Nero AAC Codec 1.3.3.0 Options

[/mnt]

q 0.41 is ABXable. There is some background warm noise and as effect snare drum sounds sandy.

I notice there was drum snare on this sample to at -q 0.41.

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.5
2009/05/06 11:52:57

File A: C:\Downloads\Descending_Darkness.flac
File B: C:\Temp\Descending Darkness.mp4

11:52:57 : Test started.
11:53:22 : 01/01 50.0%
11:53:51 : 02/02 25.0%
11:54:20 : 03/03 12.5%
11:54:39 : 03/04 31.3%
11:54:48 : 04/05 18.8%
11:54:54 : 05/06 10.9%
11:55:01 : 06/07 6.3%
11:55:09 : 07/08 3.5%
11:55:17 : 08/09 2.0%
11:55:34 : 09/10 1.1%
11:56:03 : 10/11 0.6%
11:56:16 : 11/12 0.3%
11:56:28 : 12/13 0.2%
11:56:56 : 13/14 0.1%
11:56:58 : Test finished.

----------
Total: 13/14 (0.1%)

[Sylph]

New Nero is out, any news for the codec?

[menno]

New Nero is out, any news for the codec?

No, still 1.3.3.x

[/mnt]

I've sometimes come across a track that is not transparent at 0.55, but sounds fine at 0.50.

For example Rage Against The Machine's Bulls On Parade:

Nero AAC 1.3.3.0 -q 0.55

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/25 22:29:15

File A: C:\Temp\Samples\Bulls On Parade (Edit).flac
File B: C:\Temp\Samples\Bulls On Parade (Edit) q 0.55.m4a

22:29:15 : Test started.
22:29:27 : 01/01 50.0%
22:29:32 : 02/02 25.0%
22:29:43 : 03/03 12.5%
22:29:49 : 03/04 31.3%
22:29:55 : 04/05 18.8%
22:30:05 : 05/06 10.9%
22:30:19 : 06/07 6.3%
22:30:27 : 07/08 3.5%
22:30:34 : 08/09 2.0%
22:30:39 : 09/10 1.1%
22:30:43 : 10/11 0.6%
22:30:46 : 11/12 0.3%
22:30:51 : 12/13 0.2%
22:30:55 : 13/14 0.1%
22:31:00 : 14/15 0.0%
22:31:05 : 15/16 0.0%
22:31:11 : 16/17 0.0%
22:31:13 : Test finished.

----------
Total: 16/17 (0.0%)

Stratch noise like artifact at 0:08.

Nero AAC 1.3.3.0 -q 0.50

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/25 22:31:40

File A: C:\Temp\Samples\Bulls On Parade (Edit).flac
File B: C:\Temp\Samples\Bulls On Parade (Edit) q 0.50.m4a

22:31:40 : Test started.
22:32:02 : 01/01 50.0%
22:32:09 : 02/02 25.0%
22:32:17 : 02/03 50.0%
22:32:26 : 02/04 68.8%
22:32:34 : 03/05 50.0%
22:32:40 : 03/06 65.6%
22:32:44 : 03/07 77.3%
22:32:53 : 03/08 85.5%
22:33:11 : 04/09 74.6%
22:33:20 : 05/10 62.3%
22:33:30 : 05/11 72.6%
22:33:35 : 05/12 80.6%
22:33:36 : Test finished.

----------
Total: 5/12 (80.6%)

Seems to sound ok to me.

q 0.50 vs q 0.55

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/25 22:34:01

File A: C:\Temp\Samples\Bulls On Parade (Edit) q 0.50.m4a
File B: C:\Temp\Samples\Bulls On Parade (Edit) q 0.55.m4a

22:34:01 : Test started.
22:34:17 : 01/01 50.0%
22:34:20 : 02/02 25.0%
22:34:24 : 03/03 12.5%
22:34:28 : 04/04 6.3%
22:34:33 : 05/05 3.1%
22:34:37 : 06/06 1.6%
22:34:43 : 07/07 0.8%
22:34:47 : 08/08 0.4%
22:34:51 : 09/09 0.2%
22:34:56 : 10/10 0.1%
22:35:00 : 11/11 0.0%
22:35:06 : 12/12 0.0%
22:35:11 : 13/13 0.0%
22:35:20 : 14/14 0.0%
22:35:24 : 15/15 0.0%
22:35:25 : Test finished.

