furnace/extern/flac/man/flac.1

.\" Automatically generated by Pandoc 3.1.3
.\"
.\" Define V font for inline verbatim, using C font in formats
.\" that render this, and otherwise B font.
.ie "\f[CB]x\f[]"x" \{\
. ftr V B
. ftr VI BI
. ftr VB B
. ftr VBI BI
.\}
.el \{\
. ftr V CR
. ftr VI CI
. ftr VB CB
. ftr VBI CBI
.\}
.TH "flac" "1" "" "Version 1.5.0" "Free Lossless Audio Codec conversion tool"
.hy
.SH NAME
.PP
flac - Free Lossless Audio Codec
.SH SYNOPSIS
.PP
\f[B]flac\f[R] [ \f[I]OPTIONS\f[R] ] [ \f[I]infile.wav\f[R] |
\f[I]infile.rf64\f[R] | \f[I]infile.aiff\f[R] | \f[I]infile.raw\f[R] |
\f[I]infile.flac\f[R] | \f[I]infile.oga\f[R] | \f[I]infile.ogg\f[R] |
\f[B]-\f[R] \f[I]\&...\f[R] ]
.PP
\f[B]flac\f[R] [ \f[B]-d\f[R] | \f[B]--decode\f[R] | \f[B]-t\f[R] |
\f[B]--test\f[R] | \f[B]-a\f[R] | \f[B]--analyze\f[R] ] [
\f[I]OPTIONS\f[R] ] [ \f[I]infile.flac\f[R] | \f[I]infile.oga\f[R] |
\f[I]infile.ogg\f[R] | \f[B]-\f[R] \f[I]\&...\f[R] ]
.SH DESCRIPTION
.PP
\f[B]flac\f[R] is a command-line tool for encoding, decoding, testing
and analyzing FLAC streams.
.SH GENERAL USAGE
.PP
\f[B]flac\f[R] supports as input RIFF WAVE, Wave64, RF64, AIFF, FLAC or
Ogg FLAC format, or raw interleaved samples.
The decoder currently can output to RIFF WAVE, Wave64, RF64, or AIFF
format, or raw interleaved samples.
flac only supports linear PCM samples (in other words, no A-LAW, uLAW,
etc.), and the input must be between 4 and 32 bits per sample.
.PP
flac assumes that files ending in \[lq].wav\[rq] or that have the RIFF
WAVE header present are WAVE files, files ending in \[lq].w64\[rq] or
have the Wave64 header present are Wave64 files, files ending in
\[lq].rf64\[rq] or have the RF64 header present are RF64 files, files
ending in \[lq].aif\[rq] or \[lq].aiff\[rq] or have the AIFF header
present are AIFF files, files ending in \[lq].flac\[rq] or have the FLAC
header present are FLAC files and files ending in \[lq].oga\[rq] or
\[lq].ogg\[rq] or have the Ogg FLAC header present are Ogg FLAC files.
.PP
Other than this, flac makes no assumptions about file extensions, though
the convention is that FLAC files have the extension \[lq].flac\[rq] (or
\[lq].fla\[rq] on ancient \[lq]8.3\[rq] file systems like FAT-16).
.PP
Before going into the full command-line description, a few other things
help to sort it out:
.IP "1." 3
flac encodes by default, so you must use -d to decode
.IP "2." 3
Encoding options -0 ..
-8 (or --fast and --best) that control the compression level actually
are just synonyms for different groups of specific encoding options
(described later).
.PD 0
.P
.PD
.IP "3." 3
The order in which options are specified is generally not important
except when they contradict each other, then the latter takes precedence
except that compression presets are overridden by any option given
before or after.
For example, -0M, -M0, -M2 and -2M are all the same as -1, and -l 12 -6
the same as -7.
.IP "4." 3
flac behaves similarly to gzip in the way it handles input and output
files
.PP
Skip to the EXAMPLES section below for examples of some typical tasks.
.PP
flac will be invoked one of four ways, depending on whether you are
encoding, decoding, testing, or analyzing.
Encoding is the default invocation, but can be switch to decoding with
\f[B]-d\f[R], analysis with \f[B]-a\f[R] or testing with \f[B]-t\f[R].
Depending on which way is chosen, encoding, decoding, analysis or
testing options can be used, see section OPTIONS for details.
General options can be used for all.
.PP
If only one inputfile is specified, it may be \[lq]-\[rq] for stdin.
When stdin is used as input, flac will write to stdout.
Otherwise flac will perform the desired operation on each input file to
similarly named output files (meaning for encoding, the extension will
be replaced with \[lq].flac\[rq], or appended with \[lq].flac\[rq] if
the input file has no extension, and for decoding, the extension will be
\[lq].wav\[rq] for WAVE output and \[lq].raw\[rq] for raw output).
The original file is not deleted unless --delete-input-file is
specified.
.PP
If you are encoding/decoding from stdin to a file, you should use the -o
option like so:
.IP
.nf
\f[C]
flac [options] -o outputfile
flac -d [options] -o outputfile
\f[R]
.fi
.PP
which are better than:
.IP
.nf
\f[C]
flac [options] > outputfile
flac -d [options] > outputfile
\f[R]
.fi
.PP
since the former allows flac to seek backwards to write the STREAMINFO
or RIFF WAVE header contents when necessary.
