QAAC: discussion, questions, feature requests, etc., [originally a thread for a feature request] Options

[nu774]

Hey, does tvbr support 16 bitor 24 bit depth?

The concept of bit depths is not defined for AAC, at least in the same sense as of PCM.
--bits-per-sample of qaac is for WAV/ALAC output only; It will be simply ignored for AAC.

[pururin]

Does anyone knows whether Apple's aac (coreaudiotoolbox.dll) get any update recently?
As I read the changelog and found nothing for long time. Thanks.

[nu774]

Does anyone knows whether Apple's aac (coreaudiotoolbox.dll) get any update recently?
As I read the changelog and found nothing for long time. Thanks.

On 7.9.7.8 (QuickTime 7.7.1 / iTunes10.5), there were updates on AAC encoder, and also a function named ACMP4AACHighEfficiencyEncoderFactory was added to the DLL, which enables iTunes10.5 to run HE-AAC codec without QuickTime installation.
AFAIK that was the last update on AAC encoder so far.

[pururin]

Thanks a lot for the fast reply, nu774.

[kareha]

Hmm, I've got a CoreAudioToolBox.dll version 7.9.7.9 dated 20.02.12, not sure what was updated in this one though.

[nu774]

Hmm, I've got a CoreAudioToolBox.dll version 7.9.7.9 dated 20.02.12, not sure what was updated in this one though.

Yes, it is the latest. Actually, there are at least two different binaries under the same version number "7.9.7.9". Same for 7.9.7.8.
As far as I can see, AAC encoder is not updated, though.

By the way, I sent PM to squo (a developper of QuickTime AAC) about HE-AAC iTuneSMPB problem in the past, but I haven't received a reply.
At the time of 7.9.7.8 update he was still there, but he might have been transferred/retired or something.

This post has been edited by nu774: Apr 23 2012, 02:14

[pururin]

skuo retired? That's sad.

Thought about the pre-echo on sharp transients issue, have it been fixed in the latter version?

[nu774]

Thought about the pre-echo on sharp transients issue, have it been fixed in the latter version?

I only know that output of Avisynth ColorBars() function was a killer sample to QT AAC in the past, and was fixed on 7.9.7.8 update.
ColorBars() mainly outputs SMPTE color bars (video test pattern), but also outputs audio test tone, which is repeatedly turned on/off once every seconds. QT AAC was producing quite obvious artifacts at the boundaries even at the highest settings.

[pururin]

I only know that output of Avisynth ColorBars() function was a killer sample to QT AAC in the past, and was fixed on 7.9.7.8 update.

It seems some issues that was reported got fixed on 7.9.7.8.
I'm appreciate that the developers are quite active.(qaac too, Many thanks to nu774 ) A little wonder when the next update will come.
About pre-echo, may have to ask someone like /mnt.

[2304p]

I only know that output of Avisynth ColorBars() function was a killer sample to QT AAC in the past, and was fixed on 7.9.7.8 update.

It seems some issues that was reported got fixed on 7.9.7.8.
I'm appreciate that the developers are quite active.(qaac too, Many thanks to nu774 ) A little wonder when the next update will come.
About pre-echo, may have to ask someone like /mnt.

I have iTunes 10.6.1.7 and still coreaudiotoolbox 7.9.7.3

question:
How to get a current coreaudiotollbox version?
How to updated it?

[nu774]

I have iTunes 10.6.1.7 and still coreaudiotoolbox 7.9.7.3

question:
How to get a current coreaudiotollbox version?
How to updated it?

Try

CODE

qaac --check

from command prompt, instead of explorer property window.

CoreAudioToolbox.dll is internationalized, and contains multiple resources for different locales.
The problem is, for many locales Apple has not been properly updating version resources, and for such locales Windows explorer will show file version as 7.9.7.3;
On the other hand, qaac --check will always pick version number from en-US locale resource.

[nu774]

Basically speaking, if you are using qaac 1.x branch and you can use HE-AAC encoder, your CoreAudioToolbox.dll is 7.9.7.8 or newer.

[2304p]

I have iTunes 10.6.1.7 and still coreaudiotoolbox 7.9.7.3

question:
How to get a current coreaudiotollbox version?
How to updated it?

Try

CODE

qaac --check

from command prompt, instead of explorer property window.