----------
Total: 15/15 (0.0%)

The q 0.55 encode sounds worse then the 0.50 encode at 0:08.

Also anyone testing out Nero AAC might be interested in this major precho killer sample:

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/25 22:46:46

File A: C:\Temp\Samples\Show Me Your Spine (Edit).flac
File B: C:\Temp\Samples\Show Me Your Spine (Edit) q 0.50.m4a

22:46:46 : Test started.
22:46:51 : 01/01 50.0%
22:46:58 : 02/02 25.0%
22:47:06 : 03/03 12.5%
22:47:09 : 04/04 6.3%
22:47:17 : 05/05 3.1%
22:47:26 : 06/06 1.6%
22:47:40 : 07/07 0.8%
22:47:56 : 08/08 0.4%
22:48:09 : 09/09 0.2%
22:48:16 : 09/10 1.1%
22:48:23 : 10/11 0.6%
22:48:32 : 11/12 0.3%
22:48:37 : 12/13 0.2%
22:48:55 : 13/14 0.1%
22:49:03 : 14/15 0.0%
22:49:05 : Test finished.

----------
Total: 14/15 (0.0%)

Major precho artifacts from the start, sorta like eig. Sounds miles better then any 320 CBR Mp3 encodes though.

Samples:
Bulls On Parade
Show Me Your Spine

This post has been edited by /mnt: Jul 25 2009, 23:08

[2E7AH]

I converted WAV file (6.041.498 samples; 2:16.995 length) to AAC using neroaacenc 1.3.3.0
foobar reports correctly encoded file: 2:16.995 (6.041.498 samples)
Audition reports: 2:17.020 (6.042.624 samples)

Then editing AAC file in Audition shows approx. 40 milliseconds shift

Is it something wrong with Audition AAC filter or?

[edit] parameters: -q 0.50 -ignorelength

This post has been edited by 2E7AH: Jul 26 2009, 08:14

[Alexxander]

I've ABXed "Show Me Your Spine" flac against Nero AAC 1.3.3.0 q0.50 (166kbps) and when listening to the sample with headphone I noticed inmediately some problem with stereo balance at around 4.5 sec:

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/26 15:35:40

File A: F:\Temp\Show_Me_Your_Spine__Sample_.flac
File B: F:\Temp\Show_Me_Your_Spine__Sample_.mp4

15:35:40 : Test started.
15:37:17 : 01/01 50.0%
15:37:34 : 02/02 25.0%
15:37:45 : 03/03 12.5%
15:37:53 : 04/04 6.3%
15:37:58 : 05/05 3.1%
15:38:03 : 06/06 1.6%
15:38:13 : 07/07 0.8%
15:38:20 : 08/08 0.4%
15:38:24 : 09/09 0.2%
15:38:29 : 10/10 0.1%
15:38:41 : 11/11 0.0%
15:38:47 : 12/12 0.0%
15:38:55 : 13/13 0.0%
15:39:04 : 14/14 0.0%
15:39:13 : 15/15 0.0%
15:39:18 : 16/16 0.0%
15:39:26 : Test finished.

----------
Total: 16/16 (0.0%)

Edit: Added results for q0.55, q0.60 and q0.65.

bitrate q0.55=183kbps, ABXed 13/13
bitrate q0.60=208kbps, ABXed 10/10
bitrate q0.65=231kbps, ABXed 19/20

Increasing q value makes the problem harder to detect/less pronounced, at q0.65 good concentration was necessary. It doesn't sound as distortion and it isn't anoying at all, it just sounds different from the flac file.

For sake of breveness I only include complete ABX log for q0.65:

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/26 15:57:24

File A: F:\Temp\Show_Me_Your_Spine__Sample_.flac
File B: F:\Temp\Show_Me_Your_Spine__Sample_Nero1330q065.mp4

15:57:24 : Test started.
15:58:09 : 01/01 50.0%
15:58:17 : 02/02 25.0%
15:58:26 : 02/03 50.0%
15:58:43 : 03/04 31.3%
15:58:47 : 04/05 18.8%
15:58:52 : 05/06 10.9%
15:58:58 : 06/07 6.3%
15:59:11 : 07/08 3.5%
15:59:17 : 08/09 2.0%
15:59:22 : 09/10 1.1%
15:59:36 : 10/11 0.6%
15:59:46 : 11/12 0.3%
15:59:56 : 12/13 0.2%
16:00:07 : 13/14 0.1%
16:00:35 : 14/15 0.0%
16:00:42 : 15/16 0.0%
16:00:49 : 16/17 0.0%
16:00:56 : 17/18 0.0%
16:01:03 : 18/19 0.0%
16:01:12 : 19/20 0.0%
16:01:14 : Test finished.