.PP
Also, you can force output data to go to stdout using -c.
.PP
To encode or decode files that start with a dash, use -- to signal the
end of options, to keep the filenames themselves from being treated as
options:
.IP
.nf
\f[C]
flac -V -- -01-filename.wav
\f[R]
.fi
.PP
The encoding options affect the compression ratio and encoding speed.
The format options are used to tell flac the arrangement of samples if
the input file (or output file when decoding) is a raw file.
If it is a RIFF WAVE, Wave64, RF64, or AIFF file the format options are
not needed since they are read from the file\[cq]s header.
.PP
In test mode, flac acts just like in decode mode, except no output file
is written.
Both decode and test modes detect errors in the stream, but they also
detect when the MD5 signature of the decoded audio does not match the
stored MD5 signature, even when the bitstream is valid.
.PP
flac can also re-encode FLAC files.
In other words, you can specify a FLAC or Ogg FLAC file as an input to
the encoder and it will decoder it and re-encode it according to the
options you specify.
It will also preserve all the metadata unless you override it with other
options (e.g.
specifying new tags, seekpoints, cuesheet, padding, etc.).
.PP
flac has been tuned so that the default settings yield a good speed vs.
compression tradeoff for many kinds of input.
However, if you are looking to maximize the compression rate or speed,
or want to use the full power of FLAC\[cq]s metadata system, see the
page titled `About the FLAC Format' on the FLAC website.
.SH EXAMPLES
.PP
Some typical encoding and decoding tasks using flac:
.SS Encoding examples
.TP
\f[V]flac abc.wav\f[R]
Encode abc.wav to abc.flac using the default compression setting.
abc.wav is not deleted.
.TP
\f[V]flac --delete-input-file abc.wav\f[R]
Like above, except abc.wav is deleted if there were no errors.
.TP
\f[V]flac --delete-input-file -w abc.wav\f[R]
Like above, except abc.wav is deleted if there were no errors and no
warnings.
.TP
\f[V]flac --best abc.wav\f[R] or \f[V]flac -8 abc.wav\f[R]
Encode abc.wav to abc.flac using the highest compression preset.
.TP
\f[V]flac --verify abc.wav\f[R] or \f[V]flac -V abc.wav\f[R]
Encode abc.wav to abc.flac and internally decode abc.flac to make sure
it matches abc.wav.
.TP
\f[V]flac -o my.flac abc.wav\f[R]
Encode abc.wav to my.flac.
.TP
\f[V]flac abc.aiff foo.rf64 bar.w64\f[R]
Encode abc.aiff to abc.flac, foo.rf64 to foo.flac and bar.w64 to
bar.flac
.TP
\f[V]flac *.wav *.aif?\f[R]
Wildcards are supported.
This command will encode all .wav files and all .aif/.aiff/.aifc files
(as well as other supported files ending in .aif+one character) in the
current directory.
.TP
\f[V]flac abc.flac --force\f[R] or \f[V]flac abc.flac -f\f[R]
Recompresses, keeping metadata like tags.
The syntax is a little tricky: this is an \f[I]encoding\f[R] command
(which is the default: you need to specify -d for decoded output), and
will thus want to output the file abc.flac - which already exists.
flac will require the --force or shortform -f option to overwrite an
existing file.
Recompression will first write a temporary file, which afterwards
replaces the old abc.flac (provided flac has write access to that file).
The above example uses default settings.
More often, recompression is combined with a different - usually higher
- compression option.
Note: If the FLAC file does not end with .flac - say, it is abc.fla -
the -f is not needed: A new abc.flac will be created and the old kept,
just like for an uncompressed input file.
.TP
\f[V]flac --tag-from-file=\[dq]ALBUM=albumtitle.txt\[dq] -T \[dq]ARTIST=Queen\[dq] *.wav\f[R]
Encode every .wav file in the directory and add some tags.
Every file will get the same set of tags.
Warning: Will wipe all existing tags, when the input file is (Ogg) FLAC
- not just those tags listed in the option.
Use the metaflac utility to tag FLAC files.
.TP
\f[V]flac --keep-foreign-metadata-if-present abc.wav\f[R]
FLAC files can store non-audio chunks of input WAVE/AIFF/RF64/W64 files.
The related option --keep-foreign-metadata works the same way, but will
instead exit with an error if the input has no such non-audio chunks.
The encoder only stores the chunks as they are, it cannot import the
content into its own tags (vorbis comments).
To transfer such tags from a source file, use tagging software which
supports them.
.TP
\f[V]flac -Vj2 -m3fo Track07.flac  -- -7.wav\f[R]
flac employs the commonplace convention that options in a short version
- invoked with single dash - can be shortened together until one that
takes an argument.
Here -j and -o do, and after the \[lq]2\[rq] a whitespace is needed to
start new options with single/double dash.
The -m option does not, and the following \[lq]3\[rq] is the -3
compression setting.
The options could equally well have been written out as -V -j 2 -m -3 -f
-o Track04.flac , or as -fo Track04.flac -3mVj2.
flac also employs the convention that \f[V]--\f[R] (with whitespace!)
signifies end of options, treating everything to follow as filename.
That is needed when an input filenames could otherwise be read as an
option, and \[lq]-7\[rq] is one such.
In total, this line takes the input file -7.wav as input; -o will give
output filename as Track07.flac, and the -f will overwrite if the file
Track04.flac is already present.
