. reasons to choose a more expensive DAC but those are not relevant to our opamp analogy. This is purely about performance.
Ok,
Maybe this analogy works...

Imagine you have a 9mm pistol. You have two boxes of ammo.
Both weight the same. 115gr. Both have the same velocity and therefore roughly the same ballistics. (Keep in mind that there is a spread in velocities)
So you want to test this. You pull a round from each box at random. You pull the bullet, you weigh the bullet. the charge, and you verify that the numbers on the side of the box are correct.

You then take your chronograph and measure 10 rounds from each box. You get the average velocities and verify the data from the side of the box.

Now the difference between the rounds... one bullet is FMJ, the other is a critical defense round.

If we just looked at the ballistic measurements... you'd argue that there was no measurable difference therefore they are the same.
(Even though the critical defense costs twice as much.)
You've measured the ballistics, not its terminal effects. The critical defense round will expand/deform and dump more of its energy into the target (bad guy).

And this is like Amir's measurements. Note that I'm not arguing that his measurements are wrong but that they don't tell the complete story.
 
Unlike the Opamp the DAC is tested in isolation to get meaningful results.
if you would test DACs in a whole system with speakers and so on its unlikely you would be able to resolve DAC differences.

This dose not change the fact the some DACs are better then others.
The claim being tested here is that "rolling opamps" makes an audible (or even measureable) difference. It is entirely appropriate to measure the output of the device. Who cares if you get a different voltage off the pins of one opamp vs. another if there's no measureable different in the device when you swap one out for another?
 
Ok,
Maybe this analogy works...

Imagine you have a 9mm pistol. You have two boxes of ammo.
Both weight the same. 115gr. Both have the same velocity and therefore roughly the same ballistics. (Keep in mind that there is a spread in velocities)
So you want to test this. You pull a round from each box at random. You pull the bullet, you weigh the bullet. the charge, and you verify that the numbers on the side of the box are correct.

You then take your chronograph and measure 10 rounds from each box. You get the average velocities and verify the data from the side of the box.

Now the difference between the rounds... one bullet is FMJ, the other is a critical defense round.

If we just looked at the ballistic measurements... you'd argue that there was no measurable difference therefore they are the same.
(Even though the critical defense costs twice as much.)
You've measured the ballistics, not its terminal effects. The critical defense round will expand/deform and dump more of its energy into the target (bad guy).

And this is like Amir's measurements. Note that I'm not arguing that his measurements are wrong but that they don't tell the complete story.
Generally speaking if you're going to analogize it's a good idea to analogize to something that doesn't require specialized knowledge. I have absolutely no idea what you're trying to convey here except "measurements don't tell the whole story" which is an argument that's been had ad nauseum and isn't applicable to this kind of test - measurements do, in fact, tell the whole story here.
 
Note that I'm not arguing that his measurements are wrong but that they don't tell the complete story.
You are going to need to present some evidence for that.

Take your analogy - anyone who understands about how guns, ammo, ballistics and the terminal/catastrophic results of such would know that a critical measurement was missing.

Now assume that a bunch of highly qualified electronic and audio engineers are equally able to identify missing metrics from the current suite of tests AND explain how and why they are important.

What makes you think you know something they don't? Why do you think we would accept your statement that there are tests missing?

Where is your technical argument or engineering analysis of these missing measurements?


EDIT - Consider also that in all the history of audio, no-one has been able to demonstrate, using properly controlled blind listening, that they are able to hear anything that is not measurable in current audio measurements. If you think this is wrong you are free to prove us all wrong by being the first person to do so.
 
Last edited:
Generally speaking if you're going to analogize it's a good idea to analogize to something that doesn't require specialized knowledge. I have absolutely no idea what you're trying to convey here except "measurements don't tell the whole story" which is an argument that's been had ad nauseum and isn't applicable to this kind of test - measurements do, in fact, tell the whole story here.
Ok, lets look at it this way..
You have two bullets that have the same mass and are fired at the same velocity. So they have the same trajectory and will hit roughly in the same space on the target.
You would argue that because the ballistics are the same that there's no difference or advantage to buying the more expensive critical defense round.

After all your measurements are the same, which you verified w your testing....
 
You are going to need to present some evidence for that.

Take your analogy - anyone who understands about how guns, ammo, ballistics and the terminal/catastrophic results of such would know that a critical measurement was missing.

Now assume that a bunch of highly qualified electronic and audio engineers are equally able to identify missing metrics from the current suite of tests AND explain how and why they are important.

What makes you think you know something they don't? Why do you think we would accept your statement that there are tests missing?

Where is your technical argument or engineering analysis of these missing measurements?
Stopping power wasn't measured. Just the ballistics.

And yes that's part of the point.
You're measuring the ballistics, not the xfer of energy.
 
Maybe this analogy works...

Imagine you have a 9mm pistol. You have two boxes of ammo.
Both weight the same. 115gr. Both have the same velocity and therefore roughly the same ballistics. (Keep in mind that there is a spread in velocities)
So you want to test this. You pull a round from each box at random. You pull the bullet, you weigh the bullet. the charge, and you verify that the numbers on the side of the box are correct.

