Quicktime TVBR vs Nero AAC Options

[list]

sory, but i am new in here ^^ .
I didn't know how a ABX test is.
So i tried it with a foobar2k abx component with a sample at 128 and got 13/15.
is that worth?

This post has been edited by list: Feb 9 2011, 19:14

[greynol]

www.mediafire.com/download.php?ffj8cm332bq7a7e

This link doesn't work.

[list]

sources+log www.mediafire.com/download.php?7xq5qsas9kky3lk
is the foo_abx component the right software for testing?

This post has been edited by list: Feb 9 2011, 19:33

[greynol]

Thanks.

Are you able to ABX the QT version from the lossless version?

[list]

Are you able to ABX the QT version from the lossless version?

? I made ABX between nero@128 and QT@128. Is that right?

[greynol]

ABX is supposed to be performed between lossless and lossy. If neither version is transparent to you then you need to use ABC(/HR) instead of ABX.

If you cannot tell the difference between QT and lossless but can tell the difference between nero and lossless then TOS #8 has been satisfied.

[list]

I tried abx to highlight the existing difference from one to other.
Does it mean something?

[greynol]

It means you can distinguish one from the other. It does not directly demonstrate which one sounds closer to the original.

[IgorC]

Hola, List.
Try this http://ff123.net/abchr/abchr.html

The easy way to use it is to feed the comparator with original and decoded lossy files codec_A.wav, codec_B.wav and you are ready for blind test ( Original vs Codec_A, Original vs Codec B). It's also good to do blind test on Codec_A vs Codec_B to be sure that you actually hear the difference.

So i tried it with a foobar2k abx component with a sample at 128 and got 13/15.
is that worth?

While the probability values (in %) is less than 5% your results are valid.

See the table of probability values here http://en.wikipedia.org/wiki/ABX_test

This post has been edited by IgorC: Feb 9 2011, 20:39

[punkrockdude]

I also find Quicktime to be a lot better than Nero around 128kbps. It's easy to hear when you compare so that's why I haven't done an ABX.

ABX tests are needed, it doesn't matter how "easy" you think it is to hear as you could very well be suffering from the placebo affect. Proper testing is required before anyone makes subjective audio claims, period. That is in the TOS (specifically TOS #8) that everyone agrees upon when joining the site. So proper test results are needed from both you and list before any audio quality claims can be taken seriously. That and the Nero devs would like to know the performance of their encoder and just what samples their encoder is having issues with. That way they can fine tune their encoder so that Apple's is not "much better" and comparing the two won't be so "easy to hear."

foo_abx 1.3.4 report
foobar2000 v1.1.2
2011/02/09 22:37:54

File A: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 prodigy\prodigy nero.m4a
File B: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 prodigy\prodigy.wv

22:37:54 : Test started.
22:38:41 : 01/01 50.0%
22:39:17 : 02/02 25.0%
22:40:05 : 03/03 12.5%
22:40:57 : 04/04 6.3%
22:41:45 : 05/05 3.1%
22:42:09 : 06/06 1.6%
22:42:46 : 07/07 0.8%
22:43:09 : 08/08 0.4%
22:43:49 : 09/09 0.2%
22:44:14 : 10/10 0.1%
22:44:19 : Test finished.

----------
Total: 10/10 (0.1%)

foo_abx 1.3.4 report
foobar2000 v1.1.2
2011/02/09 22:45:50

File A: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 prodigy\prodigy qt.m4a
File B: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 prodigy\prodigy.wv

22:45:50 : Test started.
22:46:32 : 01/01 50.0%
22:46:58 : 01/02 75.0%
22:47:29 : 02/03 50.0%
22:47:38 : 03/04 31.3%
22:47:51 : 04/05 18.8%
22:48:19 : 05/06 10.9%
22:48:51 : 06/07 6.3%
22:49:38 : 07/08 3.5%
22:50:40 : 07/09 9.0%
22:51:34 : 08/10 5.5%
22:51:48 : 09/11 3.3%
22:52:32 : 10/12 1.9%
22:52:36 : Test finished.

----------
Total: 10/12 (1.9%)

[punkrockdude]

I am going to try the classical piece some day but i guess it will be a lot tougher from the little glim that I had at them. Regards. The prodigy track with nero had more high frequencies and the hi hat seemed more open (like not pressing down the pedal as much) and degraded. Quicktime had less high frequencies but more focused hi hat. The hi hat "sustain" seemed to end faster than with Nero. Maybe Nero can archieve about the same quality if a lower low pass is used? Regards.