----------
Total: 19/20 (0.0%)

This post has been edited by Alexxander: Jul 26 2009, 15:12

[/mnt]

I've tried the Show Me Your Spine sample with the iTunes encoder. I thought it might be interesting since i remember seeing this track on the iTunes store at 128 DRM AAC a few years ago.

iTunes 8.2.16 AAC 160kbps VBR

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/26 16:06:26

File A: C:\Temp\Samples\Show Me Your Spine (Edit).flac
File B: C:\Temp\Samples\Show Me Your Spine (Edit) [iTunes 160kbps VBR].m4a

16:06:26 : Test started.
16:06:32 : 01/01 50.0%
16:06:37 : 02/02 25.0%
16:06:41 : 03/03 12.5%
16:06:44 : 04/04 6.3%
16:06:48 : 05/05 3.1%
16:06:53 : 06/06 1.6%
16:07:07 : 07/07 0.8%
16:07:13 : 08/08 0.4%
16:07:18 : 09/09 0.2%
16:07:23 : 10/10 0.1%
16:07:31 : 11/11 0.0%
16:07:40 : 11/12 0.3%
16:07:46 : Test finished.

----------
Total: 11/12 (0.3%)

I've ABX the first few seconds, same bitrate as Nero AAC at 0.50. Sounds worse then Nero AAC at 0.50 IMO.

iTunes 8.2.16 AAC iTunes Plus (256kbps)

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/07/26 16:08:17

File A: C:\Temp\Samples\Show Me Your Spine (Edit).flac
File B: C:\Temp\Samples\Show Me Your Spine (Edit) [iTunes Plus 256kbps].m4a

16:08:17 : Test started.
16:08:28 : 01/01 50.0%
16:08:34 : 01/02 75.0%
16:08:39 : 02/03 50.0%
16:08:44 : 03/04 31.3%
16:08:51 : 04/05 18.8%
16:08:55 : 05/06 10.9%
16:09:04 : 06/07 6.3%
16:09:25 : 07/08 3.5%
16:09:39 : 08/09 2.0%
16:09:55 : 09/10 1.1%
16:10:19 : 10/11 0.6%
16:10:28 : 11/12 0.3%
16:10:35 : 12/13 0.2%
16:10:42 : 13/14 0.1%
16:10:52 : 14/15 0.0%
16:11:09 : 15/16 0.0%
16:11:19 : Test finished.

----------
Total: 15/16 (0.0%)

I've ABXed the first few secs again, sounds better but different then the 160 VBR version. Very easy to ABX unlike Nero at higher settings.

[2E7AH]

Is it something wrong with Audition AAC filter or?

I found somewhat old Nero MPEG-4 filter, and thought to try it. Then to get it to work some additional files are needed: aac.dll; aacenc32.dll; NeroIPP.dll and placed in "C:\Program Files\Common Files\Ahead\AudioPlugins" folder (Nero 6). But unfortunately, the result was same as with FAAC/FAAD filter:

Nero MPEG-4 filter: +36ms
FAAC/FAAD filter: +36ms
QuickTime filter: +11ms

I guess Nero Digital AAC filter is missing here

[/mnt]

I've found a yet another problem sample that effects Nero AAC at 0.50 greatly:

-q 0.50

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/08/18 13:35:50

File A: C:\Music\Nine Inch Nails\Closer To God (Single)\01. Closer To God.m4a
File B: E:\Music\Albums\Nine Inch Nails - Closer To God (Single)\01. Closer To God.flac

13:35:50 : Test started.
13:36:06 : 01/01 50.0%
13:36:18 : 02/02 25.0%
13:36:26 : 03/03 12.5%
13:36:33 : 04/04 6.3%
13:36:43 : 05/05 3.1%
13:36:52 : 06/06 1.6%
13:37:02 : 07/07 0.8%
13:37:12 : 08/08 0.4%
13:37:24 : 09/09 0.2%
13:37:36 : 10/10 0.1%
13:37:37 : Test finished.

