Skip to main content

Topic: QAAC: discussion, questions, feature requests, etc. (Read 357519 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.
  • Big_Berny
  • [*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #400
Great, thanks a lot! 
The online problem I have now is that I can't seek. The stream Ing app dsub shows the correct length but the progressbar doesn't work.

Ok, this one is fixed too. Problem is solved when song is fully loaded/cached.

Last thing: afaik replaygain is lost when decoding aac, right? Is it possible to change the volume when decoding mp3 and aac by adding replaygain correction directly to the raw stream? As I read it's possible to change volume when decoding a flac to wave with flac-codec but is this also possible with a decoder for mp3 and aac?
  • Last Edit: 09 May, 2013, 07:39:27 PM by Big_Berny

  • Big_Berny
  • [*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #401
Last thing: afaik replaygain is lost when decoding aac, right? Is it possible to change the volume when decoding mp3 and aac by adding replaygain correction directly to the raw stream? As I read it's possible to change volume when decoding a flac to wave with flac-codec but is this also possible with a decoder for mp3 and aac?

Made a new topic as it has nothing to do with qaac anymore. (Unfortunately I can't edit last post anymore)

  • Pulstar
  • [*]
QAAC: discussion, questions, feature requests, etc.
Reply #402
I've noticed that with Apple Application Support v2.3.4 the encoding is slower even with multithreading enabled. Not that big of an issue considering audio takes less cycles to encode than video but I'm curious as to why. Perhaps it's just my machine?
  • Last Edit: 23 May, 2013, 08:32:45 PM by Pulstar

  • deej_1977
  • [*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #403
--- Problems that solve themselves are great, carry on, as you were :-) ---
  • Last Edit: 15 June, 2013, 04:53:50 PM by deej_1977
No inspiration

  • Boulder
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #404
I'm having problems encoding the audio track of "Edward Scissorhands" with qaac. eac3to reports it as 3/1 which I believe means FL, C and FR channels for the front and BC for the back. qaac assumes "FL FR BL BR" which causes incorrect channel mapping. I couldn't find what chanmask parameter I should use, so can you help me with this?

  • nu774
  • [*][*][*][*][*]
  • Developer
QAAC: discussion, questions, feature requests, etc.
Reply #405
I'm having problems encoding the audio track of "Edward Scissorhands" with qaac. eac3to reports it as 3/1 which I believe means FL, C and FR channels for the front and BC for the back. qaac assumes "FL FR BL BR" which causes incorrect channel mapping. I couldn't find what chanmask parameter I should use, so can you help me with this?

If it is actually "FL FR FC BC" but just channel mask is missing in the wav header, use --chanmask=0x107.

  • nu774
  • [*][*][*][*][*]
  • Developer
QAAC: discussion, questions, feature requests, etc.
Reply #406
You can calculate channel mask by summing the following values.
In this case, FL + FR + FC + BC == 1 + 2 + 4 + 256 == 263 == 0x107.
You can give both hex value (with 0x prefix) and decimal value to --chanmask option.
Code: [Select]
FL == 1<<0 == 1 == 0x1
FR == 1<<1 == 2 == 0x2
FC == 1<<2 == 4 == 0x4
LF == 1<<3 == 8 == 0x8
BL == 1<<4 == 16 == 0x10
BR == 1<<5 == 32 == 0x20
FLC == 1<<6 == 64 == 0x40
FRC == 1<<7 == 128 == 0x80
BC == 1<<8 == 256 == 0x100
SL == 1<<9 == 512 == 0x200
SR == 1<<10 == 1024 == 0x400

  • Boulder
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #407
Thanks, that table is really useful. Got it working now

  • eahm
  • [*][*][*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #408
[qaac] release 2.19 (refalac 1.19)
posted 2 hours ago by nu 774

Fixed: attempt to set one of stik, rtng, akID, sfID tags caused qaac hang.

Well, actually not hanging but waiting for console input in vain... due to a silly bug calling scanf() instead of sscanf().

Fixed: --tag akID:fra was writing USA country code (not France).

https://sites.google.com/site/qaacpage/
  • Last Edit: 13 July, 2013, 12:33:07 AM by eahm

  • eahm
  • [*][*][*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #409
qaac.exe --tvbr 87 --quality 2 --ignorelength - -o f:\temp\captures\test_qaac.m4a

Is the ignorelength switch ever really needed?
  • Last Edit: 24 July, 2013, 12:56:58 PM by eahm

  • nu774
  • [*][*][*][*][*]
  • Developer
QAAC: discussion, questions, feature requests, etc.
Reply #410
qaac.exe --tvbr 87 --quality 2 --ignorelength - -o f:\temp\captures\test_qaac.m4a

Is the ignorelength switch ever really needed?

Well, it depends.
If you are encoding from fb2k, and if blockalign (length per one sample, in bytes) of the input is even (usually it will be even), qaac will work as if --ignorelength is specified, and surely you don't need to explicitly specify that switch.

More precisely, qaac will work in ignorelength mode in the following cases:
  • Length of data (in headr) is equal to zero.
  • Length of data is not divisible by blockalign.


  • Carsi
  • [*]
QAAC: discussion, questions, feature requests, etc.
Reply #411
Hi.

I wanted to ask if I can convert from FLAC/ALAC to m4a without creating temporary wav first? It really slows down the process.

My parameters:

--tvbr 53 --no-optimize -o %d %s

I tried to remove %s but then it displays en error message?

And yes I went down til tvbr 53, I think it sounds great, want to store as much as possible on my 32gb iphone 5

  • detmek
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #412
If you are using foobar try this:
--tvbr 53 --no-optimize --ignorelength - -o %d
You need to replace %s with -.