[punkrockdude]

I tried ogg vorbis which i favour and it was a little more difficult at times to destingish (spelling?) from the lossless. I used the following settings and encoder and the averge bitrate is 130kbps:
BS; LancerMod(SSE2) [Nov 25 2009] (based on aoTuV b5d [20090301]) (...)
-q 3.5 --advanced-encode-option impulse_noisetune=-7 --advanced-encode-option lowpass_frequency=99


foo_abx 1.3.4 report
foobar2000 v1.1.2
2011/02/09 23:04:44

File A: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 prodigy\prodigy.wv
File B: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 prodigy\prodigy.ogg

23:04:44 : Test started.
23:05:05 : 01/01 50.0%
23:05:31 : 02/02 25.0%
23:06:28 : 03/03 12.5%
23:06:54 : 03/04 31.3%
23:07:13 : 04/05 18.8%
23:07:26 : 05/06 10.9%
23:07:50 : 06/07 6.3%
23:08:07 : 07/08 3.5%
23:08:21 : 07/09 9.0%
23:09:07 : 08/10 5.5%
23:09:55 : 09/11 3.3%
23:10:08 : Test finished.

----------
Total: 9/11 (3.3%)

Sorry for some ABX test being different amount of times I abx:ed them. Sometimes I just forget how many I had done. Regards.

This post has been edited by punkrockdude: Feb 9 2011, 23:19

[punkrockdude]

I now saw on the wiki site that 16 attempts is "minimum" but my ears will kill me if I try to do the tests again. Sorry guys and gals. Regards.

This post has been edited by punkrockdude: Feb 9 2011, 23:23

[db1989]

It'd be interesting to know if you can detect such differences on many/most tracks in your library, or just certain ones?

[punkrockdude]

It'd be interesting to know if you can detect such differences on many/most tracks in your library, or just certain ones?

Well, I haven't ABXed the classical piece yet but from what the quick listen that I had they seem more difficult than the Prodigy tracks. I will try a quick ABX right away to see. The type of music that I listen to (Punk Rock, Metal, Blues) is usually quite easy to ABX (please don't tell me to ABX my whole collection, hehe). Regards.

[list]

I tried the same sample with QT at 192kbps and it was hard to hear the difference from the losless:

foo_abx 1.3.4 report
foobar2000 v1.1
2011/02/09 19:39:16

File A: E:\Documents and Settings\Administrador\Escritorio\prodigy.wv
File B: E:\Documents and Settings\Administrador\Escritorio\prodigy qt 192.m4a

19:39:16 : Test started.
19:39:39 : 01/01 50.0%
19:40:07 : 02/02 25.0%
19:40:31 : 03/03 12.5%
19:40:52 : 04/04 6.3%
19:41:20 : 05/05 3.1%
19:41:54 : 06/06 1.6%
19:42:42 : 07/07 0.8%
19:43:32 : 08/08 0.4%
19:44:54 : 08/09 2.0%
19:45:18 : 09/10 1.1%
19:45:26 : Test finished.

----------
Total: 9/10 (1.1%)

[punkrockdude]

Here is my ABX of the classical (ludwig piece). When I found out what to listen for it all got a lot easier but I found it out when I was doing the QT comparison which was the last on. I think Quicktimes low pass gave this one away. It was the violin at the end at the forte fizzo where the last percent/sustain of notes had clearly less high frequencies. Maybe I will try to ABX Nero again now when I know what to look for.

foo_abx 1.3.4 report
foobar2000 v1.1.2
2011/02/09 23:30:40

File A: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 ludwig\ludwig nero.m4a
File B: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 ludwig\ludwig.wv

23:30:40 : Test started.
23:31:49 : 00/01 100.0%
23:32:42 : 00/02 100.0%
23:33:51 : 01/03 87.5%
23:34:35 : 01/04 93.8%
23:35:02 : 01/05 96.9%
23:35:19 : Trial reset.
23:35:59 : 00/01 100.0%
23:36:13 : 00/02 100.0%
23:37:58 : 01/03 87.5%
23:38:12 : 02/04 68.8%
23:38:42 : 03/05 50.0%
23:39:02 : 04/06 34.4%
23:39:47 : 05/07 22.7%
23:40:15 : 05/08 36.3%
23:40:39 : 06/09 25.4%
23:41:09 : 07/10 17.2%
23:41:45 : 08/11 11.3%
23:42:18 : 08/12 19.4%
23:42:56 : 09/13 13.3%
23:43:29 : 10/14 9.0%
23:44:04 : 10/15 15.1%
23:44:36 : 11/16 10.5%
23:45:05 : 11/17 16.6%
23:45:56 : 12/18 11.9%
23:46:42 : 12/19 18.0%
23:46:47 : Test finished.

----------
Total: 13/24 (41.9%)

foo_abx 1.3.4 report
foobar2000 v1.1.2
2011/02/09 23:47:31

File A: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 ludwig\ludwig qt.m4a
File B: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 ludwig\ludwig.wv

23:47:31 : Test started.
23:48:12 : 01/01 50.0%
23:48:38 : 02/02 25.0%
23:49:03 : 02/03 50.0%
23:49:25 : 02/04 68.8%
23:49:51 : 03/05 50.0%
23:50:17 : 04/06 34.4%
23:51:00 : 04/07 50.0%
23:51:30 : 04/08 63.7%
23:52:06 : 05/09 50.0%
23:52:28 : 06/10 37.7%
23:53:11 : 06/11 50.0%
23:53:46 : 07/12 38.7%
23:54:21 : 08/13 29.1%
23:54:43 : 09/14 21.2%
23:55:20 : 10/15 15.1%
23:55:49 : 11/16 10.5%
23:56:26 : 12/17 7.2%
23:56:53 : 13/18 4.8%
23:57:15 : 14/19 3.2%
23:57:32 : 15/20 2.1%
23:57:56 : 16/21 1.3%
23:58:33 : 17/22 0.8%
23:58:37 : Test finished.