The encoder will select encoding preset -3 modified with the -m switch,
and use two CPU threads.
Afterwards, the -V will make it decode the flac file and compare the
audio to the input, to ensure they are indeed equal.
.SS Decoding examples
.TP
\f[V]flac --decode abc.flac\f[R] or \f[V]flac -d abc.flac\f[R]
Decode abc.flac to abc.wav.
abc.flac is not deleted.
If abc.wav is already present, the process will exit with an error
instead of overwriting; use \[en]force / -f to force overwrite.
NOTE: A mere flac abc.flac \f[I]without \[en]decode or its shortform
-d\f[R], would mean to re-encode abc.flac to abc.flac (see above), and
that command would err out because abc.flac already exists.
.TP
\f[V]flac -d --force-aiff-format abc.flac\f[R] or \f[V]flac -d -o abc.aiff abc.flac\f[R]
Two different ways of decoding abc.flac to abc.aiff (AIFF format).
abc.flac is not deleted.
-d -o could be shortened to -do.
The decoder can force other output formats, or different versions of the
WAVE/AIFF formats, see the options below.
.TP
\f[V]flac -d --keep-foreign-metadata-if-present abc.flac\f[R]
If the FLAC file has non-audio chunks stored from the original input
file, this option will restore both audio and non-audio.
The chunks will reveal the original file type, and the decoder will
select output format and output file extension accordingly - note that
this is not compatible with forcing a particular output format except if
it coincides with the original, as the decoder cannot transcode
non-audio between formats.
If there are no such chunks stored, it will decode to abc.wav.
The related option --keep-foreign-metadata will instead exit with an
error if no such non-audio chunks are found.
.TP
\f[V]flac -d -F abc.flac\f[R]
Decode abc.flac to abc.wav and don\[cq]t abort if errors are found.
This is potentially useful for recovering as much as possible from a
corrupted file.
Note: Be careful about trying to \[lq]repair\[rq] files this way.
Often it will only conceal an error, and not play any subjectively
\[lq]better\[rq] than the corrupted file.
It is a good idea to at least keep it, and possibly try several
decoders, including the one that generated the file, and hear if one has
less detrimental audible errors than another.
Make sure output volume is limited, as corrupted audio can generate loud
noises.
.SH OPTIONS
.PP
A summary of options is included below.
Several of the options can be negated, see the \f[B]Negative
options\f[R] section below.
.SS GENERAL OPTIONS
.TP
\f[B]-v\f[R], \f[B]--version\f[R]
Show the flac version number, and quit.
.TP
\f[B]-h\f[R], \f[B]--help\f[R]
Show basic usage and a list of all options, and quit.
.TP
\f[B]-d\f[R], \f[B]--decode\f[R]
Decode (the default behavior is to encode)
.TP
\f[B]-t\f[R], \f[B]--test\f[R]
Test a flac encoded file.
This works the same as -d except no decoded file is written, and with
some additional checks like parsing of all metadata blocks.
.TP
\f[B]-a\f[R], \f[B]--analyze\f[R]
Analyze a FLAC encoded file.
This works the same as -d except the output is an analysis file, not a
decoded file.
.TP
\f[B]-c\f[R], \f[B]--stdout\f[R]
Write output to stdout
.TP
\f[B]-f\f[R], \f[B]--force\f[R]
Force overwriting of output files.
By default, flac warns that the output file already exists and continues
to the next file.
.TP
\f[B]--delete-input-file\f[R]
Automatically delete the input file after a successful encode or decode.
If there was an error (including a verify error) the input file is left
intact.
.TP
\f[B]-o\f[R] \f[I]FILENAME\f[R], \f[B]--output-name\f[R]=\f[I]FILENAME\f[R]
Force the output file name (usually flac just changes the extension).
May only be used when encoding a single file.
May not be used in conjunction with --output-prefix.
.TP
\f[B]--output-prefix\f[R]=\f[I]STRING\f[R]
Prefix each output file name with the given string.
This can be useful for encoding or decoding files to a different
directory.
Make sure if your string is a path name that it ends with a trailing
\[ga]/\[cq] (slash).
.TP
\f[B]--preserve-modtime\f[R]
(Enabled by default.)
Output files have their timestamps/permissions set to match those of
their inputs.
Use --no-preserve-modtime to make output files have the current time and
default permissions.
.TP
\f[B]--keep-foreign-metadata\f[R]
If encoding, save WAVE, RF64, or AIFF non-audio chunks in FLAC metadata.
If decoding, restore any saved non-audio chunks from FLAC metadata when
writing the decoded file.
Foreign metadata cannot be transcoded, e.g.\ WAVE chunks saved in a FLAC
file cannot be restored when decoding to AIFF.
Input and output must be regular files (not stdin or stdout).
With this option, FLAC will pick the right output format on decoding.
It will exit with error if no such chunks are found.
.TP
\f[B]--keep-foreign-metadata-if-present\f[R]
Like --keep-foreign-metadata, but without throwing an error if foreign
metadata cannot be found or restored.
Instead, prints a warning.
.TP
\f[B]--skip\f[R]={#|\f[I]MM:SS\f[R]}
Skip the first number of samples of the input.
To skip over a given initial time, specify instead minutes and seconds:
there must then be at least one digit on each side of the colon sign.