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

Nasty disortion and possibly stereo corruption. Very easy to ABX.

Edit sample:

-q 0.5

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/08/18 13:55:48

File A: C:\Temp\Closer To God [Edit].flac
File B: C:\Temp\Closer To God [Edit] q 0.5.mp4

13:55:48 : Test started.
13:55:55 : 01/01 50.0%
13:56:01 : 02/02 25.0%
13:56:07 : 03/03 12.5%
13:56:10 : 04/04 6.3%
13:56:15 : 05/05 3.1%
13:56:21 : 06/06 1.6%
13:56:26 : 07/07 0.8%
13:56:31 : 08/08 0.4%
13:56:35 : 09/09 0.2%
13:56:41 : 10/10 0.1%
13:56:43 : Test finished.

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

Still present on the sample.

-q 0.55

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/08/18 13:57:52

File A: C:\Temp\Closer To God [Edit].flac
File B: C:\Temp\Closer To God [Edit] q 0.55.mp4

13:57:52 : Test started.
13:58:11 : 01/01 50.0%
13:58:18 : 02/02 25.0%
13:58:28 : 03/03 12.5%
13:58:40 : 04/04 6.3%
13:58:56 : 05/05 3.1%
13:59:11 : 06/06 1.6%
13:59:22 : 07/07 0.8%
13:59:38 : 08/08 0.4%
13:59:48 : 09/09 0.2%
13:59:58 : 10/10 0.1%
14:00:00 : Test finished.

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

Better then 0.50, but still far from being transparent.

iTunes AAC 160kbps VBR

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/08/18 14:02:57

File A: C:\Temp\Closer To God [Edit] iTunes 160 VBR.m4a
File B: C:\Temp\Closer To God [Edit].flac

14:02:57 : Test started.
14:03:42 : 01/01 50.0%
14:03:57 : 02/02 25.0%
14:04:24 : 03/03 12.5%
14:04:40 : 04/04 6.3%
14:04:54 : 05/05 3.1%
14:05:09 : 05/06 10.9%
14:05:19 : 05/07 22.7%
14:05:35 : 05/08 36.3%
14:05:36 : Test finished.

----------
Total: 5/8 (36.3%)

Possible but i find it hard to ABX on the iTunes encode. Sounds better then the Nero encodes.

Sample:

Closer To God

This post has been edited by /mnt: Aug 18 2009, 14:27

[muaddib]

Thank you for reporting it.
What were the bitrates?

[/mnt]

It's also ABXable at -q 0.60:

CODE

foo_abx 1.3.4 report
foobar2000 v0.9.6.8
2009/08/19 12:57:00

File A: C:\Temp\Closer To God [Edit] -q 0.6.mp4
File B: C:\Temp\Closer To God [Edit].flac

12:57:00 : Test started.
12:57:48 : 01/01 50.0%
12:58:00 : 02/02 25.0%
12:58:21 : 03/03 12.5%
12:58:35 : 04/04 6.3%
12:58:43 : 05/05 3.1%
12:58:58 : 06/06 1.6%
12:59:18 : 07/07 0.8%
12:59:31 : 08/08 0.4%
12:59:39 : 09/09 0.2%
12:59:49 : 10/10 0.1%
12:59:58 : 11/11 0.0%
13:00:06 : 12/12 0.0%
13:00:07 : Test finished.

----------
Total: 12/12 (0.0%)

A artifact appears at 0:02. Then a really nasty artifact appears at 0:06.

Thank you for reporting it.
What were the bitrates?

Nero -q 50: 145 kbps
Nero -q 55: 172 kbps
Nero -q 60: 192 kbps
iTunes 160 VBR: 188 kbps

Judging from the bitrate display under foobar2000 (1 sec update display), Nero AAC seems to use low bitrates on this sample.

[stereotype]

New Nero is out, any news for the codec?

No, still 1.3.3.x

@menno

Any chance of releasing a 64-bit binary for linux?
Shouldn't be too hard or too much work I think, no?

If not, any tips on what is needed to run the 32-bit binary on a 64-bit machine (Fedora 12 in my case)

Thanks

[stereotype]

In case anyone else needs it, a simple # yum install glibc.i686 libstdc++.i686 solved my problem...

64-bit binaries would still be nice though