----------
Total: 17/22 (0.8%)

I will try it with ogg later or maybe tomorrow. Regards.

[punkrockdude]

I did a second ABX of the classical (ludwig) piece with Nero and this time is wa very easy to ABX. Some elements lost ambience/reverb compared to the lossless:

foo_abx 1.3.4 report
foobar2000 v1.1.2
2011/02/10 00:05:40

File A: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 ludwig\ludwig nero.m4a
File B: E:\Documents and Settings\Administratör\Mina dokument\128kbps test\128 ludwig\ludwig.wv

00:05:40 : Test started.
00:07:00 : 00/01 100.0%
00:08:19 : 00/02 100.0%
00:09:09 : 00/03 100.0%
00:09:59 : 01/04 93.8%
00:10:21 : 02/05 81.3%
00:10:53 : 02/06 89.1%
00:11:07 : 03/07 77.3%
00:11:28 : 03/08 85.5%
00:11:54 : 04/09 74.6%
00:12:07 : 05/10 62.3%
00:12:26 : 05/11 72.6%
00:12:41 : 06/12 61.3%
00:12:57 : 07/13 50.0%
00:13:16 : 08/14 39.5%
00:13:30 : 09/15 30.4%
00:13:41 : 10/16 22.7%
00:14:31 : 11/17 16.6%
00:14:42 : 12/18 11.9%
00:15:00 : 13/19 8.4%
00:15:11 : 14/20 5.8%
00:15:19 : 15/21 3.9%
00:15:44 : 16/22 2.6%
00:15:53 : 17/23 1.7%
00:15:59 : 18/24 1.1%
00:16:07 : 19/25 0.7%
00:16:23 : 20/26 0.5%
00:16:39 : 21/27 0.3%
00:16:51 : 22/28 0.2%
00:17:14 : 23/29 0.1%
00:17:26 : 24/30 0.1%
00:17:34 : 25/31 0.0%
00:17:54 : 26/32 0.0%
00:17:59 : Test finished.

----------
Total: 26/32 (0.0%)

If I am ging to to be honest, I would NOT had heard the difference if I didn't listen really carefully. A normal listening situation probably wouldn't relieve Nero from the lossless. Regards.

Edit: Corrected the worst english.

This post has been edited by punkrockdude: Feb 10 2011, 00:26

[C.R.Helmrich]

Am I blind, or where did you and list get the prodigy.wv and ludwig.wv from? Please link to or upload the original .wv files in this forum so we can try it ourselves!

Edit: Regarding prodigy.wv, I am blind. Sorry. But do you have ludwig.wv for me?

Thanks,

Chris

This post has been edited by C.R.Helmrich: Feb 10 2011, 00:39

[punkrockdude]

Am I blind, or where did you and list get the prodigy.wv and ludwig.wv from? Please link to or upload the original .wv files in this forum so we can try it ourselves!

Thanks,

Chris

I think it is on the first page of the thread where it is linked to on a mediafire.com site. Regards.

This post has been edited by punkrockdude: Feb 10 2011, 00:42

[Martel]

NeroAAC is still suffering from the bug I reported a while ago (I just downloaded the latest encoder from their site and ran the test at q=0.5 and 0.55):
http://www.hydrogenaudio.org/forums/index....showtopic=83246

Unless they messed up the Quick Time encoder, it should encode the sample just fine.

Cymbal rides are used a lot in rock, metal, jazz, blues and whatnot and they get smeared by NeroAAC. You don't need to ridicule the guy for thinking it's inferior.

[greynol]

1) No one is being ridiculed.

2) It is readily apparent that a member who agreed to our TOS made claims about quality on this forum without knowing about double-blind testing methodology. It doesn't matter if it the claims were correct or not, this is unacceptable.

[punkrockdude]

C.R.: Both files were included in the file that I downloaded. www.mediafire.com/download.php?ffj8cm332bq7a7e . Regards.

This post has been edited by punkrockdude: Feb 10 2011, 21:57

[C.R.Helmrich]

Ah, thanks. By the way, since list supposedly ABXed prodigy at 192 kbps, you or him should upload the original file on HA directly (so it's preserved here), for example in this thread:

http://www.hydrogenaudio.org/forums/index....showtopic=70598

and preferably in FLAC format since all other lossless files in that thread are in that format. When you do, please link to this thread.

Thanks,

Chris

[3biga]

My choice is QT (--tvbr 120). Nero has good detailed focused highs, but lows and mids are blurry and depressed. It isn't transparent for me (ABX 8/8 (0.4%)).
QT sounds little bit pushy compared to original, but still transparent.