These two listening tests (this test is for passives, the other thread is HERE and is for powered active circuitry) are an attempt to see if the listeners (that's you 
) can detect any audible differences between the two reference files in each set. Unlike the other tests there is a bit more to this one.
The links to the files are at the end of this post.
How the test works.
This is conducted very differently to all the other tests that have been run on the forum. Firstly, I am going to tell you what you are listening to. The passives test is a 4.7uf Shizuki film cap vs a chain of five cheap electroylitics in series. The circuit is as shown here and you can see that given ideal cap values the AC performance of each is virtually identical.
The response of the two circuits track each other with different loadings and source impedance's (such as the A/D convertor used to produce the files and the D/A convertor used to provide the input signal).
I would suggest you listen to the two reference files first, A and B (listen, or use something like Foobar to do an ABX test. Whatever you are happy with). One of these files is the film cap, the other the electrolytic.
If you think you can hear a difference, well that's great, but that isn't statistically good enough. Having identified a difference (if you think there is
one) then you should be able to match that difference to the rest of the files presented, Test01, Test02 etc. It would also be good if you could say which you prefer sonically, either A or B.
So what are the actual files.
As explained, there are two reference files, A and B. One is the film cap, one the electrolytic chain. The other twelve files are a mixture where any one file could be either the film or the electrolyitic. To produce the files an online random number generator was used (random.org) and this was used to produce a random "key" such as A, A, B, A, B, B, etc which the files were then created against.
Recording the files.
Each file was recorded in its entirety. This is important. The files are NOT COPIES of each other. After each file run, the connections to the cap networks were removed and remade, also swapping the networks if the random key denoted a change. All the files were then cropped to the same bit length and all have had the timestamps equalised. This means the files will look identical if you right click them and check the file properties.
The easy bit.
Just listen and let your ears decide. Because that is what this kind of test is all about.
The results.
So that we get as much interest as possible it would be good to see shots of your ABX results of the two reference files. That's the fun part and they can be posted in the thread. If you identify a preference for one over the other then a pm to me is best. Please mark any results by pm clearly as "PASSIVES" to easily differentiate the two tests. Such comments on the thread influence others. I would collate the results to post for all to see later. If you don't want your name associated with your choice then just say so and you will be marked "anonymous". And for those of you that can really really tell a difference then the above results should be backed up by identification of the random files. That last part is the most time consuming but is also the only real test of the validity of your results. With two files there is a 50/50 chance you can guess the one you might hope for.
So a pm something like,
PASSIVES, and keep me anonymous (if you wish). A sounds best because of wxyz, and I think A is the same as Test05, Test09 etc etc would be brilliant.
And above all.
ENJOY
Links,
https://www.dropbox.com/s/vndg41ujzw5091e/A.wav
https://www.dropbox.com/s/ger74r7f01vqb9b/B.wav
https://www.dropbox.com/s/qkxkxo3da0v4uu7/Test01.wav
https://www.dropbox.com/s/b9eoeb7lnwviaw6/Test02.wav
https://www.dropbox.com/s/dwf8wt2rxk8iqp8/Test03.wav
https://www.dropbox.com/s/38eowfts5pog0ya/Test04.wav
https://www.dropbox.com/s/4hus4fuz7hxwnkb/Test05.wav
https://www.dropbox.com/s/icdgcd9z0c5ml8m/Test06.wav
https://www.dropbox.com/s/31gmzgfdhtl0ka7/Test07.wav
https://www.dropbox.com/s/i100sdhbs0lilvq/Test08.wav
https://www.dropbox.com/s/2vo2mul7wai3h6u/Test09.wav
https://www.dropbox.com/s/4rr7l4794cwc8rw/Test10.wav
https://www.dropbox.com/s/3yr16gffyo0d0or/Test11.wav
https://www.dropbox.com/s/s20somb6g883tmi/Test12.wav
) can detect any audible differences between the two reference files in each set. Unlike the other tests there is a bit more to this one.The links to the files are at the end of this post.
How the test works.
This is conducted very differently to all the other tests that have been run on the forum. Firstly, I am going to tell you what you are listening to. The passives test is a 4.7uf Shizuki film cap vs a chain of five cheap electroylitics in series. The circuit is as shown here and you can see that given ideal cap values the AC performance of each is virtually identical.