You then take your chronograph and measure 10 rounds from each box. You get the average velocities and verify the data from the side of the box.

Now the difference between the rounds... one bullet is FMJ, the other is a critical defense round.

If we just looked at the ballistic measurements... you'd argue that there was no measurable difference therefore they are the same.
(Even though the critical defense costs twice as much.)
You've measured the ballistics, not its terminal effects. The critical defense round will expand/deform and dump more of its energy into the target (bad guy).

And this is like Amir's measurements. Note that I'm not arguing that his measurements are wrong but that they don't tell the complete story.
Well, no single measurement can cover everything that can go wrong, but Amir's usual suite is going to be to be pretty hard to sneak something obvious past. If a significant deviation from "accurate" can be heard, it's entirely not obvious how it could have no correlate in frequency/phase response, non-linearity, or noise. An audio-frequency electrical signal just doesn't have that many dimensions. The most problematic sort of thing to measure would be "smart" devices (for example, some Cirrus Logic low-power chips that generate a quiet "click" when they switch power rails to optimize power consumption based on output level - if you are testing constant-level sine waves, you might not catch it if you aren't set up to catch it). Or something that is very sensitive to external loading. Shouldn't be an issue with internal audio op-amps, they are linear devices and their behavior is pretty deterministic/systematic.
 
First, consider that what you measure may not capture the differences in these op amps.
May not? So you don't know? Well I do. I know that you can't change the soundstage without the signal being fundamentally changed which vast array of measurements capture. Lower noise floor? Blacker backgrounds? Both would show up in SNR/DNR tests. Smoother highs? Frequency response.

"Slam?" There is no such thing.

Fact is that an amplifier just sees varying signals. You can't have it modify the output in such a way that simpler signals are not modified.

Or rather the perceived sound differences and people's preferences.
There is no discussion of preference when there is no difference. And perception is not equal to actual difference. Have you not heard the expression, "people eat with their eyes?" You think fancy restaurants are confused about this and should not decorate their plates of food when they send them out?

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.
You perceived a difference. You can't equate it with actual difference in sound waves until you do a proper controlled test. How many times do we have to explain this?

We all know that you perceive a difference. No need to keep repeating and clinging to this argument. We also all experience the same thing. The difference between us is that we know when to trust that impression vs not. If science, engineering and measurements all agree that there is no difference, then there is extremely high chance that this is the case. To counter it, you don't just make repeated word arguments. Come back with a listening test or for heaven's sake, please stop wasting our time with these arguments.
 
Well, no single measurement can cover everything that can go wrong, but Amir's usual suite is going to be to be pretty hard to sneak something obvious past. If a significant deviation from "accurate" can be heard, it's entirely not obvious how it could have no correlate in frequency/phase response, non-linearity, or noise. An audio-frequency electrical signal just doesn't have that many dimensions. The most problematic sort of thing to measure would be "smart" devices (for example, some Cirrus Logic low-power chips that generate a quiet "click" when they switch power rails to optimize power consumption based on output level - if you are testing constant-level sine waves, you might not catch it if you aren't set up to catch it). Or something that is very sensitive to external loading. Shouldn't be an issue with internal audio op-amps, they are linear devices and their behavior is pretty deterministic/systematic.
Ok, but I'm not questioning Amir's testing. I mean you can accept what he measured was correct.
But that it may not tell the whole story.

That may require ABX testing.
 
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.
That is folklore as well because they have not performed any controlled testing. Nor is their perception of fidelity proven or even demonstrated. They perceive a difference, then attempt, without any knowledge of electronics or audio science, attribute it to distortion being good. I have listened to high distortion tube amps. The distortion is sometimes not audible. Other times, it muddies the sound by stepping on low level detail. And clips the amp which really becomes audible. Never have I found distortion to be a good 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.
Just accept that these folks are wrong with their claims. There is a reason no official research entity such as AES, ASA, IEEE, etc. will ever accept these claims in a paper. We know what they perceive. So we are very familiar with the observations you are making. It is just that we have performed controlled testing and demonstrated to ourselves that such differences are not real. Just as science and engineering predicts.

I continue to do these tests but at some level, you need to accept the evidence and realize these folks are just wrong. Just as there are many people who are wrong in other fields.
 
May not? So you don't know? Well I do. I know that you can't change the soundstage without the signal being fundamentally changed which vast array of measurements capture. Lower noise floor? Blacker backgrounds? Both would show up in SNR/DNR tests. Smoother highs? Frequency response.

"Slam?" There is no such thing.

Fact is that an amplifier just sees varying signals. You can't have it modify the output in such a way that simpler signals are not modified.


There is no discussion of preference when there is no difference. And perception is not equal to actual difference. Have you not heard the expression, "people eat with their eyes?" You think fancy restaurants are confused about this and should not decorate their plates of food when they send them out?