CoreAudioToolbox.dll is internationalized, and contains multiple resources for different locales.
The problem is, for many locales Apple has not been properly updating version resources, and for such locales Windows explorer will show file version as 7.9.7.3;
On the other hand, qaac --check will always pick version number from en-US locale resource.

thanks

[Dario]

Is there a place where I can download the CoreAudioToolbox library independently? I am unable to download iTunes at the moment, as my connection is very slow.

[kareha]

Is there a place where I can download the CoreAudioToolbox library independently? I am unable to download iTunes at the moment, as my connection is very slow.

Uploaded 7.9.7.9 to my Dropbox:

http://dl.dropbox.com/u/26532689/CoreAudioToolbox.dll

[Dario]

Thank you very much!

[Anakunda]

Hiyas, any seem of approximate quality relation table between QuickTime TVBR values and LAME V x values, eventually Vorbis Q values?
Something that look like
Lame V4 ≈ TVBR 40
Lame V2 ≈ TVBR 60
Lame V0 ≈ TVBR 80
etc..

[LigH]

Not really. The three algorithms differ in specific cases, it may be hard to rate them generally, and the only "reliable" way would be a mass ABX test in quality ranges most people would call "transparent".

As listed in this post about Vorbis accelleration, my own humble opinion is that qaac -V 80 might be comparable to oggenc2 -q 5, based on the assumption that they both are close in some bitrate/quality quotient. The MP3 technology is less efficient compared to both of them, may need about 20-30% more bitrate for a similar quality, and I found that lame -V 2 matches this range quite well.

The opinion above is personal and weak. Other members here may have more founded facts.

[LigH]

I exchanged the CoreAudioToolbox.dll from a freshly installed QuickTime 7.7.2 (DLL file date: Feb. 20 2012) with the one kareha uploaded. "qaac --check" still reports version 7.9.7.3 (on a german Windows XP SP3).

When I realized they have the same size, I did a byte comparison, and guess what ... I found they were identical.

I opened it in ResHacker and checked the "Version Info" tables. There are several for different language codes: 1033 = "7.9.7.9"; 1025 = "7.9.7.7"; 1028,1030,1031,... = "7.9.7.3" (and on top, they mix up InternalName and OriginalFileName vice versa).

So ... don't care about numbers if you don't use an english Windows.
__

P.S.:

Oh, I did the check with an older qaac version; qaac 1.37 reports CoreAudioToolbox version 7.9.7.9 correctly.

This post has been edited by LigH: May 31 2012, 09:37

[eahm]

1.38 released. I finally completely switched from every other encoding software to just foobar2000 + qaac (TVBR 91/100), thank you dev/s!

Changelog:
Updated libsoxrate to 0.21 (merged upstream update on SoX rate effect).
Fixed not to write WAVE_FORMAT_EXTENSIBLE header unnecessarily on WAV output.

https://sites.google.com/site/qaacpage/cabinet

This post has been edited by eahm: Jun 9 2012, 19:23

[Anakunda]

I finally completely switched from every other encoding software to just foobar2000 + qaac (TVBR 91/100), thank you dev/s!

Is there noticeable difference between TVBR 91 and 100, and can you ABX tvbr 91 from original?

[eahm]

I finally completely switched from every other encoding software to just foobar2000 + qaac (TVBR 91/100), thank you dev/s!

Is there noticeable difference between TVBR 91 and 100, and can you ABX tvbr 91 from original?

You'll be amazed how low you can go with AAC. I keep ~192/224 just to be safe

This post has been edited by eahm: Jun 9 2012, 19:38

[LigH]

IMHO, TVBR 82 is similar to Vorbis q5, that seems to be enough for "easy listening"; see the Vorbis CPU optimization thread for an explanation...

This post has been edited by LigH: Jun 9 2012, 19:44

[eahm]

IMHO, TVBR 82 is similar to Vorbis q5, that seems to be enough for "easy listening"; see the Vorbis CPU optimization thread for an explanation...

I don't go lower than 160Kbps. TVBR 91 is the lowest I go. I keep ~1000 songs on my iPhone, no space problem.

This post has been edited by eahm: Jun 9 2012, 19:52

[LigH]

Don't care about bitrates. 160 kbps silence is not "better" than 32 kbps silence. There is easily compressible sound which you can't directly compare with harder "noise". If the quality mode ensures a threshold of maximum loss, trust in it.