Serious Question: How can Op Amp rolling improve the sound of an already well measuring device? Are that many confused? (Master Thread)

[Mort]

The obvious suspicion is really sloppy test setup, making it worthless. The worse suspicion is deliberate manipulation to fabricate perceivable differences.

¿por qué no los dos?

 

[dark132]

I believe more in the measurement data of the instrument, if the instrument shows no change, then there is no change, or the change is beyond the degree that the human ear can distinguish, and the human feeling is indeed unreliable

 

[Geert]

Since we love to have our facts straight I checked myself (a bit handicapped working on a Linux PC with only Audaciity installed).

@pkane, I made an attempt to compare Danny's opamp recordings via DeltaWave but the delta's are that large that it's difficult to believe they result from swapping opamps. Can you have a look at these graphs (file 1 versus 2) and check if you see something that might point to processing issues (incl. me making errors)? (Result file is also in attachment).

Edit: I trimmed the start and the end of the files manually because it looks like there's an alignment issue, but I can't get a better result.

Zoomed in alignment

With file 1 versus file 3 DeltaWave had issues with matching clock drift, which is also strange.

 

[Cbdb2]

@pkane, I made an attempt to compare Danny's opamp recordings via DeltaWave but the delta's are that large that it's difficult to believe they result from swapping opamps. Can you have a look at these graphs (file 1 versus 2) and check if you see something that might point to processing issues (incl. me making errors)? (Result file is also in attachment).

Edit: I trimmed the start and the end of the files manually because it looks like there's an alignment issue, but I can't get a better result.

View attachment 437738
View attachment 437739
View attachment 437740

Zoomed in alignment
View attachment 437776

With file 1 versus file 3 DeltaWave had issues with matching clock drift, which is also strange.

If douchebag manipulated the files and is lying, the levels are different and the video commenter said it sounded like he was using a hpf, which would not surprise me, your wasting your time. Hes a scammer who knows little real electronics. Come on, different levels for a comparison test? Hes either scamming or knows nothing about this subject. I think its both.

 

[Scytales]

Since we love to have our facts straight I checked myself (a bit handicapped working on a Linux PC with only Audaciity installed). The 3 recordings available on Danny's video, start of the tracks perfectly aligned. You can even see the left channel of the 3th track clipping just by looking at the waveform:

View attachment 437238

@pkane, I made an attempt to compare Danny's opamp recordings via DeltaWave but the delta's are that large that it's difficult to believe they result from swapping opamps. Can you have a look at these graphs (file 1 versus 2) and check if you see something that might point to processing issues (incl. me making errors)? (Result file is also in attachment).

Edit: I trimmed the start and the end of the files manually because it looks like there's an alignment issue, but I can't get a better result.

View attachment 437738
View attachment 437739
View attachment 437740

Zoomed in alignment
View attachment 437776

With file 1 versus file 3 DeltaWave had issues with matching clock drift, which is also strange.

As I understand it, the 3 soundtracks are all analogue to digital recordings of the output of the Device Under Test (DUT) made with an ADC or sound card of some kind.

So, assuming no errors or cheating were made by the author with the test or the recording set-up, couldn't the differences be explained by analogue distortion of the waveform shape produced by the DUTs ?

For example, If we suppose one or more of the op-amps under scrutiny are not stable and are on the brink of oscillation, or outright oscillates, perhaps the analogue to digital recorder just capture a distorted, exaggerated amplitude on some transients due to oscillation ?

Sound engineer Ian Sheperd has illustrated a somewhat similar phenomenon when he compared the CD release and the vinyl release both made from his own same digital master with the notorious tool used to built the online Dynamic Range Database:

The vinyl appears more dynamic than the original master put unscathed on the CD release ! It is my belief that Sheperd's observation is due to the fact that vinyls are made and played back with electro-mechanical devices which can be unfaithful to the original waveform shape, under-damped, and somewhat act like an expander that can artificially enlarge the waveform amplitude.