You perceived a difference. You can't equate it with actual difference in sound waves until you do a proper controlled test. How many times do we have to explain this?

We all know that you perceive a difference. No need to keep repeating and clinging to this argument. We also all experience the same thing. The difference between us is that we know when to trust that impression vs not. If science, engineering and measurements all agree that there is no difference, then there is extremely high chance that this is the case. To counter it, you don't just make repeated word arguments. Come back with a listening test or for heaven's sake, please stop wasting our time with these arguments.
Ok,
You're the expert here.
So how do you measure the sound stage?

What part of the signal represents sound stage? How much change is required for it to be perceived?

That's what I've been asking...
 
But that it may not tell the whole story.
So what? We don't know the whole story about the Universe but know that earth is not flat. And that the moon exists and is not cheese. In your profession, do you know everything? You don't. But you know hell of a lot more than lay people outside of your field. And you make decisions based on proper knowledge that produce positive results. Do the same in audio. Don't go looking for evidence of something that doesn't exist. And be dismissive of proper explanation and data of what works just because so and so says otherwise.
 
Stopping power wasn't measured. Just the ballistics.

And yes that's part of the point.
You're measuring the ballistics, not the xfer of energy.

Read my post again - ANYONE - even lay people would recognise that that measurement was missing.


Yet all the much better qualified enineers who would be just as able - or more so - to recognise missing audio measurements agree that the current measurements are sufficient.

AGAIN - where is your technical/engineering analysis to show they are wrong.


Your repeated failure to engage with points raised is causing me to question if your involvement here is in good faith.
 
Last edited:
So how do you measure the sound stage?
In electronics (the subject here)? Crosstalk, but it has to be really bad to be audible.
 
So how do you measure the sound stage?
How do you measure soundstage? Go ahead and explain to me how you do that. Bring the details. Don't just tell me it got wider, deeper, etc. Tell me how exactly you measure it.
 
Ok, but I'm not questioning Amir's testing. I mean you can accept what he measured was correct.
But that it may not tell the whole story.

That may require ABX testing.
But all of your analogies are flawed because you are handicapping them each time you use the "lets say/look at it this way" phrase. If someone was going to test the difference between two different bullets, entrance/exit wounds, penetrating power, armor plate level pentration, etc are all things we know should be objectively tested to0, along with other metrics you listed.

What you keep trying to do is imply there is some untested/unknown physical property to sound that is the reason people hear differences between equally measured components but which no one has tested for yet.

In terms of grasping at straws....10/10
 
Last edited:
So how do you measure the sound stage?
"sound stage" isn't a function of opamps, it's created by speakers interacting in a room and can be predicted by measuring frequency response and dispersion. Unless you've got an extraordinarily broken design, your preamp and amplifier are not going to be altering the frequency response and cannot alter dispersion.

Everything you've posted here seems to boil down to "there must be a difference, so you must not be measuring properly".

There's one sure-fire way to learn if the measurements aren't measuring everything, and that's to do properly controlled blind listening tests. If listeners can consistently hear a difference when the measurements say there isn't one, then we're not measuring the right things. If the measurements say there's no difference and listeners in a properly controlled test also say there's no difference, there's no difference.

I'm not aware of any properly-controlled blind tests that show a audible differences when opamp rolling, so I'm going to err on the side of the measurements and say changing out your opamps is a waste of time and money. If you've got access to some controlled test results that show otherwise, please share them, I'm sure everyone would like to see them.
 
Ok, but I'm not questioning Amir's testing. I mean you can accept what he measured was correct.
But that it may not tell the whole story.
So you have this idea. The notion/perception that there is something "more", something everybody here and the measurements are missing.
I assume you like this idea, it feels "right", convincing and clear.
For me that is totally legit.
But then you do something: you make your idea into a claim.
And a claim is a different thing, it calls for general acceptance, for undeniable reality outside of your hearing/feeling.
"It MUST be right! There HAS to be something ...".
And for this you do not present any evidence whatsoever, only more and additional claims. And the "argument" that others propose the same thing. But these people too, only have a "feeling" they bring forward. This is not how science works.
If you really think your idea is worth it, then you should investigate and try to find evidence. You cannot just go on with claims and "analogies".
One of the first things would be to check your perception in a controlled way with controlled (blind and matched) listening.

OK, I see SuidiceSquid beat me to it in just about every point.
 
This is not correct.
Ohm was teacher in some kind of Highschool (for the Jesuit Order) till 1826. One year before the "pamphlet" he went to Berlin to become teacher at a Prussian cadet academy. For some reason the cadet academy was not amused about the research and he resigned. In 1833 he became Professor (in Nuremberg) mainly because of his "pamphlet". So the scientific community obviously had a different estimation of him. Part of this estimation certainly came from him actually doing experiments and showing evidence for his theory.
Now, do you consider yourself as some kind of Georg Simon, or as a chosen one from the perceptive few, or what is this about?

In other words, you do not actually have any evidence to reference?
This was a quote from "Small Signal Audio Design" by Douglas Self.
 
Back
Top Bottom