Does Op-amp Rolling Work? [Video]

[DualTriode]

So, turning up the volume above 5 watts does make sense.

So @Ant and All, The 1 kHz FFT plot posted by @amirm does show different harmonics. The stock Op-Amp shows higher High Order Distortions. The higher order HD's that show up in in the AP Rub and Buzz plots. Can you hear driver Rub and Buzz?

The first claim was that the measurements were the same, we can see that they are not. The no difference claim is an exaggeration. True

Then came the claim that you can not hear the difference.

The 1 kHz FFT plot posted by @amirm does show different harmonics. That 1 kHz FFT plot only shows the 1 kHz harmonics. Take the next step and run a multi-tone FFT plot. Closer to the multi-tone environment of playing "real music", all the harmonics show up including the Higher Order harmonics. Quite the cacophony. Will any of this show up in the SINAD number? Not really, the value of SINAD is limited to a single measurement at 1 kHz.

Take a look at the 6:45 mark in the original A5 amplifier review. Do you think that you can hear the higher High Order Distortions of the 1 kHz tone, likely not. The multi-tone Plot shows a pile of harmonic hash between/among the test tones. Do you think that you can hear that? If you can, it will not be identified as a barley audible 5th harmonic of the 1kHz test tone.

Douk Audio A5 Stereo Budget Amplifier Review [Video]

Video review of the Douk Audio A5 amplifier is out including a bit about listening tests that was not in the text review: https://www.audiosciencereview.com/forum/index.php?threads/douk-a5-budget-stereo-amplifier-review.61410/ Please subscribe to the channel. It is not monetized but like to...

www.audiosciencereview.com

 

[amirm]

Not really, the value of SINAD is limited to a single measurement at 1 kHz.

No. You know the SINAD at ALL levels and a number of frequencies:

The difference is so small as to be due to performance differences between channels. And at any rate they say that the discrete amp was slightly worse.

 

[NTK]

Take a look at the 6:45 mark in the original A5 amplifier review. Do you think that you can hear the higher High Order Distortions of the 1 kHz tone, likely not. The multi-tone Plot shows a pile of harmonic hash between/among the test tones. Do you think that you can hear that? If you can, it will not be identified as a barley audible 5th harmonic of the 1kHz test tone.

There are 32 tones in the multitone test, spread over 10 octaves, therefore the spacings between each tone is about 1/3 octave. The "grass heights" are ~80 dB below the tones at 10 kHz. You think they are audible (see masking curves below)?

 

[DualTriode]

No. You know the SINAD at ALL levels and a number of frequencies:

The difference is so small as to be due to performance differences between channels. And at any rate they say that the discrete amp was slightly worse.

If you read my post the discussion was noted as the SINAD as it is shown on your dashboard. That 1 kHz SINAD is used to rank all the amplifiers that you rank on your "eye chart".

 

[amirm]

If you read my post the discussion was noted as the SINAD as it is shown on your dashboard. That 1 kHz SINAD is used to rank all the amplifiers that you rank on your "eye chart".

What? You claimed that SINAD is only shown as a single measurement. I showed you that the analysis included some 100 such SINAD measurements.

And no, SINAD alone doesn't determine the fate of an amplifier. If for example it has load dependency, it would rank lower, all else being equal.

 

[DualTriode]

There are 32 tones in the multitone test, spread over 10 octaves, therefore the spacings between each tone is about 1/3 octave. The "grass heights" are ~80 dB below the tones at 10 kHz. You think they are audible (see masking curves below)?

I am not saying that you can hear anything.

As I said, it will not likely be the 5th harmonic of the 1 kHz test tone, or any other single harmonic.

That pile of harmonic hash is much more likely to be something that you can hear.

That pile of harmonic hash is not grass, it is not random, it is not noise. That pile is distortion. Can you hear it.

This part no one has commented on or disputed:

The measurements we have looked at so far are taken at the amplifier output. The sound/distortion that we possibly can hear is the the driver output. Drivers can largely contribute to intermodulation distortion all on their own.

I like the Two-Tone test to visualize driver IMD sidebands.

Before I am run out of town. I am taking a break.

 

[craigenputtock]

Yes. When you go to change the op-amp, it is so easy to touch the volume control. Unless you measure the output, which of course they did not, you can't level match. This is what I did after changed the op-amps. Not for that reason but to make sure I was testing at the same 5 watt output as my review. It is likely that neither one of them thought that levels had to be the same. And worse yet, couldn't tell after they played the files!

Now that is giving them the benefit of the doubt. It is kind of hard to do that when files progressively got louder toward their favorite option. And even EQ changed. These point to experimenting until they found the results they wanted, not knowing that in this day and age, people do quick analysis.

We know they lied about this work as Danny claims they were experimenting with this idea for "months." Yet, they repeated my exact test of Douk A5 with these op-amps. Clearly he is lying about this. If so, then he is liable to lie about other things as well.

Is it only a matter of volume, though? Did you detect any difference in tonal qualities? Do you not think that two circuits of different topologies cannot sound different even though they measure the same? I think that's really the crux the matter.

 

[Jimbob54]

If you did not get it, my post was meant to be sarcastic, numbnuts.

If you're going to be sarcastic in a wall or text you probably need to make it clearer .

Putting /s after a sarcastic comment can emphasise .

Let's leave the insults out of it please. Thanks

Edit I see @Mort said it first and better.

 

[lateralous]

Is it only a matter of volume, though? Did you detect any difference in tonal qualities? Do you not think that two circuits of different topologies cannot sound different even though they measure the same? I think that's really the crux the matter.

