Does Op-amp Rolling Work?

[amirm]

It's good to know that most of the tests generally carried out here are of no interest...

If you look at my audio tweak reviews, they are similar. When from engineering perspective we know there is no difference, and then the first few measurements confirm the same, we don't act dumb and keep going. If by then you don't believe, well, that is it. We have lost you as a reader/viewer.

Further, I have often gone quite very far, producing null tests and such with a ton of effort, yet the people who didn't want to believe, still didn't. Still keep saying, "but did you listen?"

As an engineer, I have to optimize what I do or I would be sitting here forever testing one thing.

Finally, remember a few of these tests were added just to quiet people. Multitone is one of those. "Oh you only test with one sine wave." OK, that has 32. Then I added 19+20 Khz because that is what stereophile uses, not because we needed it. As I have said before, the dashboard almost entirely predicts the performance of a product. If there is great engineering, you see it there. If it is poor, you see it there. And in this case, when not a hair seems to have changed, then the rest of the test won't show anything either.

 

[amirm]

Adding on, I wasn't even going to run any tests for this video! I thought I just explain why swapping op-amps in an already optimized design is not of value. And that if you want something better, get another product with better design and implementation. Then I thought I should add some "meat" to the plate and ran the dashboard. Thinking someone would complain about noise, I ran that. Finally to address the "only tested at 1 kHz," I ran the power sweeps. By then I was ready to do the video, thinking it would be a low-key affair.

 

[Heinrich]

I am still waiting for the explanation on the fact that ... would not make sense in terms of methodology here...

Sorry to say, but I dropped in some anchor points for a reasonable discussion in this thread already. If you are willing to explain what you expect for a test, for what reason, we can do. Even I personally can do. I've got a 24bit digitizing apparatus, that should be sufficiant. Again, if not, please explain why. Please be aware that you deal with other people's precious time.

I only ask, because admittedly I don't see anything that you could tell. What is the meaning of measurements you propose (related to HD measurements)?

 

[morillon]

If you look at my audio tweak reviews, they are similar. When from engineering perspective we know there is no difference, and then the first few measurements confirm the same, we don't act dumb and keep going. If by then you don't believe, well, that is it. We have lost you as a reader/viewer.

Further, I have often gone quite very far, producing null tests and such with a ton of effort, yet the people who didn't want to believe, still didn't. Still keep saying, "but did you listen?"

As an engineer, I have to optimize what I do or I would be sitting here forever testing one thing.

Finally, remember a few of these tests were added just to quiet people. Multitone is one of those. "Oh you only test with one sine wave." OK, that has 32. Then I added 19+20 Khz because that is what stereophile uses, not because we needed it. As I have said before, the dashboard almost entirely predicts the performance of a product. If there is great engineering, you see it there. If it is poor, you see it there. And in this case, when not a hair seems to have changed, then the rest of the test won't show anything either.

I'm a bit like you...
"I know in advance"...
I know in advance the result "transparent, at the top... etc" of a good number of machines tested here...
we were developed-tested upstream by their manufacturers to respond to the protocol known here, on the same benches...
and yet they are entitled to a complete test including these famous imd tests which seem in your opinion hardly useful etc ( and a sinad ranking which would no longer have any meaning for a while with machines which could no longer be distinguished...)
funny ...
;-)

(as a reminder...the fosi phono...or a protocol was not so relevant to predict the result in certain cases...including not relevant...... It calls for humility, doesn't it?..a story of AOP from elsewhere...
;-) )

 

[SIY]

My point, however, is that since the descriptors clearly align with EQ characteristics -and these are the very features many claim to hear -a frequency response comparison would likely be more effective in persuading (at least some of) those individuals.

If one is a True Believer, nothing will persuade. Including the basic explanation that the op-amp can't actually affect the frequency response (and the True Believers sniff at the idea that these magical qualities they hear are matters of EQ rather than the mysterious black magic of faith-based designers) and a Bode plot of the gain vs. frequency to bring that point home.

Scott Wurcer is rolling in his grave.

He was a dear friend and I miss him horribly. Yeah, zero tolerance for audio superstition. I hope he and my wife (who were great friends as well) are drinking La Tache in Heaven.

 

[amirm]

I know in advance the result "transparent, at the top... etc" of a good number of machines tested here...

No, we don't know that. I have tested products that surprised me in both directions. I don't have visibility into design differences of the complete products I test.

