IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
iTunes Plus AAC encoding flawy, High and Best settings cause corrupt audio
Danny_DJ
post Sep 7 2011, 16:28
Post #1





Group: Members
Posts: 10
Joined: 5-September 11
Member No.: 93498



Hi,

by chance i discovered audio glitches in AAC encoding. I used 256 CBR Stereo AAC w/ XLD 20110821 on Mac OS 10.6.8 with high or best mode, but it gave me a eventually "scratchy sound". I contacted the developer of XLD, "tmkk". He cross-checked it w/ iTunes Plus internal encoding and confirmed this bug and suggested me to contact this forum since here are some of the iTunes' AAC encoder's developers. So could anyone please have a look into this?

The affected source file can be found here:
http://soundcloud.com/dannywho/amelie-went...eview-1/s-FrDeC

Already at the intro there is an alteration of the sound recognizable. There it sounds quite musical (it seens a little bit like conga drums on top of the bassdrum) but since I composed it on my own I know which sounds I've put in and which not.
And then very noticably from 2:35 through the end, five times very significantly.

Try to convert it on your own.
Thanks for your efforts and response on this in advance smile.gif

Regards,
Danny

This post has been edited by db1989: Sep 7 2011, 18:23
Reason for edit: restoring link; self-composed so not a TOS #9 violation


--------------------
My Music: www.soundcloud.com/dannywho
Go to the top of the page
+Quote Post
db1989
post Sep 7 2011, 18:07
Post #2





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



QUOTE (Danny_DJ @ Sep 7 2011, 16:28) *
I contacted the developer of XLD, "tmkk". He cross-checked it w/ iTunes Plus internal encoding and confirmed this bug and suggested me to contact this forum since here are some of the iTunes' AAC encoder's developers.
Are there? I know some people who work on Neroís AAC codec are members here, but Iíve never heard of there being anyone from Apple (or whoever provides their encoder). I could well be wrong, but Iíd be interested in confirmation either way.
Go to the top of the page
+Quote Post
mixminus1
post Sep 7 2011, 18:37
Post #3





Group: Members
Posts: 689
Joined: 23-February 05
Member No.: 20097



Confirmed under Windows XP via qaac and qtaacenc with foobar2000, QT 7.7.0.

Weird "ghost notes" at the beginning, and very distinct distortion on the bass notes after the 2:35 mark.

I didn't try qaac's "quality" switch, as I don't know what its 0-2 range maps to, but trying the various quality settings in qtaacenc, the artifacts were consistent with --best, --higher, and --high (the default), but disappeared with --normal.

This was true at CBR 256 and 192, as well as --tvbr 90.

Good find!


--------------------
"Not sure what the question is, but the answer is probably no."
Go to the top of the page
+Quote Post
Danny_DJ
post Sep 7 2011, 18:49
Post #4





Group: Members
Posts: 10
Joined: 5-September 11
Member No.: 93498



@db1989: don't remove this entry, it is required for other users to be able to make comparisons and confirmations on their own.

The unflawed original AIFF source can be found here:
http://soundcloud.com/dannywho/amelie-went...preview/s-Sdqx4


--------------------
My Music: www.soundcloud.com/dannywho
Go to the top of the page
+Quote Post
b66pak
post Sep 7 2011, 19:14
Post #5





Group: Members
Posts: 58
Joined: 2-February 10
Member No.: 77800



QUOTE (mixminus1 @ Sep 7 2011, 20:37) *
I didn't try qaac's "quality" switch, as I don't know what its 0-2 range maps to...


tested for "--tvbr 127" and compared the output aac lenght:

qaac quality 0 = qtaacenc fastest

qaac quality 0 = qtaacenc fast

qaac quality 1 = qtaacenc normal

qaac quality 2 = qtaacenc high

qaac quality 2 = qtaacenc highest
_
Go to the top of the page
+Quote Post
/mnt
post Sep 7 2011, 19:46
Post #6





Group: Members
Posts: 697
Joined: 22-April 06
Member No.: 29877



Sounds like it's a problem with the pre-echo handling on QT. This problem is much worse on certain Kraftwerk tracks (Die Roboter / The Robots and Spacelab), even at tvbr 127.

This post has been edited by /mnt: Sep 7 2011, 19:47


--------------------
"I never thought I'd see this much candy in one mission!"
Go to the top of the page
+Quote Post
Danny_DJ
post Sep 9 2011, 09:10
Post #7





Group: Members
Posts: 10
Joined: 5-September 11
Member No.: 93498



I got informed that the reason for that problem has been found and the fix should come in a future release, but w/o any mention about when exactly.