This is a stream of random questions, where is the crux? What is only a matter of volume? The comparison of the provided files is invalid because the recordings are not level matched, full stop. Any perceived difference in tonal, or other <insert colorful word> quality can never be decoupled from the change in level. Your question regarding amplifier topologies is irrelevant to the discussion and unanswerable without definition.

 

[amirm]

Is it only a matter of volume, though? Did you detect any difference in tonal qualities? Do you not think that two circuits of different topologies cannot sound different even though they measure the same? I think that's really the crux the matter.

You don't get to "pass go" if volumes are not matched. As to measurements, we don't have that for their configuration. They did not bother to do this. So there is nothing to conclude. They need to learn what it takes to make proper comparisons.

 

[craigenputtock]

This is a stream of random questions, where is the crux? What is only a matter of volume? The comparison of the provided files is invalid because the recordings are not level matched, full stop. Any perceived difference in tonal, or other <insert colorful word> quality can never be decoupled from the change in level. Your question regarding amplifier topologies is irrelevant to the discussion and unanswerable without definition.

"Your question regarding amplifier topologies is irrelevant to the discussion ...." --On the contrary, that's the whole point of using a different op-amp.

 

[Scrivs]

"Your question regarding amplifier topologies is irrelevant to the discussion ...." --On the contrary, that's the whole point of using a different op-amp.

Fellas, there's no reaching the willfully stupid.

 

[solderdude]

Do you not think that two circuits of different topologies cannot sound different even though they measure the same?

If they measure the same (not just one single measurement) then the output signal is the same.

• Where would any audible difference come from ?
• Why would we 'think' op-amps would have properties that cannot be measured but are audible ?

Of course one can use op-amps that are not really suitable for usage in an amplifier. There are plenty of examples measured here.
We are talking about suitable op-amps here in this particular circuit (Douk amp) not in a specialized circuit.

Op-amps can have different topologies which can make them more or less suitable for the job and not all topologies are easily interchangeable without looking at components around them. Voltage feedback and current feedback are the possible topologies.
They both are designed to do the same job but take a different approach for the feedback mechanism. Ultimately it is the components around the op-amp that determine what the op-amp does. The different topologies often require some additional care of components around the op-amp in order to get stable operation.
When the circuits measure 'the same' within the 'audible band' and are not driven outside of the required parameters then how can the different topologies still sound different.

• What is the reasoning/belief behind this ?

Then within the different topologies there are also fundamentally different components that can be used throughout the op-amp (be it one-chip or discrete) such as BJT, MOSFET or JFET and using components that aren't available in discrete form.
The usage of these components determines some of the specifications from op-amps that may matter in certain circuits that require specific properties (low input currents, rail-to-rail in/out, low/high operation voltages for instance).
In all cases the components around the op-amp and the properties of the op-amp determine whether or not a specific op-amp can be used in a circuit and what components around it are needed.
So not all op-amps are suitable and simply can (should) not be used and makes certain op-amps not suitable for audio equipment.

You simply cannot swap out an op-amp for any other one. What can be done in some cases is swap suitable op-amps in a circuit.
But ... that device has to be checked for proper operation using measurements ... ears simply can't do that as they are way too limited.

Ears should be used to enjoy music, the analyzing is best left to test equipment (with multiple tests) revealing the total performance.

Is it only a matter of volume, though?

In the GR video/files to download... YES, level as well as the clipping.
The reason the files aren't level matched is because otherwise you can't hear any differences and the goal is to make differences audible because he says there are differences.
Thus the most likely scenario here is that the files are deliberately made with different levels and one even clipping.
The thing is that simply swapping an op-amp can NOT (as in impossible) lead to such substantial level differences nor clipping as in one of the files.

 

[antcollinet]

Do you not think that two circuits of different topologies cannot sound different even though they measure the same?

There is only the signal, then distortion, noise and frequency response to change it. All of these are measureable. What do you think these “unmeasurable” differences consist of?

 

[lateralous]

"Your question regarding amplifier topologies is irrelevant to the discussion ...." --On the contrary, that's the whole point of using a different op-amp.

Quite telling you've ignored the rest of my comment.

 

[Old_School_Brad]

I was just browsing through GR-Researchs response video comments today, and it seems like a lot of the critical ones have been deleted. I'm pretty sure there were twice as many critical comments just a few days ago...

 

[Philosoraptor]

Interesting video, Amir. I found your explanation on when discrete components are used (e.g. power & heat) over ICs interesting.

 

[DualTriode]

For grins a A5 amplifier arrived on my doorstep yesterday.

The plot is the amplifier stock at 5 watts output There several other Op-Amps sitting here, not the 50 dollar variety, all $10 or less.

Thanks DT

 

[DanTheMan]

For grins a A5 amplifier arrived on my doorstep yesterday.

The plot is the amplifier stock at 5 watts output There several other Op-Amps sitting here, not the 50 dollar variety, all $10 or less.

Thanks DT

View attachment 440271

Any chance you’re going to post some level matched sound clips?

 

[DualTriode]

For grins a A5 amplifier arrived on my doorstep yesterday.

The plot is the amplifier stock at 5 watts output There several other Op-Amps sitting here, not the 50 dollar variety, all $10 or less.

Thanks DT

View attachment 440271

Hello All,

FFT BW was 22.4 / 48kHz FFT length was 48K with 5 averages
Sweeps at 96kHz to capture more of the Harmonics. Note the 96kHz, the wider BW includes more noise so the SINAD looks a little lower.

This time a different 5532 and
OPA2134
LN4562 this one was maybe one SINAD point less than the others
OPA1656

So now after looking, no or tiny differences among the OP-Amps that I looked at. Roll on only if it is fun for you.