This case is difference because we know what change we are making and the likely outcome of it.

Regardless, you haven't show why more testing would do any good. If you are not convinced and pretending to speak for others, then make your case. Get this amp, swap the op-amps and run whatever test you want to make your point.

 

[BDWoody]

moreover, your rigor shows the lightness of the approach taken here...

That's enough.

 

[MAB]

but the very limited basis presented here does not allow it...
moreover, your rigor shows the lightness of the approach taken here...

You've done this FUD-casting before, I pointed it out then, you even said you would remember. Well, I'm still here and you are still unnecessarily casting FUD on other's work, over and over, with no work of your own to demonstrate anything. You seem to be incapable of not casting FUD, and incapable of backing up any of the doubt you cast with actual data. You are repeating the FUD-casting behavior here, ad infinitum.

For crying out loud, it's not like any of this requires a fancy analyzer... a sound card and good technique will do, instead you are asking for someone to follow up on your fantasy with an analyzer that costs more than a car and a large amount of their time which you seem to not value. Get a sound card, figure out how to measure this stuff, roll some opamps, and report back on what you found!

well yes.....to observe if there are differences on tests recognized for this type of use...
(I don't predict the results, really... but it would be a bit serious to do them... at least push the tests to the maximum of what we know how to do... that's all...)
stop

Yes, please stop and reflect. And if you can, contribute. Otherwise you are one of the many trolls who come here to offer nothing but uncertainty. And this is among the easier tests, even if a bit time consuming.

 

[oleg87]

I'm a bit like you...
"I know in advance"...
I know in advance the result "transparent, at the top... etc" of a good number of machines tested here...
we were developed-tested upstream by their manufacturers to respond to the protocol known here, on the same benches...
and yet they are entitled to a complete test including these famous imd tests which seem in your opinion hardly useful etc ( and a sinad ranking which would no longer have any meaning for a while with machines which could no longer be distinguished...)
funny ...
;-)

(as a reminder...the fosi phono...or a protocol was not so relevant to predict the result in certain cases...including not relevant...... It calls for humility, doesn't it?..a story of AOP from elsewhere...
;-) )

While personally I have nothing against a well-defined and perhaps automated panel of tests that includes various IMD metrics (but I don't run ASR nor have an AP box), considering in this case the Douk seem to be overwhelmingly limited by the class D amp in any case, especially at high frequencies, the IMD is unlikely to reveal anything new.
Ultimately IMD is caused by non-linearity, if overall THD is low, IMD is not going to be an issue either, and at frequencies where that may not be the case, the op-amp isn't going to be dominant nonlinearity contributor by a mile.

 

[antcollinet]

if overall THD is low, IMD is not going to be an issue either,

More to the point - if THD doesn't change then nor will IMD. Whether it is high or low, if it is the same with two different op amps, then so will IMD be with those two op amps. As will the multi-tone.

 

[aagstn]

I just had some random YouTuber I had never heard of recommended to me and he must have seen Amir's video as he went on a measurements are meaningless and I hear a difference in Op Amps rant. They miss the whole point of why do they hear a difference. Is there a volume difference, are they focused more on the audio after the switch. All they care about is they hear a difference so it is okay for them to tell all their viewers to spend stupid money rolling Op Amps. There is no defeating the ignorance in the audio world but at least the sanity is out there for people that care to look for it.

 

[JSmith]

All they care about is they hear a difference so it is okay for them to tell all their viewers to spend stupid money rolling Op Amps.

I think there is also a few that seem to consider anything to the contrary as some kind of personal attack, or questioning their hearing abilities... when it's nothing of the sort.


JSmith

 

[Old_School_Brad]

If one is a True Believer, nothing will persuade. Including the basic explanation that the op-amp can't actually affect the frequency response (and the True Believers sniff at the idea that these magical qualities they hear are matters of EQ rather than the mysterious black magic of faith-based designers) and a Bode plot of the gain vs. frequency to bring that point home.

Sure, those types exist. However, some people have doubts and only need a slight push to challenge their beliefs.

Amir conducted this test -presumably to provide the most definitive proof that there are no audible differences between op-amps. Yet, in my opinion, an opportunity was missed. The main characteristic often attributed to different op-amps -their frequency response, was not directly compared, leaving room for continued skepticism.

