Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Succesful ABX of 24/96 vs. 16/44.1 (Read 34948 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Succesful ABX of 24/96 vs. 16/44.1

Reply #25
I tried these samples with my HD-590 and X-Fi XtremeMusic in Win7, playback 24/96, audio producing mode.

foo_abx 1.3.4 report
foobar2000 v1.1.8
2011/10/20 01:35:37

File A: C:\testfiles\16_441to24_96.wav
File B: C:\testfiles\24_96.wav

01:35:37 : Test started.
01:36:17 : 00/01  100.0%
01:36:23 : Trial reset.
01:38:58 : 01/01  50.0%
01:39:39 : 02/02  25.0%
01:40:15 : 03/03  12.5%
01:41:11 : 04/04  6.3%
01:42:38 : 05/05  3.1%
01:42:43 : Test finished.

----------
Total: 5/6 (10.9%)

Not perfect and i really need a quiete room at night  + a very good day i sadly have to few of atm. I didn´t try for to long. On the re-resample i hear the high note synths not as clean as with the native 96kHz resample in the moments the bass beats come.
I listened second 11 to 14 in loop.
This is quite surprising since i did lately tons of imho serious listening to find a resample setting i want to use with sox. I for myself ended up using not the sox recommendation. I use a slower filter, allow aliasing above the passband and use the lower shaped shibata dither. Maybe it now comes out this wasn´t a to silly choice  Of cause i don´t want to enter audiophile territory with this claim but if anyone wants to try it this way when he thinks sox standard has problems and may report here?

More likely is we have a problem here with applying 2x dither and 2 times a lowpass at exactly the same spot. One time on downsampling and one time on upsampling.
Like said before, on all my tries before i only tried some 24/96 directly to 16/44.1 and did not find any problem.

Edit: i never realy understood how the X-Fi works in win7 and what other evil things may interact with it, so you may have an idea. Also i didn´t find a working link to the 192kHz original.
I also want to mention i once failed to abx a prepared sample from 2bdecided of "Jazz t the Pawnshop" that he simply created with applying a lowpass to the 96kHz file simulating a resample.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Succesful ABX of 24/96 vs. 16/44.1

Reply #26
I find your position plausible in general but have had less ABX success than you. What equipment do you use? Are you successful with both loudspeakers and headphones? Are these samples you have made yourself, and how are you downsampling?


Sennheiser HD 650 and EMU Pre Tracker USB or onboard sound card HD Realtek 24/192.  I'm not into loudspeakers. The samples are from OP and also different variations SOX/SSRC with/without foobar's dither.

On the re-resample i hear the high note synths not as clean as with the native 96kHz resample in the moments the bass beats come.

That's basicly the description of my experience too.

Succesful ABX of 24/96 vs. 16/44.1

Reply #27
Now that baffled me.
I loaded the 2 samples into Audio Diffmaker. It seldom gives a real hint on differences and how audible they are but the difference here is that non-existand we really must have another problem.
There is really only dither noise as difference below 21kHz. I can upload the result file if interest. The thread the original samples were offered is locked.
Only plausible thing to me is the logic Arnold already mentioned. I only want to know where this problem creeps in.
Since there is noise we hear on the file without highs the playback must add some nasty things that it doesn´t when there are highs. Maybe an imperfect resample even when setting none takes place we can´t influence that is really borked.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Succesful ABX of 24/96 vs. 16/44.1

Reply #28
I did a quick test and couldn't ABX them.
But for those who can easily hear it, I'd like to propose testing 24/44.1 instead of 16/44.1, just to see if the difference is due to SRC or to bit reduction/dithering.

And eventually 16/96 vs 16/44.1.

And perhaps with a different resampler.

 

Succesful ABX of 24/96 vs. 16/44.1

Reply #29
Let's eliminate resampling completely to begin with:

Code: [Select]
sox 24_96.flac lpf.wav sinc -21k spectrogram -h -o lpf.png
sox 24_96.flac hpf.wav sinc 21k spectrogram -h -o hpf.png

Theory says that if lpf.wav sounds different to the original, you either have golden ears, or problems with your playback chain.
Similarly, if hpf.wav is audible at all.

Also try Monty's Intermod Tests here: http://xiph.org/~xiphmont/demo/neil-young.html

Succesful ABX of 24/96 vs. 16/44.1

Reply #30
Let's eliminate resampling completely to begin with:

Code: [Select]
sox 24_96.flac lpf.wav sinc -21k spectrogram -h -o lpf.png
sox 24_96.flac hpf.wav sinc 21k spectrogram -h -o hpf.png

Theory says that if lpf.wav sounds different to the original, you either have golden ears, or problems with your playback chain.
Similarly, if hpf.wav is audible at all.

Also try Monty's Intermod Tests here: http://xiph.org/~xiphmont/demo/neil-young.html



Good point. IM  can cause false positives in bandpass and downsampling tests, and is not always inaudible in low cost audio interfaces.

Succesful ABX of 24/96 vs. 16/44.1

Reply #31
I must have overlooked these intermod test samples. When i play these warbles-96.wav there is no content at all below ~23khz even deep in the noisefloor. When recording the output i see small spikes in the recording at ~21kHz that were not there before. This is at very low volume and is not audible to me. Maybe these effects can indeed influence listening alone for the fact we have high frequency content. This still is not explaining the listening tests of before but shows my system adds own sounds.
After all this is what the monty article want to teach us
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!