Fractions of a second can be specified, with locale-dependent decimal
point, e.g.
--skip=123:9,867 if your decimal point is a comma.
A --skip option is applied to each input file if more are given.
This option cannot be used with -t.
When used with -a, the analysis file will enumerate frames from starting
point.
.TP
\f[B]--until\f[R]={#|[+|]\f[I]MM:SS\f[R]}
Stop at the given sample number (which is not included).
A negative number is taken relative to the end of the audio, a
\[ga]+\[cq] (plus) sign means that the --until point is taken relative
to the --skip point.
For other considerations, see --skip.
.TP
\f[B]--no-utf8-convert\f[R]
Do not convert tags from local charset to UTF-8.
This is useful for scripts, and setting tags in situations where the
locale is wrong.
This option must appear before any tag options!
.TP
\f[B]-s\f[R], \f[B]--silent\f[R]
Silent mode (do not write runtime encode/decode statistics to stderr)
.TP
\f[B]--totally-silent\f[R]
Do not print anything of any kind, including warnings or errors.
The exit code will be the only way to determine successful completion.
.TP
\f[B]-w\f[R], \f[B]--warnings-as-errors\f[R]
Treat all warnings as errors (which cause flac to terminate with a
non-zero exit code).
.SS DECODING OPTIONS
.TP
\f[B]-F\f[R], \f[B]--decode-through-errors\f[R]
By default flac stops decoding with an error message and removes the
partially decoded file if it encounters a bitstream error.
With -F, errors are still printed but flac will continue decoding to
completion.
Note that errors may cause the decoded audio to be missing some samples
or have silent sections.
.TP
\f[B]--cue\f[R]=[#.#][-[#.#]]
Set the beginning and ending cuepoints to decode.
Decimal points are locale-dependent (dot or comma).
The optional first #.# is the track and index point at which decoding
will start; the default is the beginning of the stream.
The optional second #.# is the track and index point at which decoding
will end; the default is the end of the stream.
If the cuepoint does not exist, the closest one before it (for the start
point) or after it (for the end point) will be used.
If those don\[cq]t exist , the start of the stream (for the start point)
or end of the stream (for the end point) will be used.
The cuepoints are merely translated into sample numbers then used as
--skip and --until.
A CD track can always be cued by, for example, --cue=9.1-10.1 for track
9, even if the CD has no 10th track.
.TP
\f[B]\[en]decode-chained-stream\f[R]
Decode all links in a chained Ogg stream, not just the first one.
.TP
\f[B]--apply-replaygain-which-is-not-lossless\f[R][=\f[I]SPECIFICATION\f[R]]
Applies ReplayGain values while decoding.
\f[B]WARNING: THIS IS NOT LOSSLESS.
DECODED AUDIO WILL NOT BE IDENTICAL TO THE ORIGINAL WITH THIS
OPTION.\f[R] This option is useful for example in transcoding media
servers, where the client does not support ReplayGain.
For details on the use of this option, see the section \f[B]ReplayGain
application specification\f[R].
.SS ENCODING OPTIONS
.PP
Encoding will default to -5, -A \[lq]tukey(5e-1)\[rq] and one CPU
thread.
.TP
\f[B]-V\f[R], \f[B]--verify\f[R]
Verify a correct encoding by decoding the output in parallel and
comparing to the original.
.TP
\f[B]-0\f[R], \f[B]--compression-level-0\f[R], \f[B]--fast\f[R]
Fastest compression preset.
Currently synonymous with \f[V]-l 0 -b 1152 -r 3 --no-mid-side\f[R]
.TP
\f[B]-1\f[R], \f[B]--compression-level-1\f[R]
Currently synonymous with \f[V]-l 0 -b 1152 -M -r 3\f[R]
.TP
\f[B]-2\f[R], \f[B]--compression-level-2\f[R]
Currently synonymous with \f[V]-l 0 -b 1152 -m -r 3\f[R]
.TP
\f[B]-3\f[R], \f[B]--compression-level-3\f[R]
Currently synonymous with \f[V]-l 6 -b 4096 -r 4 --no-mid-side\f[R]
.TP
\f[B]-4\f[R], \f[B]--compression-level-4\f[R]
Currently synonymous with \f[V]-l 8 -b 4096 -M -r 4\f[R]
.TP
\f[B]-5\f[R], \f[B]--compression-level-5\f[R]
Currently synonymous with \f[V]-l 8 -b 4096 -m -r 5\f[R]
.TP
\f[B]-6\f[R], \f[B]--compression-level-6\f[R]
Currently synonymous with
\f[V]-l 8 -b 4096 -m -r 6 -A \[dq]subdivide_tukey(2)\[dq]\f[R]
.TP
\f[B]-7\f[R], \f[B]--compression-level-7\f[R]
Currently synonymous with
\f[V]-l 12 -b 4096 -m -r 6 -A \[dq]subdivide_tukey(2)\[dq]\f[R]
.TP
\f[B]-8\f[R], \f[B]--compression-level-8\f[R], \f[B]--best\f[R]
Currently synonymous with
\f[V]-l 12 -b 4096 -m -r 6 -A \[dq]subdivide_tukey(3)\[dq]\f[R]
.TP
\f[B]-l\f[R] #, \f[B]--max-lpc-order\f[R]=#
Specifies the maximum LPC order.
