Listening Test Part 1. Passives.

[Mooly]

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.

That would be great Tony. Mabe try the "active test" too 🙂

 

[Mooly]

I have not concentrated on any detail. Just the overall tonality of the introductory bow part, that was all. But I have not tried headphones, I have one of my Quadral floor standing speakers 50cm next to my left ear 🙂

On Stax ES phones, I hear the same.

Dan.

Thats brilliant, and this really is the perfect opportunity to carry this kind of test forward.

Max, your Stax ES's should be very revealing. If you can choose one reference file and ABX that with a selection of the random ones... now that would move this kind of test into new territory.

 

[Mooly]

This is the component line up. The IC is for scale reference.



The two 4.7uf's were strung between another set of identical RCA sockets with the appropriate termination resistance. (Only reason I disassembled them was to string in another network for my own amusement for something different). You can see the electros are matched types between channels, and that some of them are the "cheap of the cheap" ultra miniature types of "generic" branding with no pretension to audiophile use at all.

 

[SY]

Did you experimentally confirm that the frequency responses were the same?

 

[Mooly]

Did you experimentally confirm that the frequency responses were the same?

Only with a scope and only at the lower frequency end. I used a Philips test CD as a signal source that goes down to 2Hz.

 

[wintermute]

Really struggling Mooly!! I thought I could hear a difference (with the focusrite and non usb headphones), then I started up the abx in foobar, first three attempts I got wrong! Now very much doubting that what I thought I could hear I actually could.

I can have a listen on my main system but probably not tonight. I found with pano's original interconnect test that I could pick up differences on my speakers that I couldn't on the headphones, only problem is I can't run foobar on my main system, as it is linux based...

Tony.

 

[Mooly]

Hi Tony,
listening on a really good set up is what this test demands really. Foobar with good headphones should be revealing though. See what you think on your main system 🙂

 

[fas42]

Really struggling Mooly!! I thought I could hear a difference (with the focusrite and non usb headphones), then I started up the abx in foobar, first three attempts I got wrong! Now very much doubting that what I thought I could hear I actually could.

It may be worth considering that the Foobar ABX software itself is veiling the differences - personally, I find there is too much loss of clarity using this processing, on my setup ...

 

[PMA]

It may be worth considering that the Foobar ABX software itself is veiling the differences - personally, I find there is too much loss of clarity using this processing, on my setup ...

With all respect, Frank, NO. Only in case of wrong settings or rather poor hardware used. I have it proven with numerous real tests and measurements.

 

[SY]

at the lower frequency end.

Cool, that was the part I was concerned about. Really nice job here- it's great to see someone take the trouble to try to actually do this right.

 

[PMA]

I am sure we have had more tests that were totally aligned both in amplitude and frequency response, even more that this one. Namely, mp3 x wav and 16 bit x 24 bit tests. You just have not participated. Would you mind posting your ABX screenshot result, for this last Mooly's test?

 

[SY]

I am sure we have had more tests that were totally aligned both in amplitude and frequency response, even more that this one. Namely, mp3 x wav and 16 bit x 24 bit tests.

But the test structure was poor and badly controlled because people didn't want to bother doing the legwork of independently creating and randomizing files.

 

[Mooly]

Cool, that was the part I was concerned about. Really nice job here- it's great to see someone take the trouble to try to actually do this right.

Thanks SY. There is more than a few hours work gone into producing these files... the methodology of the test structure though is something I can't take credit for... so thanks 🙂

 

[PMA]

But the test structure was poor and badly controlled because people didn't want to bother doing the legwork of independently creating and randomizing files.

Why would you think so? They were asked, same as here, to post their ABX screenshots. You are referring to previous, older tests. This is off topic here, I will send you links to the files mp3 x wav and 16 x 24 bit. I am looking forward ABX results.

 

[AudioLapDance]

It may be worth considering that the Foobar ABX software itself is veiling the differences - personally, I find there is too much loss of clarity using this processing, on my setup ...

With all respect, Frank, NO. Only in case of wrong settings or rather poor hardware used. I have it proven with numerous real tests and measurements.

I recall that Fas42 / Frank is using his on board sound card or a slot sound card, so it's more feasible that this could be happening with his setup.

I'm assuming that most of us use outboard DACs, either SPDIF (gasp) or USB.

Frank, maybe you could detail your computer setup as compared to say PMA / Pavel. Not to bash but to compare.

Cheers,
Jeff

PS Itss tough listening , I have to get my AKG K301 headphones out...

 

[PMA]

Karl, why not to use 24-bit files for a test that is supposed to have high resolution? You never enter the same river twice, as you know.

 

[SY]

I am looking forward ABX results.

That's nice, but this isn't an ABX test and can't be converted into one and have any validity. That was just a suggested way to begin, it's not the ending. You might want to give my Linear Audio article on DBT test formats a read.

You can either sort these or you can't. There's nothing wrong with a null result, that's actually useful data.

I assume that as a very skilled engineer, you know the difference between resolution and noise floor, so the 24 bit stuff is just careless use of terminology.

 

[Max Headroom]

Thats brilliant, and this really is the perfect opportunity
to carry this kind of test forward.

Yes, I think a discussion of the nature of the perceived differences between particular pairs of files would be most useful.
This allows us to confer and draw attention to particular characteristics that may be missed by any particular participant, using any particular gear.
So everybody collaborate so that all participants are well able to discern and identify the fine differences between the recordings.
IMHO, this training makes for more reliable ABX testing results.

Max, your Stax ES's should be very revealing. If you can choose one reference file and ABX that with a selection of the random ones... now that would move this kind of test into new territory.

Yes, they are very nicely revealing, but they won't glare on anything either.
They are essentially distortionless, and can go seriously big/loud.

So, whilst being very nicely precise, they can also be in a sense a little forgiving of the source wrt to higher distortion transducers/systems.
That's when less clean playback amp/loudspeaker combos can add glare/spit/resonances etc and therefore make some source recordings differences rather more apparent.

Dan.

 

[SY]

Dan, it's not an ABX test, it's a sorting test.

Discuss what you perceive as sonic differences between A and B all you like, but try not to give away your answers and sabotage Mooly's efforts to do things in a better controlled manner than most of the other "tests" I've seen posted here.

 

[PMA]

A or B files are not same to any of the further 12 files. There are ALWAYS differences. The difference can be seen both in frequency domain and sample values. Attached is the difference in 1st 3 second between A and B in frequency domain. In time domain, "B" is a bit compressed compared to "A".

I can only say that the files sound same if they are same. I assume I have proven the difference by my ABX result. How can we 'pair' A or B to other files if any of them is different???