Archimago has a new DAC listening test

[krabapple]

Seems quite clear to me that they are hires downsampled to 16/44. Although I'm not sure where does the 24/88k version of Stravinsky come from. The Reference Recording site offers only 16/44k and 24/176k. Later in the post he writes:

Four of the original six AMPT Test tracks from 2021 are CD tracks (i.e., 16/44). The other two are from hi rez sources; these were downconverted to 16/44 for the 2021 listening test.

He also writes there:

I guess that for the blind test he left them at 24-bit to avoid complaints that 16-bit was the reason for not hearing differences

For the 2024 listening test , the six CD-rate tracks were upconverted to 24 bit, the other two left at 24 bit, again for consistency's sake. (I don't know if the sample rates were normalized too).

This is how I interpret the protocol.

I can't imagine a 16-->24 conversion having any audible effect unless something was broken. 24-->16 could have an audible effect if done incompetently.

 

[danadam]

Four of the original six AMPT Test tracks from 2021 are CD tracks (i.e., 16/44). The other two are from hi rez sources; these were downconverted to 16/44 for the 2021 listening test.

AFAIU the AMPT tracks weren't created for any specific listening test but rather to be a usual part of his DACs' reviews.

 

[krabapple]

AFAIU the AMPT tracks weren't created for any specific listening test but rather to be a usual part of his DACs' review

Arch specifically wrote "this time" he would use 24 bit versions, so they've been used before in their 2021 16/44 versions.

 

[board]

That's not remarkable. You did the equivalent of raising the noise floor, so if there's a quite enough passage and you crank the level high enough...

Discontinuity at the end the track yields audible ticks.

I know this is an old topic by now, but it wasn't until now that I got around to doing an ABX test after making a new version of the two files I originally ABX'ed (Mozart mixed with the black metal band Dark Funeral at -70 dB), so if you're up for it, I would be very happy to hear if you, or anyone else, would be able to ABX these new files, where I've tried to remove the clicks at the end.



https://mega.nz/folder/IVZBEbqb#llqi3-Tzn5R12UajgN9beQ



If you find another tell, please let me know what it is and where.
Oddly, in the original versions that you successfully ABX'ed I couldn't hear any clicks when playing the songs, but with a null-test in Audacity I could hear a click at the end. So, hopefully that has been removed now (if not, please let me know).

My new ABX test wasn't quite as good as the last one. I feel like I've had a bit of an ear ache for a while, but maybe my hearing has simply become a bit worse in the last two months.
If you can't successfully ABX them, nor hear any tells, I would be happy to reveal where I was able to hear Dark Funeral under the Mozart music (because that was what I was listening for). It wasn't in the most obvious place, actually, but it was certainly difficult to make out.
Anyway, here's the result:




foo_abx 2.0.6d report
foobar2000 v1.6.7
2024-07-29 21:12:53

File A: Martin Pearlman - Boston Baroque - Mozart Requiem (Completion by Robert Levin) [8] - med Dark Fungeral ved 2 sekunder, -70 dB mixet ind, begge sange peaket ved 0 før mix - begge faded.wav
SHA1: df504045c0d88c0181b5a1ee144033f54c9a236e
File B: Martin Pearlman - Boston Baroque - Mozart Requiem (Completion by Robert Levin) [8] - normalized - faded.wav
SHA1: d00df1c6971a2553ab67e501352f9864f9680bfa

Output:
Default : Speakers (ODAC-revB USB DAC)
Crossfading: YES

21:12:53 : Test started.
21:13:46 : 01/01
21:14:19 : 02/02
21:15:12 : 02/03
21:15:37 : 03/04
21:17:03 : 04/05
21:20:07 : 05/06
21:21:56 : 06/07
21:22:17 : 07/08
21:22:40 : 08/09
21:24:52 : 09/10
21:29:56 : 10/11
21:32:47 : 11/12
21:36:35 : 11/13
21:40:31 : 12/14
21:40:50 : 13/15
21:43:46 : 14/16
21:43:46 : Test finished.

----------
Total: 14/16
p-value: 0.0021 (0.21%)

-- signature --
c0ba92c38614af7a1d868bf22e16e1d772e35fdd

 

[board]

On JP's suggestion I will now reveal what I heard and where, in the hopes that others will then give an ABX test a go:
In the section around 0:23 I could indeed hear Dark Funeral underneath the Mozart music.
Are others able to hear the same?

I will be the first one to admit that this test has been mostly academic and very case-specific, but nevertheless it showed that another signal mixed in at -70 dB can in specific cases be audible underneath the music.