  • eahm
  • [*][*][*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #413
And the number is 54 anyway to be even more precise.

Code: [Select]
AAC (Apple True VBR / qaac)

Q0 - Q4 (0) = ~40 kbps
Q5 - Q13 (9) = ~45 kbps
Q14 - Q22 (18) = ~75 kbps
Q23 - Q31 (27) = ~80 kbps
Q32 - Q40 (36) = ~95 kbps
Q41 - Q49 (45) = ~105 kbps
Q50 - Q58 (54) = ~115 kbps
Q59 - Q68 (63) = ~135 kbps
Q69 - Q77 (73) = ~150 kbps
Q78 - Q86 (82) = ~165 kbps
Q87 - Q95 (91) = ~195 kbps
Q96 - Q104 (100) = ~225 kbps
Q105 - Q113 (109) = ~255 kbps
Q114 - Q122 (118) = ~285 kbps
Q123 - Q127 (127) = ~320 kbps
  • Last Edit: 05 August, 2013, 12:28:09 PM by eahm

  • Boulder
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #414
Has anyone tested the compression efficiency of multichannel ALAC?

I was thinking of switching from FLAC to ALAC for my movie archives (because mkvtoolnix sometimes breaks FLAC files when segmented mkv files are created), but some multichannel ALAC files are huge.

For example, the audio track of Notting Hill is ~970MB with FLAC but 2,9GB with ALAC. Some multichannel (5.1ch) tracks are very close to the FLAC file in size.

  • db1989
  • [*][*][*][*][*]
  • Global Moderator
QAAC: discussion, questions, feature requests, etc.
Reply #415
the audio track of Notting Hill is ~970MB with FLAC but 2,9GB with ALAC.
This seems peculiar, well beyond the usual expected variance between lossless codecs. Did you encode both of these yourself, yes? Which encoder was used to create the ALAC? Have you observed this trend on other files compressed with ALAC compared to other lossless codecs? I just find it hard to believe this can be normal. And I cannot recall reading of any such large bugs in ALAC.

  • Boulder
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #416
Yes, they are my encodes. I used refalac, which is in the QAAC package. There's nothing wrong with the file what comes to playback, there's no static or anything else that could explain the huge size.

  • eahm
  • [*][*][*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #417
Yes, they are my encodes. I used refalac, which is in the QAAC package. There's nothing wrong with the file what comes to playback, there's no static or anything else that could explain the huge size.

Wow, I have to try a multichannel track. Two days ago I converted ~31000 audio files from FLAC to ALAC. If they come up with a 128GB iPhone I am done going lossy.

edit:
Downloading the first, Mozart (FLAC 470MB) from here: http://www.2l.no/hires/

edit2:
ALAC is 494MB (TAK -p1 456MB, TAK -p2 454MB, TAK -p4m 447MB, WavPack normal 469MB, TTA 509MB, WMA lsl 494MB, WAV 929MB)
  • Last Edit: 16 August, 2013, 11:17:14 AM by eahm

  • lvqcl
  • [*][*][*][*][*]
  • Developer
QAAC: discussion, questions, feature requests, etc.
Reply #418
Maybe the file is 24 bit but the real bit depth is less?

  • Boulder
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #419
That could be the case here. This is what eac3to reports when decoding the FLAC file: Original audio track: max 24 bits, average 17 bits, most common 16 bits. Apparently ALAC is not as efficient in such cases.

  • testyou
  • [*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #420
But 3 times as large?  I'm guessing there's something else going on there.

  • nu774
  • [*][*][*][*][*]
  • Developer
QAAC: discussion, questions, feature requests, etc.
Reply #421
Simply put, when valid bit depth is 16/24 and compression ratio is around 25% for both codec on the effective 16bits part, ALAC file will become 3 times larger than FLAC one.
24bit ALAC file always stores LSB-side 8bits uncompressed, while FLAC won't consume bits due to it's wasted bits feature in this case.
Code: [Select]
16×0.25 : 16×0.25+8 = 1 : 3


  • sluggy
  • [*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #422
[qaac] release 2.20 + refalac 1.20
posted 2 hours ago by nu 774
Add optional libsoxr support. It's basically the same as libsoxrate (both derives from SoX), but is more optimized (fast), and is maintained by a developer of SoX.

When libsoxr.dll is present, now qaac/refalac will use it for sample rate conversion. libsoxr binary is on the cabinet page. Note that you still need libsoxrate when you want --lowpass or mixing option.

libsoxr binary built with GCC doesn't usually work with qaac/refalac due to a few ABI compatibility issues. The binary at cabinet page is built with GCC _with care_.

I have reported this issue to author of libsoxr, so it might be fixed in the future.

Decreased refresh rate of progress on title bar of console window.

Explicitly check presence of BOM when reading text files, since MLang often does a wrong guess even when BOM is present.

  • testyou
  • [*][*][*]
QAAC: discussion, questions, feature requests, etc.
Reply #423
When I have libsoxr.dll in the directory and run "--check", it fails and says libgcc_s_sjlj-1.dll is missing.

  • nu774
  • [*][*][*][*][*]
  • Developer
QAAC: discussion, questions, feature requests, etc.
Reply #424
When I have libsoxr.dll in the directory and run "--check", it fails and says libgcc_s_sjlj-1.dll is missing.

Oh thanks for reporting it. Updated libsoxr archive on the cabinet page.
libsoxr is a pure C DLL and I can't understand why it needs to be dependent on that DLL, but this might be the case: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57120.