--------------------
My Music: www.soundcloud.com/dannywho
Go to the top of the page
+Quote Post
kawaiigardiner
post Sep 30 2011, 04:48
Post #8





Group: Members
Posts: 32
Joined: 3-August 09
Member No.: 71985



I know I am a little late to the conversation but has the original poster tried the exact same audio file with Mac OS X Lion? I've encoded around 1/2 dozen different files so far and haven't come across it so I wonder whether it is a 'Snow Leopard only bug'.
Go to the top of the page
+Quote Post
mixminus1
post Sep 30 2011, 05:17
Post #9





Group: Members
Posts: 689
Joined: 23-February 05
Member No.: 20097



Is the version of QuickTime included with Lion newer than 7.7.0? It's definitely not exclusive to Snow Leopard as I was able to reproduce it under Windows with the same version of QuickTime.

Note that this particular artifact is very program-dependent, and since it affects all AAC encoding methods of QuickTime from the default quality setting on up, it theoretically could show up on tracks currently on the iTunes Music Store.

However, of the several dozen I've downloaded over the past couple years, I've never heard anything like it, nor anything else I could positively identify as an encoding artifact.


--------------------
"Not sure what the question is, but the answer is probably no."
Go to the top of the page
+Quote Post
kaoruAngel
post Oct 18 2011, 22:26
Post #10





Group: Members
Posts: 23
Joined: 27-May 05
Member No.: 22323



SoundCloud reports that it "can't find that page!" for both links posted [ http://soundcloud.com/dannywho/amelie-went...eview-1/s-FrDeC and http://soundcloud.com/dannywho/amelie-went...preview/s-Sdqx4 ]. Where can I find this source track to test with?

I made the decision this past week to switch to QT's AAC encoder after reading IgorC's latest AAC listening test in which Nero performed poorly, but now your report concerns me! Am I wrong to use QT?
Go to the top of the page
+Quote Post
Danny_DJ
post Oct 18 2011, 22:32
Post #11





Group: Members
Posts: 10
Joined: 5-September 11
Member No.: 93498



I took these files offline since "skuo" as part of the Apple AAC dev team reported that he found the reason for that issue and that a solution will be implemented in furter QT engine updates.

I add that this issue was the only one I ever encountered w/ AAC. I am buying a lot of iTunes music which is encoded the same way and never had any sound issues.


--------------------
My Music: www.soundcloud.com/dannywho
Go to the top of the page
+Quote Post
matt the cat
post Jan 23 2012, 18:10
Post #12





Group: Members
Posts: 30
Joined: 28-December 10
Member No.: 86874



To be honest, I dont really trust Apple, it would be useful to be able to test this with the newer versions.
Go to the top of the page
+Quote Post
HTS
post Jan 29 2012, 04:02
Post #13





Group: Members
Posts: 356
Joined: 13-October 07
Member No.: 47799



QUOTE (mixminus1 @ Sep 7 2011, 12:37) *
Confirmed under Windows XP via qaac and qtaacenc with foobar2000, QT 7.7.0.

Weird "ghost notes" at the beginning, and very distinct distortion on the bass notes after the 2:35 mark.

I didn't try qaac's "quality" switch, as I don't know what its 0-2 range maps to, but trying the various quality settings in qtaacenc, the artifacts were consistent with --best, --higher, and --high (the default), but disappeared with --normal.

This was true at CBR 256 and 192, as well as --tvbr 90.

Good find!

Qaac's site says:

QUOTE
Since 1.00, qaac directly uses CoreAudioToolbox.dll. Therefore, QuickTime installation is no more required. However, Apple Application Support is required.


Does this make a difference?
Go to the top of the page
+Quote Post
Danny_DJ
post Feb 16 2012, 18:26
Post #14





Group: Members
Posts: 10
Joined: 5-September 11
Member No.: 93498



Hey guys,

DeaDruiD PM'd me and told me the following:

QUOTE
The glitches were easily hearable when encoding using qaac+CoreAudioToolbox.dll 7.9.7.3 (included in QuickTime 7.7.0 for Windows) and disappeared completely when using CoreAudioToolbox.dll 7.9.7.9 (included in Safari 5.1.4 for Windows). The newest version does not include any changes to the AAC encoder, so it is safe to assume that version 7.9.7.8 (included in QuickTime 7.7.1, iTunes 10.5.3 and Safari 5.1.2) is glitch-free as well. Using newest version of QuickTime/iTunes for Mac should work as well.


But now comes my problem: I am on Snow Leopard 10.6.8, my System Profiler tells me that my QT framework is on 7.6.6.
There seems to be no way to seperately get any higher version for SL. Or am I wrong?
My System Updater gives me no QT option and the Apple download section tells me that QT is always up to date w/ my system updater.

Are QT updates discontinued w/ the introduction of Lion?

Regards,

Danny


--------------------
My Music: www.soundcloud.com/dannywho
Go to the top of the page
+Quote Post
Danny_DJ
post Feb 16 2012, 18:56
Post #15





Group: Members
Posts: 10
Joined: 5-September 11
Member No.: 93498



Update:

I just went for the latest security update which was available and that now fixed the glitches, although my system profiler still tells me that I am on 7.6.6.

Thanks for the developers around that they investigated in that issue, found the error and finally have implemented the solution with the latest update smile.gif


--------------------
My Music: www.soundcloud.com/dannywho
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 23rd December 2014 - 05:00