Skip to main content

Topic: New iTunes, new CoreAudioToolbox.dll (Read 83601 times) previous topic - next topic

0 Members and 2 Guests are viewing this topic.
  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
iTunes 10.7.0.21, CoreAudioToolbox.dll 7.9.8.1. Waiting for some expert to tell what's new.

Pink Floyd - The Dark Side of The Moon -V63 with 7.9.7.9 = 43.2MB, avg. bitrate 139 Kbps / -V63 with 7.9.8.1 = 42.6MB, avg. bitrate 137 Kbps

Pink Floyd - Wish You Were Here -V63 with 7.9.7.9 = 43.2MB, avg. bitrate 135 Kbps / -V63 with 7.9.8.1 = 41.7MB, avg. bitrate 130 Kbps
  • Last Edit: 12 September, 2012, 04:16:45 PM by eahm

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #1
1406 -V63 songs + covers = 7.9.7.9 5.48GB / 7.9.8.1 5.36GB

I wanted to try to compare ALAC but the command -A doesn't work anymore:

Manual
Code: [Select]
qaac.exe -A "03. Pigs (Three Different Ones).wav"
qaac 1.40, CoreAudioToolbox 7.9.8.1

03. Pigs (Three Different Ones).wav
Apple Lossless Encoder
ERROR: Invalid ALACSpecificConfig!


foobar2000
Code: [Select]
Conversion failed: The encoder has terminated prematurely with code 2 (0x00000002); please re-check parameters
  • Last Edit: 12 September, 2012, 05:38:46 PM by eahm

New iTunes, new CoreAudioToolbox.dll
Reply #2
According to Apple, today's update was only incremental as it "just" added support for iOS 6 and iDevices that support said OS.  The big update, which isn't coming until October, may or may not upgrade the encoder that they use.  So I too am curious to see why things are a little different with this version of iTunes.

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #3
Right after I've reconverted everything I was thinking about iTunes 11, and maybe do it all over again, well... 20mins for ~1400 songs is not that bad.

I don't even think we needed a 10.7 version, I've been using iOS 6 since the first day of the beta and it was perfectly functional with iTunes 10.6, of course they have their reasons.

Let's wait for a reply from nu774 and see what's new on this version.
  • Last Edit: 12 September, 2012, 07:18:44 PM by eahm

  • nu774
  • [*][*][*][*][*]
  • Developer
New iTunes, new CoreAudioToolbox.dll
Reply #4
Released qaac 1.41 now.

There was a format change of ALAC magic cookie on the new CoreAudioToolbox (7.9.8.1).
New version returns the same format as the open sourced ALAC encoder, which is documented in ALACMagicCookieDescription.txt.
Probably this update was done to make them consistent.

qaac 1.41 takes care of both old and new format.

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #5
Thank you for the update.

Did anything change in terms of quality for AAC? I'm asking because the avg. bitrate for True VBR is lower and even the size is.
  • Last Edit: 13 September, 2012, 01:34:23 AM by eahm

  • nu774
  • [*][*][*][*][*]
  • Developer
New iTunes, new CoreAudioToolbox.dll
Reply #6
Did anything change in terms of quality for AAC? I'm asking because the avg. bitrate for True VBR is lower and even the size is.

I confirmed resulting bitstream (and size) is different from 7.9.7.9, but I don't know what has actually changed (quality, bug-fix, etc).

  • Anakunda
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #7
I reencoded an album earlier encoded with CoreAudio 7.9.7.9 and the bitrate of newer CoreAudio is slightly lower.
Also I have made a spectral views of encoded audio short part, they are slightly different. Any conclusion which snapshot could be better?

7.9.7.9:


7.9.8.1:

  • mixminus1
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #8
The lowpass frequency appears to be the same - beyond that, there is nothing useful that can be gained from those.
"Not sure what the question is, but the answer is probably no."

  • dhromed
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #9
Any conclusion which snapshot could be better?


Depends. Which is the one that makes sound?

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #10
Bit-Comparing two files made with 7.9.7.9 and 7.9.8.1:


-a128
Differences found in 1 out of 1 track pairs.

Comparing:
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -a128 7979\04. Time.m4a"
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -a128 7981\04. Time.m4a"
Differences found: 36437644 sample(s), starting at 0.0000000 second(s), peak: 0.3619905 at 13.0465986 second(s), 2ch


-a128 -q1 (same as iTunes High Quality (128kbps) setting)
Differences found in 1 out of 1 track pairs.

Comparing:
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -a128 -q1 7979\04. Time.m4a"
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -a128 -q1 7981\04. Time.m4a"
Differences found: 36372859 sample(s), starting at 0.0000000 second(s), peak: 0.3286602 at 15.3048526 second(s), 1ch


-c128
Differences found in 1 out of 1 track pairs.

Comparing:
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -c128 7979\04. Time.m4a"
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -c128 7981\04. Time.m4a"
Differences found: 36442639 sample(s), starting at 0.0000000 second(s), peak: 0.3739062 at 302.3941043 second(s), 1ch


-V63
Differences found in 1 out of 1 track pairs.

Comparing:
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -V63 7979\04. Time.m4a"
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -V63 7981\04. Time.m4a"
Differences found: 36430926 sample(s), starting at 0.0000000 second(s), peak: 0.2852085 at 337.7100000 second(s), 1ch



-v128
Differences found in 1 out of 1 track pairs.