This number must be <= 32.
For subset streams, it must be <=12 if the sample rate is <=48kHz.
If 0, the encoder will not attempt generic linear prediction, and only
choose among a set of fixed (hard-coded) predictors.
Restricting to fixed predictors only is faster, but compresses weaker -
typically five percentage points / ten percent larger files.
.TP
\f[B]-b\f[R] #, \f[B]--blocksize\f[R]=#
Specify the blocksize in samples.
The current default is 1152 for -l 0, else 4096.
Blocksize must be between 16 and 65535 (inclusive).
For subset streams it must be <= 4608 if the samplerate is <= 48kHz, for
subset streams with higher samplerates it must be <= 16384.
.TP
\f[B]-m\f[R], \f[B]--mid-side\f[R]
Try mid-side coding for each frame (stereo only, otherwise ignored).
.TP
\f[B]-M\f[R], \f[B]--adaptive-mid-side\f[R]
Adaptive mid-side coding for all frames (stereo only, otherwise
ignored).
.TP
\f[B]-r\f[R] [#,]#, \f[B]--rice-partition-order\f[R]=[#,]#
Set the [min,]max residual partition order (0..15).
For subset streams, max must be <=8.
min defaults to 0.
Default is -r 5.
Actual partitioning will be restricted by block size and prediction
order, and the encoder will silently reduce too high values.
.TP
\f[B]-A\f[R] \f[I]FUNCTION(S)\f[R], \f[B]--apodization\f[R]=\f[I]FUNCTION(S)\f[R]
Window audio data with given apodization function.
More can be given, comma-separated.
See section \f[B]Apodization functions\f[R] for details.
.TP
\f[B]-e\f[R], \f[B]--exhaustive-model-search\f[R]
Do exhaustive model search (expensive!).
.TP
\f[B]-q\f[R] #, \f[B]--qlp-coeff-precision\f[R]=#
Precision of the quantized linear-predictor coefficients.
This number must be in between 5 and 16, or 0 (the default) to let
encoder decide.
Does nothing if using -l 0.
.TP
\f[B]-p\f[R], \f[B]--qlp-coeff-precision-search\f[R]
Do exhaustive search of LP coefficient quantization (expensive!).
Overrides -q; does nothing if using -l 0.
.TP
\f[B]--lax\f[R]
Allow encoder to generate non-Subset files.
The resulting FLAC file may not be streamable or might have trouble
being played in all players (especially hardware devices), so you should
only use this option in combination with custom encoding options meant
for archival.
.TP
\f[B]--limit-min-bitrate\f[R]
Limit minimum bitrate by not allowing frames consisting of only constant
subframes.
This ensures a bitrate of at least 1 bit/sample, for example 48kbit/s
for 48kHz input.
This is mainly useful for internet streaming.
.TP
\f[B]-j\f[R] #, \f[B]--threads\f[R]=#
Try to set a maximum number of threads to use for encoding.
If multithreading was not enabled on compilation or when setting a
number of threads that is too high, this fails with a warning.
The value of 0 means a default set by the encoder; currently that is 1
thread (i.e.\ no multithreading), but that could change in the future.
Currently, up to 128 threads are supported.
Using a value higher than the number of available CPU threads harms
performance.
.TP
\f[B]--ignore-chunk-sizes\f[R]
When encoding to flac, ignore the file size headers in WAV and AIFF
files to attempt to work around problems with over-sized or malformed
files.
WAV and AIFF files both specifies length of audio data with an unsigned
32-bit number, limiting audio to just over 4 gigabytes.
Files larger than this are malformed, but should be read correctly using
this option.
Beware however, it could misinterpret any data following the audio
chunk, as audio.
.TP
\f[B]--replay-gain\f[R]
Calculate ReplayGain values and store them as FLAC tags, similar to
vorbisgain.
Title gains/peaks will be computed for each input file, and an album
gain/peak will be computed for all files.
All input files must have the same resolution, sample rate, and number
of channels.
Only mono and stereo files are allowed, and the sample rate must be 8,
11.025, 12, 16, 18.9, 22.05, 24, 28, 32, 36, 37.8, 44.1, 48, 56, 64, 72,
75.6, 88.2, 96, 112, 128, 144, 151.2, 176.4, 192, 224, 256, 288, 302.4,
352.8, 384, 448, 512, 576, or 604.8 kHz.
Also note that this option may leave a few extra bytes in a PADDING
block as the exact size of the tags is not known until all files are
processed.
Note that this option cannot be used when encoding to standard output
(stdout).
.TP
\f[B]--cuesheet\f[R]=\f[I]FILENAME\f[R]
Import the given cuesheet file and store it in a CUESHEET metadata
block.
This option may only be used when encoding a single file.
A seekpoint will be added for each index point in the cuesheet to the
SEEKTABLE unless --no-cued-seekpoints is specified.
.TP
\f[B]--picture\f[R]={\f[I]FILENAME|SPECIFICATION\f[R]}
Import a picture and store it in a PICTURE metadata block.
More than one --picture option can be specified.
Either a filename for the picture file or a more complete specification
form can be used.
The \f[I]SPECIFICATION\f[R] is a string whose parts are separated by |
(pipe) characters.
Some parts may be left empty to invoke default values.
Specifying only \f[I]FILENAME\f[R] is just shorthand for
\[lq]||||FILENAME\[rq].