Going back to the op-amps under scrutiny in the sound samples provided with the GR-Research youtube video, I think that we cannot rule out the possibility that some of those discrete op-amps have been deliberately designed with additional time constants in order for them to act like a subtle tone control or equalizer or components put in parallel with the outboard feedback loop to alter the gain. Perhaps the output impedance of some of this op amps is way out of the value expected by the following stage in the Douk Audio amplifier, adding another layer of possible distortion. Reverse engineering those op-amps could resolved any doubt.

 

[SIY]

I think that we cannot rule out the possibility that some of those discrete op-amps have been deliberately designed with additional time constants in order for them to act like a subtle tone control or equalizer.

Do you understand how opamps work? The gain and frequency response are set by the external components.

Here's the open loop response of a common high performance opamp.

All sorts of "time constants" happening there, none of which show up in an actual functional circuit.

The guy is a fraud and it's naïve to try to defend him.

 

[pkane]

@pkane, I made an attempt to compare Danny's opamp recordings via DeltaWave but the delta's are that large that it's difficult to believe they result from swapping opamps. Can you have a look at these graphs (file 1 versus 2) and check if you see something that might point to processing issues (incl. me making errors)? (Result file is also in attachment).

Edit: I trimmed the start and the end of the files manually because it looks like there's an alignment issue, but I can't get a better result.

View attachment 437738
View attachment 437739
View attachment 437740

Zoomed in alignment
View attachment 437776

With file 1 versus file 3 DeltaWave had issues with matching clock drift, which is also strange.

Sorry, wasn't following the discussion here. Where can I get the clips to be compared?

 

[Scytales]

Do you understand how opamps work? The gain and frequency response are set by the external components.

[...]

The guy is a fraud and it's naïve to try to defend him.

I am not trying to defend anyone.

I just search for reasonable explanations for the measured differences under the following condition I have outlined : 'assuming no errors or cheating were made by the author with the test or the recording set-up'.

The op-amps used in the GR-Research youtube video are not off-the shelf op-amp from reputable established manufacturers of ICs, but discrete op-amps built by some companies that address a particular clientele. I for myself cannot rule out the possibility that those discrete op-amp are actually not plain simple op-amps but that their circuit boards are deliberately designed with built-in components to alter the gain or frequency response and/or that those discrete op-amps cannot be used unconditionally in any existing circuit and remain stable.

 

[SIY]

I for myself cannot rule out the possibility

But anyone who understands electronics can. It doesn't matter what you do in those modules, you cannot increase the gain of the circuit.

He's a fraud.

 

[Geert]

Sorry, wasn't following the discussion here. Where can I get the clips to be compared?

They are linked to the YouTube video referenced earlier. Links:

1. https://bit.ly/3FPAKde (Stacatto + Stock)
2. https://bit.ly/4bJvsw4 (Stacatto + Sparkos)
3. https://bit.ly/4kxKlpb (Sparkos + Sparkos)

 

[solderdude]

Blind test the original files.
Then ...normalize all 3 files... check for same volume/LUFS etc to be within 0.1dB.
Blind test these equal volume samples.
Report here.

 

[Geert]

The op-amps used in the GR-Research youtube video are not off-the shelf op-amp from reputable established manufacturers of ICs, but discrete op-amps built by some companies that address a particular clientele. I for myself cannot rule out the possibility that those discrete op-amp are actually not plain simple op-amps but that their circuit boards are deliberately designed with built-in components to alter the gain or frequency response and/or that those discrete op-amps cannot be used unconditionally in any existing circuit and remain stable.

From the SparkosLabs website: "We designed these devices to be as sonically transparent as possible. This means that the op amp amplifies the incoming signal, and produces an exact, larger copy at the output without coloring the sound or creating any distortion products (THD)". The opamps are designed to easily integrate (remain stable) . The website also shows THD plots made by ATM Audio and the performance is very good. ATM tested these opamps in the input buffer of a Hypex NC500 amplifier.

