Help - Search - Members - Calendar
Full Version: FLAC 1.1.4 command line in foobar2000 0.8.3
Hydrogenaudio Forums > Hosted Forums > foobar2000 > Support - (fb2k)
Music Man Mark
Hi all - long time reader, permanent learner. Here is what I'm struggling with, and have searched my tail off but cannot find an answer (first time in years that searching didn't get the solution).

I am using PlexToolsLE 3.13 to rip CDs to .wav files, then encode to FLAC 1.1.4 via foobar 2000 v0.8.3 on my HDD. Foobar performs the encoding but when the progress bar shows it is complete (a slight pause, the progress window does not close), I get the message:


QUOTE
ERROR (foo_clienc) : Encoding failed


The command line I am using is (not worried about tags, I will use directory structure to extract later):

CODE
-5 %s %d


Any thoughts? I must be missing something pretty easy.

Thanks,
Mark

Moderation: Moved to "Support - (fb2k)". Duplicate topic deleted.
JensRex
Upgrade to latest foobar2000, then ask again.
Music Man Mark
QUOTE(JensRex @ Oct 30 2007, 13:18) *

Upgrade to latest foobar2000, then ask again.


Sorry about the duplicate post, but is the only answer to upgrade foobar? I thought using a command line encoder bypassed version issues with programs, maybe that's misinformation on my part.

Thanks,
Mark
JensRex
The only answer might not be to upgrade foobar2000, but I'd venture a guess that most people are disinclined to provide support for a version that has been outdated for a very very long time.

Check out the latest beta version. I promise you'll like it.
guruboolez
Try this:
-5 -V - -o %d


N.B. foobar2000 0.83 is working better than 0.9x with flac.exe so it can make sense using it (with pipe encoding at least):
http://www.hydrogenaudio.org/forums/index....showtopic=43160
http://www.hydrogenaudio.org/forums/index....showtopic=41287

Difference is of course very small...
Music Man Mark
QUOTE(guruboolez @ Oct 30 2007, 14:54) *

Try this:
-5 -V - -o %d


N.B. foobar2000 0.83 is working better than 0.9x with flac.exe so it can make sense using it (with pipe encoding at least):
http://www.hydrogenaudio.org/forums/index....showtopic=43160
http://www.hydrogenaudio.org/forums/index....showtopic=41287

Difference is of course very small...



That worked! Thank you so much guruboolez, that's a huge help.

Mark
Peter
QUOTE(guruboolez @ Oct 30 2007, 22:54) *
N.B. foobar2000 0.83 is working better than 0.9x with flac.exe so it can make sense using it (with pipe encoding at least):
http://www.hydrogenaudio.org/forums/index....showtopic=43160
http://www.hydrogenaudio.org/forums/index....showtopic=41287

Difference is of course very small...
Not true for 0.9.5 + latest FLAC encoder. With older 0.9.x versions you can workaround this problem by using FLAC 1.2.0 or newer and adding "--ignore-chunk-sizes" switch.
guruboolez
Weird. I made a small test before posting my previous answer to be sure that nothing changed in the meantime, and it confirmed that foobar2000 0.83 output was ~20kb smaller than foobar2000 0.95b2's one (FLAC 1.2.1):

0.83: 32,0 Mo [Mb] (33 650 573 octets [bytes])
0.95b2: 32,1 Mo [Mb] (33 671 579 octets [bytes])

setting (for both fb2k version): -8 -V -P 16000 - -o %d
Same path to flac.exe for both foobar2000.


Shouldn't the output be the same? Should I add an additional switch?


EDIT:
Checked the number of seekpoint with metaflac:

0.83
CODE

  channels: 2
  bits-per-sample: 16
  total samples: 22461600
  MD5 signature: 6a1e8f112bd151cc0fe3e6ef2ad1c81a
METADATA block #1
  type: 3 (SEEKTABLE)
  is last: false
  length: 918
  seek points: 51



0.95b2
CODE
  channels: 2
  bits-per-sample: 16
  total samples: 22461600
  MD5 signature: 6a1e8f112bd151cc0fe3e6ef2ad1c81a
METADATA block #1
  type: 3 (SEEKTABLE)
  is last: false
  length: 21924
  seek points: 1218
Peter
Because the source stream has unknown length, FLAC encoder has to reserve extra space for a potentially large seektable. To produce identical files as with known source stream length, the encoder would have to rewrite entire file again when writing the seektable at the end of the encode process.
I honestly fail to see why people even care about this kind of size differences. Seektables in output files are now strictly valid, as far as I am aware of (we're telling the FLAC encoder what to expect).
guruboolez
I think I don't understand. The “problem” isn't fixed, right? Not fully at least?
I agree with you and don't think that 20 kb is worrying enough, especially with lossless, especially with flac (which is far to be the strongest compressor).

N.B. --ignore-chunk-sizes with 1.2.1 seems to discard the seektable (metaflac doesn't mention anymore the seekpoint number and details).
Peter
If you want "the problem" fixed, I suggest using a codec that accepts unknown-length streams without strange behaviors, like WavPack.
The 0.8.3 solution isn't good either because it's unreliable, the output may be lossy/truncated when reencoding from a format that does not give accurate length info such as WMA lossless.
guruboolez
... or using %s instead of - in order to avoid all possible “problems” smile.gif

Thanks for all these clarifications.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2008 Invision Power Services, Inc.