See the section \f[B]Picture specification\f[R] for
\f[I]SPECIFICATION\f[R] format.
.TP
\f[B]-S\f[R] {#|X|#x|#s}, \f[B]--seekpoint\f[R]={#|X|#x|#s}
Specifies point(s) to include in SEEKTABLE, to override the
encoder\[cq]s default choice of one per ten seconds (`-s 10s').
Using #, a seek point at that sample number is added.
Using X, a placeholder point is added at the end of a the table.
Using #x, # evenly spaced seek points will be added, the first being at
sample 0.
Using #s, a seekpoint will be added every # seconds, where decimal
points are locale-dependent, e.g.\  `-s 9.5s' or `-s 9,5s'.
Several -S options may be given; the resulting SEEKTABLE will contain
all seekpoints specified (duplicates removed).
Note: `-S #x' and `-S #s' will not work if the encoder cannot determine
the input size before starting.
Note: if you use `-S #' with # being >= the number of samples in the
input, there will be either no seek point entered (if the input size is
determinable before encoding starts) or a placeholder point (if input
size is not determinable).
Use --no-seektable for no SEEKTABLE.
.TP
\f[B]-P\f[R] #, \f[B]--padding\f[R]=#
(Default: 8192 bytes, although 65536 for input above 20 minutes.)
Tell the encoder to write a PADDING metadata block of the given length
(in bytes) after the STREAMINFO block.
This is useful for later tagging, where one can write over the PADDING
block instead of having to rewrite the entire file.
Note that a block header of 4 bytes will come on top of the length
specified.
.TP
\f[B]-T\f[R] \[lq]\f[I]FIELD=VALUE\f[R]\[rq]\f[B], --tag\f[R]=\[lq]\f[I]FIELD=VALUE\f[R]\[rq]
Add a FLAC tag.
The comment must adhere to the Vorbis comment spec; i.e.\ the FIELD must
contain only legal characters, terminated by an `equals' sign.
Make sure to quote the content if necessary.
This option may appear more than once to add several Vorbis comments.
NOTE: all tags will be added to all encoded files.
.TP
\f[B]--tag-from-file\f[R]=\[lq]\f[I]FIELD=FILENAME\f[R]\[rq]
Like --tag, except FILENAME is a file whose contents will be read
verbatim to set the tag value.
The contents will be converted to UTF-8 from the local charset.
This can be used to store a cuesheet in a tag
(e.g.\ --tag-from-file=\[lq]CUESHEET=image.cue\[rq]).
Do not try to store binary data in tag fields!
Use APPLICATION blocks for that.
.SS FORMAT OPTIONS
.PP
Encoding defaults to FLAC and not OGG.
Decoding defaults to WAVE (more specifically WAVE_FORMAT_PCM for
mono/stereo with 8/16 bits, and to WAVE_FORMAT_EXTENSIBLE otherwise),
except: will be overridden by chunks found by
--keep-foreign-metadata-if-present or --keep-foreign-metadata
.TP
\f[B]--ogg\f[R]
When encoding, generate Ogg FLAC output instead of native FLAC.
Ogg FLAC streams are FLAC streams wrapped in an Ogg transport layer.
The resulting file should have an `.oga' extension and will still be
decodable by flac.
When decoding, force the input to be treated as Ogg FLAC.
This is useful when piping input from stdin or when the filename does
not end in `.oga' or `.ogg'.
.TP
\f[B]--serial-number\f[R]=#
When used with --ogg, specifies the serial number to use for the first
Ogg FLAC stream, which is then incremented for each additional stream.
When encoding and no serial number is given, flac uses a random number
for the first stream, then increments it for each additional stream.
When decoding and no number is given, flac uses the serial number of the
first page.
.PP
\f[B]--force-aiff-format\f[R]
.PD 0
.P
.PD
\f[B]--force-rf64-format\f[R]
.PD 0
.P
.PD
\f[B]--force-wave64-format\f[R] : For decoding: Override default output
format and force output to AIFF/RF64/WAVE64, respectively.
This option is not needed if the output filename (as set by -o) ends
with \f[I].aif\f[R] or \f[I].aiff\f[R], \f[I].rf64\f[R] and
\f[I].w64\f[R] respectively.
The encoder auto-detects format and ignores this option.
.PP
\f[B]--force-legacy-wave-format\f[R]
.PD 0
.P
.PD
\f[B]--force-extensible-wave-format\f[R] : Instruct the decoder to
output a WAVE file with WAVE_FORMAT_PCM and WAVE_FORMAT_EXTENSIBLE
respectively, overriding default choice.
.PP
\f[B]--force-aiff-c-none-format\f[R]
.PD 0
.P
.PD
\f[B]--force-aiff-c-sowt-format\f[R] : Instruct the decoder to output an
AIFF-C file with format NONE and sowt respectively.
.TP
\f[B]--force-raw-format\f[R]
Force input (when encoding) or output (when decoding) to be treated as
raw samples (even if filename suggests otherwise).
.SS raw format options
.PP
When encoding from or decoding to raw PCM, format must be specified.
.TP
\f[B]--sign\f[R]={signed|unsigned}
Specify the sign of samples.
.TP
\f[B]--endian\f[R]={big|little}
Specify the byte order for samples
.TP
\f[B]--channels\f[R]=#
(Input only) specify number of channels.