 

[antcollinet]

I just search for reasonable explanations for the measured differences under the following condition I have outlined : 'assuming no errors or cheating were made by the author with the test or the recording set-up'.

And there is the problem. You start from a false premise.

 

[SIY]

And there is the problem. You start from a false premise.

“If my mother had two wheels, she’d be a bicycle.”

 

[pkane]

They are linked to the YouTube video referenced earlier. Links:

1. https://bit.ly/3FPAKde (Stacatto + Stock)
2. https://bit.ly/4bJvsw4 (Stacatto + Sparkos)
3. https://bit.ly/4kxKlpb (Sparkos + Sparkos)

Here's comparison. 2 and 3 are clipping in a number of places and there's an obvious level difference between 1 and 2 and 1 and 3. #1 is not clipping anywhere.

Comparing 1 to 2, frequency response is strange (those consistent frequency lines across the recording) and there's 0.93dB difference in levels:

Removing clipped areas, the RMS of the difference is -40dB full spectrum, and -49.6dBA, which is not awful, but still suspicious due to the frequency artifacts (and clipping).

Comparing 1 to 3 shows a larger difference in level of 2.3dB, and null RMS value of only about -22dB and 42dBA. Spectrogram shows larger, more pronounced frequency bands across the recording:

 

[Old_School_Brad]

Here's comparison. 2 and 3 are clipping in a number of places and there's an obvious level difference between 1 and 2 and 1 and 3. #1 is not clipping anywhere.

Comparing 1 to 2, frequency response is strange (those consistent frequency lines across the recording) and there's 0.93dB difference in levels:
View attachment 437977

Removing clipped areas, the RMS of the difference is -40dB full spectrum, and -49.6dBA, which is not awful, but still suspicious due to the frequency artifacts (and clipping).

Comparing 1 to 3 shows a larger difference in level of 2.3dB, and null RMS value of only about -22dB and 42dBA. Spectrogram shows larger, more pronounced frequency bands across the recording:
View attachment 437982

So, it's reasonable to conclude that the files aren't worth comparing due to multiple factors, including level differences, clipping, and unexplainable artifacts?

 

[Geert]

Removing clipped areas, the RMS of the difference is -40dB full spectrum, and -49.6dBA, which is not awful, but still suspicious due to the frequency artifacts (and clipping).

Thanks. I only used the first 40 sec of the tracks to stay out of the part that’s clipping and got similar delta value. Suspicious indeed. The root cause remains unknown, but it's safe to conclude people shouldn't use these recordings to get a reliable impression of the effect of opamp swapping.

 

[oleg87]

The op-amps used in the GR-Research youtube video are not off-the shelf op-amp from reputable established manufacturers of ICs, but discrete op-amps built by some companies that address a particular clientele. I for myself cannot rule out the possibility that those discrete op-amp are actually not plain simple op-amps but that their circuit boards are deliberately designed with built-in components to alter the gain or frequency response and/or that those discrete op-amps cannot be used unconditionally in any existing circuit and remain stable.

I only glanced at the datasheet but the Sparkos stuff seems like a quite conventional "objectivist-friendly" design. It's always possible misusing it in the wrong circuit will cause issues but it's a fine audio op-amp.

 

[pkane]

Thanks. I only used the first 40 sec of the tracks to stay out of the part that’s clipping and got similar delta value. Suspicious indeed. The root cause remains unknown, but it's safe to conclude people shouldn't use these recordings to get a reliable impression of the effect of opamp swapping.

Right. Seems to me these op amps were not configured correctly in circuit, likely resulting in oscillations and frequency response artifacts. Level differences can be corrected for in comparison, but clipping and oscillations cannot.

 

[MaxwellsEq]

Seems to me these op amps were not configured correctly in circuit, likely resulting in oscillations and frequency response artifacts

Which is exactly why swapping op amps will, at best, have no benefit, but will far more likely make things worse. It's a stupid thing to do.