The response of the two circuits track each other with different loadings and source impedance's (such as the A/D convertor used to produce the files and the D/A convertor used to provide the input signal).
I would suggest you listen to the two reference files first, A and B (listen, or use something like Foobar to do an ABX test. Whatever you are happy with). One of these files is the film cap, the other the electrolytic.
If you think you can hear a difference, well that's great, but that isn't statistically good enough. Having identified a difference (if you think there is
one) then you should be able to match that difference to the rest of the files presented, Test01, Test02 etc. It would also be good if you could say which you prefer sonically, either A or B.
So what are the actual files.
As explained, there are two reference files, A and B. One is the film cap, one the electrolytic chain. The other twelve files are a mixture where any one file could be either the film or the electrolyitic. To produce the files an online random number generator was used (random.org) and this was used to produce a random "key" such as A, A, B, A, B, B, etc which the files were then created against.
Recording the files.
Each file was recorded in its entirety. This is important. The files are NOT COPIES of each other. After each file run, the connections to the cap networks were removed and remade, also swapping the networks if the random key denoted a change. All the files were then cropped to the same bit length and all have had the timestamps equalised. This means the files will look identical if you right click them and check the file properties.
The easy bit.
Just listen and let your ears decide. Because that is what this kind of test is all about.
The results.
So that we get as much interest as possible it would be good to see shots of your ABX results of the two reference files. That's the fun part and they can be posted in the thread. If you identify a preference for one over the other then a pm to me is best. Please mark any results by pm clearly as "PASSIVES" to easily differentiate the two tests. Such comments on the thread influence others. I would collate the results to post for all to see later. If you don't want your name associated with your choice then just say so and you will be marked "anonymous". And for those of you that can really really tell a difference then the above results should be backed up by identification of the random files. That last part is the most time consuming but is also the only real test of the validity of your results. With two files there is a 50/50 chance you can guess the one you might hope for.
So a pm something like,
PASSIVES, and keep me anonymous (if you wish). A sounds best because of wxyz, and I think A is the same as Test05, Test09 etc etc would be brilliant.
And above all.
ENJOY
Links,
https://www.dropbox.com/s/vndg41ujzw5091e/A.wav
https://www.dropbox.com/s/ger74r7f01vqb9b/B.wav
https://www.dropbox.com/s/qkxkxo3da0v4uu7/Test01.wav
https://www.dropbox.com/s/b9eoeb7lnwviaw6/Test02.wav
https://www.dropbox.com/s/dwf8wt2rxk8iqp8/Test03.wav
https://www.dropbox.com/s/38eowfts5pog0ya/Test04.wav
https://www.dropbox.com/s/4hus4fuz7hxwnkb/Test05.wav
https://www.dropbox.com/s/icdgcd9z0c5ml8m/Test06.wav
https://www.dropbox.com/s/31gmzgfdhtl0ka7/Test07.wav
https://www.dropbox.com/s/i100sdhbs0lilvq/Test08.wav
https://www.dropbox.com/s/2vo2mul7wai3h6u/Test09.wav
https://www.dropbox.com/s/4rr7l4794cwc8rw/Test10.wav
https://www.dropbox.com/s/3yr16gffyo0d0or/Test11.wav
https://www.dropbox.com/s/s20somb6g883tmi/Test12.wav
Last edited by a moderator:
Mooly,
What is real performance of your signal chain?
What is distortion of the various circuits?
What is IM performance of two closely spaced tone, such as 440Hz with 450Hz?
It is very easy to concoct very time consuming test protocols.
No commercial designer is going to increase component count and circuit board area unless significant improvements are seen in measurements.
What is real performance of your signal chain?
What is distortion of the various circuits?
What is IM performance of two closely spaced tone, such as 440Hz with 450Hz?
It is very easy to concoct very time consuming test protocols.
No commercial designer is going to increase component count and circuit board area unless significant improvements are seen in measurements.
The performance, the distortion and IM distortion is a "constant" in exactly the same way it is when an electro is swapped for an expensive film in an amplifier, or when an opamp change is made in some piece of equipment.
