[X] My Assistant

[jcoalson]

edit: Kent, your original question is not clear, flac can encode or decode, and can do different things with replaygain for each.  what do you mean by interpret?

When I use flac -d to decode a file, will it read the embedded ReplayGain tags and apply them when decoding?

gotcha. yes, but only with --apply-replaygain-which-is-not-lossless which is not yet released.

All right, but based on this can one specify a relative gain adjustment for files which do not already have an existing gain tag? (That part still seems unclear. If I have FLAC without gain information stored in the tags, but I know what the gain/scale information is from having previously analyzed the WAVs in WaveGain and storing a log, can I say, "I want this much gain applied - don't ask questions!"?)

if there is no gain tag, the option will have no effect. here's the usage:

CODE

--apply-replaygain-which-is-not-lossless[=<specification>]
   Applies ReplayGain values while decoding.

   WARNING: THIS IS NOT LOSSLESS.  DECODED AUDIO WILL NOT BE IDENTICAL TO THE ORIGINAL WITH THIS OPTION.

   The equals sign and <specification> is optional.  If omitted, the default is 0aLn1 (see below).

   The <specification> is a shorthand notation for describing how to apply ReplayGain.  All components are optional but order is important.  '[]' means 'optional'.  '|' means 'or'.  '{}' means required.  The format is:

       [<preamp>][a|t][l|L][n{0|1|2|3}]

   <preamp>
       A floating point number in dB.  This is added to the existing gain value.

   a|t
       Specify 'a' to use the album gain, or 't' to use the track gain

   l|L
       Specify 'l' to peak-limit the output, so that the ReplayGain peak value is full-scale.  Specify 'L' to use a 6dB hard limiter that kicks in when the signal approaches full-scale.

   n{0|1|2|3}
       Specify the amount of noise shaping.  ReplayGain synthesis happens in floating point; the result is dithered before converting back to integer.  This quantization adds noise.  Noise shaping tries to move the noise where you won't hear it as much.  0 means no noise shaping, 1 means 'low', 2 means 'medium', 3 means 'high'.

   For example, the default of 0aLn1 means 0dB preamp, use album gain, 6dB hard limit, low noise shaping.

   --apply-replaygain-which-is-not-lossless=3 means 3dB preamp, use album gain, no limiting, no noise shaping.

   flac uses the ReplayGain tags for the calculation.  If a stream does not have the required tags or they can't be parsed, decoding will continue with a warning, and no ReplayGain is applied to that stream.

... and for us non-programmers, is there a downloadable CVS compile with the --apply-replaygain-which-is-not-lossless switch enabled?

nope, sorry. Cerebus was asking about it in another thread, for use with the Slim Server; maybe those guys have a binary.

Josh

[john33]

For anyone that wants to play, I've just uploaded a complete set of compiles of the current FLAC CVS to Rarewares/Lossless page.

Be Aware these are untested and, unless Josh says otherwise, should not be used for anything other than testing new features.

[jcoalson]

yes, and especially don't encode to ogg flac with it; the files will be silently corrupt as I still have some fixes to make.

Josh

[atici]

Hmm does this new version feature any size improvements, i.e. smaller filesize for same quality?

[Cerebus]

john33:

I don't see the
--apply-replaygain-which-is-not-lossless[=<specification>]

option using the binaries from Rarewares...does it not come up with --explain or --help?

C

[jcoalson]

Hmm does this new version feature any size improvements, i.e. smaller filesize for same quality?

quality is always "lossless" but there are no compression or major speed improvements.

john33:

I don't see the
--apply-replaygain-which-is-not-lossless[=<specification>]

option using the binaries from Rarewares...does it not come up with --explain or --help?

--apply-replaygain-which-is-not-lossless is not documented and won't show up in the help. if you want to know if it's there or not, try --apply-replaygain-which-is-not-lossless=dfjsdkfjsl and see if it prints an error about the specification, or unknown option.

Josh

[Cerebus]

Is there any way that we can get this added to FLACDrop, in the same sort of fashion that Musedrop has the 'custom decoding' option? I don't know how to get a hold of the author (Layer3Maniac).

[MrBubble1337]

Hmm does this new version feature any size improvements, i.e. smaller filesize for same quality?

quality is always "lossless" but there are no compression or major speed improvements.

john33:

I don't see the
--apply-replaygain-which-is-not-lossless[=<specification>]

option using the binaries from Rarewares...does it not come up with --explain or --help?

--apply-replaygain-which-is-not-lossless is not documented and won't show up in the help. if you want to know if it's there or not, try --apply-replaygain-which-is-not-lossless=dfjsdkfjsl and see if it prints an error about the specification, or unknown option.

Josh

God damn it, I'd been debugging a weird issue with FLAC where I hadn't been able to find documentation for an option in the --help MAN, never even thought to check whether applying the option directly would throw an error for the option specification or an unknown option error. Feel like a moron but that post just saved me a hell of a lot of time and confusion - cheers! Looks like a decent forum too, think I might stick around