Rather than moving on to review the next SINAD-breaking little box, perhaps a greater effort could be made to put these "snake oil" claims to rest once and for all. This could be done through a full suite of thorough tests, volume- and time-matched recordings, and a structured blind listening test using Foobar. A simple, step-by-step guide on how to install Foobar, set up the required plugin, and conduct an A/B comparison would go a long way in making such tests accessible to the broader audience.

Amir has often stated that the focus should be on the real issues in the industry. This is undeniably one of them. Maybe he says this when he’s tired of certain debates, but if he truly means it, now is the time to put this topic to rest once and for all. I think many cling to their beliefs simply because they don’t know how to conduct proper blind tests, and, perhaps, are too proud to admit it. By making the process easy and accessible, more people might finally see the value of putting science in audio and start reading objective reviews.

 

[pma]

Whether the opamp rolling makes difference or not is strictly defined by application circuit and opamp parameters. In the test like the one made here, changing opamps probably makes no difference. In case of microphone preamp with gain up to 60 dB or in case of MM phono preamp operating in real world conditions (with the cartridge), the opamp rolling makes crucial differences. To know if there would be a difference, it is necessary to have circuit schematics and opamps datasheets. Then, any good analog designer would tell if there would be a difference or not. Datasheet parameters like open loop gain plots, voltage noise density and current noise density plots, output current capability etc give the full scale view of the problem and all answers. On the other hand, to generalize based on the test posted here is impossible. It covers only few real life circuit applications.

 

[Blumlein 88]

If one is a True Believer, nothing will persuade. Including the basic explanation that the op-amp can't actually affect the frequency response (and the True Believers sniff at the idea that these magical qualities they hear are matters of EQ rather than the mysterious black magic of faith-based designers) and a Bode plot of the gain vs. frequency to bring that point home.

He was a dear friend and I miss him horribly. Yeah, zero tolerance for audio superstition. I hope he and my wife (who were great friends as well) are drinking La Tache in Heaven.

I wonder how many op amp rollers know what a Bode plot is?

 

[amirm]

Amir conducted this test -presumably to provide the most definitive proof that there are no audible differences between op-amps. Yet, in my opinion, an opportunity was missed. The main characteristic often attributed to different op-amps -their frequency response, was not directly compared, leaving room for continued skepticism.

Not a single person brought this up as a negative. They just talk about soundstage being wider, only ears matter, etc. None would care if I had or had not shown the frequency response. Go and read the comments before opining on backhanded complaints about my testing.

 

[amirm]

Rather than moving on to review the next SINAD-breaking little box, perhaps a greater effort could be made to put these "snake oil" claims to rest once and for all.

Once and for all? If that is the case, why don't you do us a favor and buy this amp, the opamps and perform said frequency response measurement.

The measurements and video have made a significant impact as evidenced by nearly 1000 comments. Many are complimentary saying I saved them money. Whoever needed convincing, is onboard.

 

[amirm]

On the other hand, to generalize based on the test posted here is impossible. It covers only few real life circuit applications.

The "real-life" application is in both desktop and Hypex/Purifi class D amps. Many allow selection of op-amps at the point of purchase or after the fact. I had tested Hypex/Purifi and now I have done the little boxes. No one is talking about what a single opamp does by itself. If you think the buffer/gain stage sets the audible performance of the output stage and not the other way around, let's hear it. For now, this is what you are dealing with:

Tell us in what way the opamp contributed to those graphs.

 

[JSmith]

MM phono preamp

Unsure why you bring up that topic into this quite different discussion?


JSmith

 

[MAB]

Whether the opamp rolling makes difference or not is strictly defined by application circuit and opamp parameters. In the test like the one made here, changing opamps probably makes no difference. In case of microphone preamp with gain up to 60 dB or in case of MM phono preamp operating in real world conditions (with the cartridge), the opamp rolling makes crucial differences. To know if there would be a difference, it is necessary to have circuit schematics and opamps datasheets. Then, any good analog designer would tell if there would be a difference or not. Datasheet parameters like open loop gain plots, voltage noise density and current noise density plots, output current capability etc give the full scale view of the problem and all answers. On the other hand, to generalize based on the test posted here is impossible. It covers only few real life circuit applications.

I agree with what you say here.
I just don't know of any instrumentation amplifiers or mic preamps socketed for opamp rolling.
So within the narrow confines of consumer electronics for people who wish to randomly swap opamps, this device and these opamps in this test seems more than sufficient to determine swapping opamps in this amp is a waste of time and money.