Comparing:
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -v128 7979\04. Time.m4a"
"D:\Music\ToDo\Pink Floyd\1973 - The Dark Side of the Moon (2011 Remaster) -v128 7981\04. Time.m4a"
Differences found: 36442702 sample(s), starting at 0.0000000 second(s), peak: 0.3185556 at 12.5673243 second(s), 1ch



I don't know if anyone can tell what's different from this.
  • Last Edit: 13 September, 2012, 07:26:00 PM by eahm

New iTunes, new CoreAudioToolbox.dll
Reply #11
I don't think much can be said other than the results that were found conducting that test.  It might be all for not though since iTunes 11 is about a month away from being released.  Even then, that might not change anything with the encoder that Apple is implementing.  I have my reservations about upgrading to iTunes since it appears they are getting rid of the list view of an entire library but that is a different discussion.

  • neothe0ne
  • [*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #12
TVBR 91 appears to have no difference between qaac 1.39 and 1.41.

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #13
neothe0ne, you may want to research how qaac works.
  • Last Edit: 14 September, 2012, 02:33:00 AM by eahm

  • Anakunda
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #14
If I understand good then audio data is related to apple's encoding library not to qaac.
Although I think there's little difference from previous version (compare bitrates and spectral graphs between versions)

  • neothe0ne
  • [*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #15
neothe0ne, you may want to research how qaac works.

Core Audio Toolbox was changed between those 2 versions, so what are you implying that I'm doing wrong here?

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #16
neothe0ne, you may want to research how qaac works.

Core Audio Toolbox was changed between those 2 versions, so what are you implying that I'm doing wrong here?

Well, like you said it was wrong, two versions of qaac don't change/mean anything for the decoder but now you've changed the reply and of course it makes the difference, the DLL from iTunes/QuickTime was different.
  • Last Edit: 14 September, 2012, 04:30:25 AM by eahm

  • neothe0ne
  • [*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #17
neothe0ne, you may want to research how qaac works.

Core Audio Toolbox was changed between those 2 versions, so what are you implying that I'm doing wrong here?

Well, like you said it was wrong, two versions of qaac don't change/mean anything for the decoder but now you've changed the reply and of course it makes the difference, the DLL from iTunes/QuickTime was different.


Just for the record:
Underlined is your assumption of what I was talking about (I have no idea why you or how you would assume that... because it's obviously very very very wrong)
Bolded is untrue.

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #18
neothe0ne, my last reply to this discussion.

I was just telling you that the DLL (CoreAudioToolbox.dll) is what matters for the encoding of AAC with QAAC, based on what you said here: "TVBR 91 appears to have no difference between qaac 1.39 and 1.41."

Also yes, you didn't change your reply, you changed your statement, correcting what you said before and saying: "Core Audio Toolbox was changed between those 2 versions". Of course the quality was going to be different, not because of QAAC though.


And talking about the topic again, it's funny higher encoding result in slightly smaller files, I just encoded at -V36 for a project and the files with 7.9.8.1 are bigger, Apple may have improved the quality at lower bitrates?
  • Last Edit: 14 September, 2012, 01:14:21 PM by eahm

  • IgorC
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #19
Eahm.
Can You try to do some blind tests and post them here?

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #20
IgorC, I can try but I promise you right now I will not hear anything different. AAC is so good I don't even notice anything from -V54 to Lossless, these little changes they are doing now are not for us I'm sure but for some kind of encoder optimization.

New iTunes, new CoreAudioToolbox.dll
Reply #21
I don't know if this means anything, but I converted an album back in june *using sound forge izotope to downsample and quicktime pro to encode*

I use tvbr, and granted there's a new itunes update and version of sound forge izotope since then *but not a new qt update*, I rencoded the album again with the same process and the bitrate between the two are dramatically different.

Here's an example

Song
Before 96kbps After 89

Song 2
Before 106 After 100

Song 3

Before 117 After 101

Song 4
Before 145 After 147


It's perplexing really, there's no consistency to the difference in bitrates between the 2 different encodes, the settings are exactly the same, and wondering possibly if it were an error I deleted the new encode and re-encoded all over gain twice, and the results are exactly the same. Using replaygain, the peaks are different too, quicktime uses core audio I believe, but I have no idea if it's that or the soundforge update *which updates the izotope converter too*.

If any experts can chime in I'd love to hear some theories.

  • eahm
  • [*][*][*][*][*]
New iTunes, new CoreAudioToolbox.dll
Reply #22
I would very much like to ditch Apple over Fraunhofer because they don't talk about these updates/changes.
  • Last Edit: 19 September, 2012, 07:45:55 PM by eahm

New iTunes, new CoreAudioToolbox.dll
Reply #23
I don't think much can be said other than the results that were found conducting that test.  It might be all for not though since iTunes 11 is about a month away from being released.  Even then, that might not change anything with the encoder that Apple is implementing.  I have my reservations about upgrading to iTunes since it appears they are getting rid of the list view of an entire library but that is a different discussion.


List view will still be there, slightly different but it's still there.
  • Last Edit: 19 September, 2012, 09:02:32 PM by RobertoDomenico

New iTunes, new CoreAudioToolbox.dll
Reply #24
I've went back and re-encoded other albums.

Allot show significantly lower bitrates on average than before.

What this tells me is the tweaks on coreaudio must've directly effected quicktime too, because it hasn't been updated in months but I'm already seeing a difference in bitrates.

Stuff that will be 139 will become 127, 130 will become 114, etc...

Some songs will show an increase in bitrate by like 4, but the drastic difference is in the bitrate drop.

I've used quicktime for a few years and this is the most drastic change in bitrate I've seen.

An album that was 56.3mb before now became 53mb.

Now, these albums I'm comparing are my personal needledrops and as I've mentioned before sound forge has been updated also *I downsample with izotope*, however in my experience downsampling in the past after the update while the peaks do change in the files the bitrates stay virtually the same. So I think this is all coreaudio at work...

I would love to hear an enthusiast chime in.
  • Last Edit: 19 September, 2012, 10:30:21 PM by Mix3dmessagez