There is no change in component count in any of these tests. An opamp for an opamp and a film for an electrolytic. In the case of the electrolytics five were chained together to really highlight differences... if they exist.
Thanks for the interest but the tests are what they are, and hopefully reflect the same kind of scenario as replacing or specifying one component with another in a typical unit, be that a diy'er looking for an upgrade or a designer looking to specify a specific part. We are often led to believe that hi end audio is as much an art as a science with components proving "difficult" to measure but that the audible change is clear to be heard.
A wire would be obvious when the files were tested with software with LF phase shift occurring.
This test is surely of extremes, five cheap caps vs a film cap. The final termination resistor is 56k so a small "current" is pulled through the caps hopefully highlighting non linearity's and distortion mechanisms. The test is simple. Can you identify one vs the other
A wire would be obvious when the files were tested with software with LF phase shift occurring.
This test is surely of extremes, five cheap caps vs a film cap. The final termination resistor is 56k so a small "current" is pulled through the caps hopefully highlighting non linearity's and distortion mechanisms. The test is simple. Can you identify one vs the other
Please provide measurements that prove this.
Dielectric performance and microphonics of different capacitors impact distortion performance.
Phase shift in short wire in audio band? Nonsense.
Well, if that's true (and there's a good reason that it wouldn't be), people should be able to tell one from the other, ears-only. And that's the point of the test. Analogous to Pano's wire-versus-mud test, one can then put claims of huge audible differences between brands or dielectrics of coupling capacitors into perspective.
The "good reason" is that, in the audio band, the voltage across the cap is almost zero, so even a cap with high distortion won't significantly affect the signal. This is why you can have amps with electrolytic input caps that still show 0.00x% distortion.
Phase shift in short wire vs either cap/s in the test... significant. The fact that so many analyse via software rather than ears renders inclusion of a wire vs cap as null and void.
The test is simple... just listen
I have no means or equipment available to measure down to the levels that would be needed for what you mention... but then neither do most others... yet the majority immediately claim to be able to "hear" the differences between various components.
Now we have a test that is, I believe, conducted correctly in eliminating any possibility of guessing... and... as had been pointed out to me... the silence is, well lets say deafening.
Sorry, haven't "tuned" into doing this test so far - life and all sort of thing ...
Have gone so far as to download A and B, and about half the test samples. The characteristics of A and B are quite distinctive, and, for me, line up with the described differences in the 2 setups. Put the first sample into one camp ... and that's as far as I got, ...
Will try to do the rest in the next day or so - and post off the results ...
Have gone so far as to download A and B, and about half the test samples. The characteristics of A and B are quite distinctive, and, for me, line up with the described differences in the 2 setups. Put the first sample into one camp ... and that's as far as I got,
...Will try to do the rest in the next day or so - and post off the results ...
Hi Karl,
I gave it a very, very short try. My flat is being reconstructed. So I have only ONE speaker in my small office. The floor-standing speaker stands 50 cm from my left ear, but is driven with my hi-end preamp and power amplifier. I am listening to left channel only. This is my result, 1st take, no repetitions, no re-tries.
I gave it a very, very short try. My flat is being reconstructed. So I have only ONE speaker in my small office. The floor-standing speaker stands 50 cm from my left ear, but is driven with my hi-end preamp and power amplifier
. I am listening to left channel only. This is my result, 1st take, no repetitions, no re-tries.Attachments
Hi Karl,
I gave it a very, very short try. My flat is being reconstructed. So I have only ONE speaker in my small office. The floor-standing speaker stands 50 cm from my left ear, but is driven with my hi-end preamp and power amplifier
Hi Pavel, so your final answer then, you still have phone a friend or ask the audience
Firstly and seriously though, thanks for taking part, and I must say that is a mighty impressive ABX result... but what we would really need to do now is follow that through and see if the audible differences you hear are as easy to pick out in the random files. It is on that result where a test such as this "stands or falls".
Great result though. I hope that encourages others to give this a try.
I've just had a listen with my cheap logitech usb headphones on my laptop (which are the best I have), and I definitely at this stage wouldn't try an ABX with foobar because I can't pick up on anything that I could use as a differentiator. I might try with a different set of (equally cheap) analogue headphones with my focusrite 2i2 and see if that makes any difference
Tony.
Tony.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.