The channels must be interleaved, and in the order of the FLAC format
(see the format specification); the encoder (/decoder) cannot re-order
channels.
.TP
\f[B]--bps\f[R]=#
(Input only) specify bits per sample (per channel: 16 for CDDA.)
.TP
\f[B]--sample-rate\f[R]=#
(Input only) specify sample rate (in Hz.
Only integers supported.)
.TP
\f[B]--input-size\f[R]=#
(Input only) specify the size of the raw input in bytes.
This option is only compulsory when encoding from stdin and using
options that need to know the input size beforehand (like, --skip,
--until, --cuesheet ) The encoder will truncate at the specified size if
the input stream is bigger.
If the input stream is smaller, it will complain about an unexpected
end-of-file.
.SS ANALYSIS OPTIONS
.TP
\f[B]--residual-text\f[R]
Includes the residual signal in the analysis file.
This will make the file very big, much larger than even the decoded
file.
.TP
\f[B]--residual-gnuplot\f[R]
Generates a gnuplot file for every subframe; each file will contain the
residual distribution of the subframe.
This will create a lot of files.
gnuplot must be installed separately.
.SS NEGATIVE OPTIONS
.PP
The following will negate an option previously given:
.PP
\f[B]--no-adaptive-mid-side\f[R]
.PD 0
.P
.PD
\f[B]--no-cued-seekpoints\f[R]
.PD 0
.P
.PD
\f[B]--no-decode-through-errors\f[R]
.PD 0
.P
.PD
\f[B]--no-delete-input-file\f[R]
.PD 0
.P
.PD
\f[B]--no-preserve-modtime\f[R]
.PD 0
.P
.PD
\f[B]--no-keep-foreign-metadata\f[R]
.PD 0
.P
.PD
\f[B]--no-exhaustive-model-search\f[R]
.PD 0
.P
.PD
\f[B]--no-force\f[R]
.PD 0
.P
.PD
\f[B]--no-lax\f[R]
.PD 0
.P
.PD
\f[B]--no-mid-side\f[R]
.PD 0
.P
.PD
\f[B]--no-ogg\f[R]
.PD 0
.P
.PD
\f[B]--no-padding\f[R]
.PD 0
.P
.PD
\f[B]--no-qlp-coeff-prec-search\f[R]
.PD 0
.P
.PD
\f[B]--no-replay-gain\f[R]
.PD 0
.P
.PD
\f[B]--no-residual-gnuplot\f[R]
.PD 0
.P
.PD
\f[B]--no-residual-text\f[R]
.PD 0
.P
.PD
\f[B]--no-seektable\f[R]
.PD 0
.P
.PD
\f[B]--no-silent\f[R]
.PD 0
.P
.PD
\f[B]--no-verify\f[R]
.PD 0
.P
.PD
\f[B]--no-warnings-as-errors\f[R]
.SS ReplayGain application specification
.PP
The option --apply-replaygain-which-is-not-lossless[=<specification>]
applies ReplayGain values while decoding.
\f[B]WARNING: THIS IS NOT LOSSLESS.
DECODED AUDIO WILL NOT BE IDENTICAL TO THE ORIGINAL WITH THIS
OPTION.\f[R] This option is useful for example in transcoding media
servers, where the client does not support ReplayGain.
.PP
The <specification> is a shorthand notation for describing how to apply
ReplayGain.
All elements are optional - defaulting to 0aLn1 - but order is
important.
The format is:
.PP
[<preamp>][a|t][l|L][n{0|1|2|3}]
.PP
In which the following parameters are used:
.IP \[bu] 2
\f[B]preamp\f[R]: A floating point number in dB.
This is added to the existing gain value.
.IP \[bu] 2
\f[B]a|t\f[R]: Specify `a' to use the album gain, or `t' to use the
track gain.
If tags for the preferred kind (album/track) do not exist but tags for
the other (track/album) do, those will be used instead.
.IP \[bu] 2
\f[B]l|L\f[R]: 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.
.IP \[bu] 2
\f[B]n{0|1|2|3}\f[R]: 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\[cq]t hear it as
much.
0 means no noise shaping, 1 means `low', 2 means `medium', 3 means
`high'.
.PP
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.
.PP
flac uses the ReplayGain tags for the calculation.
If a stream does not have the required tags or they can\[cq]t be parsed,
decoding will continue with a warning, and no ReplayGain is applied to
that stream.
.SS Picture specification
.PP
This described the specification used for the \f[B]--picture\f[R]
option.
[\f[I]TYPE\f[R]]|[\f[I]MIME-TYPE\f[R]]|[\f[I]DESCRIPTION\f[R]]|[\f[I]WIDTHxHEIGHTxDEPTH\f[R][/\f[I]COLORS\f[R]]]|\f[I]FILE\f[R]
.PP
\f[I]TYPE\f[R] is optional; it is a number from one of:
.IP " 0." 4
Other
.IP " 1." 4
32x32 pixels `file icon' (PNG only)
.IP " 2." 4
Other file icon
.IP " 3." 4
Cover (front)
.IP " 4." 4
Cover (back)
.IP " 5." 4
Leaflet page
.IP " 6." 4
Media (e.g.\ label side of CD)
.IP " 7." 4
Lead artist/lead performer/soloist
.IP " 8." 4
Artist/performer
.IP " 9." 4
Conductor
.IP "10." 4
Band/Orchestra
.IP "11." 4
Composer
.IP "12." 4
Lyricist/text writer
.IP "13." 4
Recording Location
.IP "14." 4
During recording
.IP "15." 4
During performance
.IP "16." 4
Movie/video screen capture
.IP "17." 4
A bright coloured fish
.IP "18." 4
Illustration
.IP "19." 4
Band/artist logotype
.IP "20." 4
Publisher/Studio logotype
.PP
The default is 3 (front cover).
