Another big opamp listening test

PMA · 2017-06-02 1:42 pm

Hello all, I have prepared another listening tests, with 4 very different and well known opamps. At the moment, I do not want to disclose any technical details, just that the test files are both time and length matched. Please feel free to participate in the test and vote which of the files you prefer, by listening. You may also try to make a foobar ABX test, if you like, to test if you are able to find differences in the double blind test. Have fun!

http://pmacura.cz/opamptest.zip

==================================

The poll has just been closed so it is time to disclose opamps under test:

aa - Tesla MA1458, equivalent of LM1458 (dual LM741), http://www.teslakatalog.cz/MA1458.html

bb - National Semiconductor LM4562

cc - SGS Thomson TL072

dd - Burr Brown OPA2134

More about test conditions in my web page: http://pmacura.cz/anothertest.htm

Thanks to everyone who participated in the test and the thread!
====================================================
Now you can try a new test with opamps, this time in 96kHz/24bit resolution

http://www.diyaudio.com/forums/everything-else/309098-hires-96-24-listening-test-opamps.html

ErnieM · 2017-06-02 2:59 pm

Doubt I can tell a difference. Should be fun. Thanks.

mjf · 2017-06-02 6:26 pm

tiny differences........hard to choose

Mooly · 2017-06-02 6:42 pm

Thanks Pavel. I hope to give these a listen over the weekend.

PMA · 2017-06-02 9:58 pm

Thank you guys for giving it a try

rayma · 2017-06-02 10:19 pm

PMA said:
Thank you guys for giving it a try

On my computer speakers [Proton 300 stereo radio], there are clear differences among the four, with B preferred.

lduarte1973 · 2017-06-03 11:56 am

B too , warmer sound and more bass , opa 2604 ?

PMA · 2017-06-03 12:23 pm

Thank you, but we have with disclosure to wait till the end of the poll - 2 weeks.

PMA · 2017-06-04 9:28 am

This is a listening test and many of you may remember we had posts with file analysis in the past that degraded the test result.

This time, the files are so well matched in time, amplitude and frequency response, that I would have no objections if you wanted to make file analysis. Despite of that, I am sure that the only hint for you will be the subjective sound preference.

jan.didden · 2017-06-04 9:33 am

Good try, but there are always morons that don't understand blind testing and start to give preferences and even opamp types they think. No longer blind. Too bad for all your efforts PMA.

Jan

PMA · 2017-06-04 9:49 am

I still would like to gain something positive and believe I will

. So everyone is welcomed to make the test and vote.

Mooly · 2017-06-04 1:02 pm

I've given these files three good listens both last night and today. There are a lot of audible cues (or clues that could be a decider

) but in practice differentiating one file from another subjectively is proving extremely difficult.

PMA · 2017-06-04 1:08 pm

Good to hear from you Karl, thank you!

PMA · 2017-06-05 7:00 am

To encourage more testers, I would add a small test description.

4 different opamps are tested, wired as non-inverting unity gain buffers. The circuit is built on a standard double-sided FR4 PCB and it is used as a line stage in my AB box testing equipment. Blue Alps potentiometers are followed with these unity gain buffers. The whole thing is inside well shielded Al box with 3mm bottom and cover plates. Opamps are in sockets so they can be easily interchanged. Complex set of measurements was done to prove there are no additional disturbing noises or oscillations in the circuit.

All the tested opamps have had exactly same conditions and there is no "cheating".

silverprout · 2017-06-05 8:02 am

Sorry i can't answer, my brain is biased by the posts 6 and 7

PS : it should be pehaps interesting to hide some information in these posts.

PMA · 2017-06-05 8:25 am

silverprout said:
Sorry i can't answer, my brain is biased by the posts 6 and 7

PS : it should be pehaps interesting to hide some information in these posts.

I understand, but I have no chance to hide the posts. Maybe just try to ignore it, and though I should not comment, I can tell you that the guess of the type was wrong, it was just a guess.

Markw4 · 2017-06-05 9:04 am

PMA said:
To encourage more testers, I would add a small test description.

4 different opamps are tested, wired as non-inverting unity gain buffers. The circuit is built on a standard double-sided FR4 PCB and it is used as a line stage in my AB box testing equipment. Blue Alps potentiometers are followed with these unity gain buffers. The whole thing is inside well shielded Al box with 3mm bottom and cover plates. Opamps are in sockets so they can be easily interchanged. Complex set of measurements was done to prove there are no additional disturbing noises or oscillations in the circuit.

All the tested opamps have had exactly same conditions and there is no "cheating".

Okay, but what data converters are being used? And what is the starting resolution of the source material? There is more to what is being listened to than the opamps alone. Although the other factors might be the same for every opamp, they still could affect listener perceptions.

Therefore, I would be in favor of posting (1)the original sound file, and (2)the original sound file passing through the data converters only, and not passing through any test opamps.

PMA · 2017-06-05 9:25 am

Markw4 said:
Okay, but what data converters are being used? And what is the starting resolution of the source material? There is more to what is being listened to than the opamps alone. Although the other factors might be the same for every opamp, they still could affect listener perceptions.

Therefore, I would be in favor of posting (1)the original sound file, and (2)the original sound file passing through the data converters only, and not passing through any test opamps.

Hello Mark,

thank you for your comments, I appreciate them, however I assume it would be counterproductive if I answered. The poll is valid until June 16 and then we may disclose more. Your questions are and the discussion might be quite complex (and maybe might open new biases as well).

What I found in my previous tests is the fact that the files must be bullet-proof regarding consistency of test conditions and time and amplitude matching. Once the files are in public, they are usually analyzed as data rather then only listened. I can only disclose that I made no manipulation in the digital domain with the files, except for throwing away the samples at the beginning and at the end to get exactly same length and starting sample of all the files.

All I wanted to say in my previous post is that I am not trying to cheat the participants, in this particular test.

Mooly · 2017-06-05 9:27 am

The source material is interesting... as I mentioned in my first post, a lot of audible 'clues' buried deep down in there.

I don't know how much Pavel wants this to turn into a discussion, or whether its preferred to just listen and tick the box, with discussion later

silverprout · 2017-06-05 10:32 am

PMA said:
I understand, but I have no chance to hide the posts. Maybe just try to ignore it, and though I should not comment, I can tell you that the guess of the type was wrong, it was just a guess.

My heart is ok to ignore it, but my brain still consider it :yinyang:

List of culprits :
https://en.wikipedia.org/wiki/List_of_cognitive_biases

Main suspect involved in the posts 6 and 7.
https://en.wikipedia.org/wiki/Confirmation_bias

Another big opamp listening test

Which of the files you prefer by listening?

aa

bb

cc

dd

Member

Member

Member

Administrator

Member

Moderator

Banned

Member

Member

AX tech editor

Member

Administrator

Member

Member

Attachments

Member

Member

Member

Member

Administrator

Member