There may only be one picture each of type 1 and 2 in a file.
.PP
\f[I]MIME-TYPE\f[R] is optional; if left blank, it will be detected from
the file.
For best compatibility with players, use pictures with MIME type
image/jpeg or image/png.
The MIME type can also be --> to mean that FILE is actually a URL to an
image, though this use is discouraged.
.PP
\f[I]DESCRIPTION\f[R] is optional; the default is an empty string.
.PP
The next part specifies the resolution and color information.
If the \f[I]MIME-TYPE\f[R] is image/jpeg, image/png, or image/gif, you
can usually leave this empty and they can be detected from the file.
Otherwise, you must specify the width in pixels, height in pixels, and
color depth in bits-per-pixel.
If the image has indexed colors you should also specify the number of
colors used.
When manually specified, it is not checked against the file for
accuracy.
.PP
\f[I]FILE\f[R] is the path to the picture file to be imported, or the
URL if MIME type is -->
.PP
\f[B]Specification examples:\f[R] \[lq]|image/jpeg|||../cover.jpg\[rq]
will embed the JPEG file at ../cover.jpg, defaulting to type 3 (front
cover) and an empty description.
The resolution and color info will be retrieved from the file itself.
\[lq]4|-->|CD|320x300x24/173|http://blah.blah/backcover.tiff\[rq] will
embed the given URL, with type 4 (back cover), description \[lq]CD\[rq],
and a manually specified resolution of 320x300, 24 bits-per-pixel, and
173 colors.
The file at the URL will not be fetched; the URL itself is stored in the
PICTURE metadata block.
.SS Apodization functions
.PP
To improve LPC analysis, the audio data is windowed.
An \f[B]-A\f[R] option applies the specified apodization function(s)
instead of the default (which is \[lq]tukey(5e-1)\[rq], though different
for presets -6 to -8.)
Specifying one more function effectively means, for each subframe, to
try another weighting of the data and see if it happens to result in a
smaller encoded subframe.
Specifying several functions is time-expensive, at typically diminishing
compression gains.
.PP
The subdivide_tukey(\f[I]N\f[R]) functions (see below) used in presets
-6 to -8 were developed to recycle calculations for speed, compared to
using a number of independent functions.
Even then, a high number like \f[I]N\f[R]>4 or 5, will often become less
efficient than other options considered expensive, like the slower -p,
though results vary with signal.
.PP
Up to 32 functions can be given as comma-separated list and/or
individual \f[B]-A\f[R] options.
Any mis-specified function is silently ignored.
Quoting a function which takes options (and has parentheses) may be
necessary, depending on shell.
Currently the following functions are implemented: bartlett,
bartlett_hann, blackman, blackman_harris_4term_92db, connes, flattop,
gauss(\f[I]STDDEV\f[R]), hamming, hann, kaiser_bessel, nuttall,
rectangle, triangle, tukey(\f[I]P\f[R]),
partial_tukey(\f[I]N\f[R][/\f[I]OV\f[R][/\f[I]P\f[R]]]),
punchout_tukey(\f[I]N\f[R][/\f[I]OV\f[R][/\f[I]P\f[R]]]),
subdivide_tukey(\f[I]N\f[R][/\f[I]P\f[R]]), welch.
.PP
For parameters \f[I]P\f[R], \f[I]STDDEV\f[R] and \f[I]OV\f[R],
scientific notation is supported, e.g.\  tukey(5e-1).
Otherwise, the decimal point must agree with the locale,
e.g.\ tukey(0.5) or tukey(0,5) depending on your system.
.IP \[bu] 2
For gauss(\f[I]STDDEV\f[R]), \f[I]STDDEV\f[R] is the standard deviation
(0<\f[I]STDDEV\f[R]<=5e-1).
.IP \[bu] 2
For tukey(\f[I]P\f[R]), \f[I]P\f[R] (between 0 and 1) specifies the
fraction of the window that is cosine-tapered; \f[I]P\f[R]=0 corresponds
to \[lq]rectangle\[rq] and \f[I]P\f[R]=1 to \[lq]hann\[rq].
.IP \[bu] 2
partial_tukey(\f[I]N\f[R]) and punchout_tukey(\f[I]N\f[R]) are largely
obsoleted by the more time-effective subdivide_tukey(\f[I]N\f[R]), see
next item.
They generate \f[I]N\f[R] functions each spanning a part of each block.
Optional arguments are an overlap \f[I]OV\f[R] (<1, may be negative),
for example partial_tukey(2/2e-1); and then a taper parameter
\f[I]P\f[R], for example partial_tukey(2/2e-1/5e-1).
.IP \[bu] 2
subdivide_tukey(\f[I]N\f[R]) is a more efficient reimplementation of
partial_tukey and punchout_tukey taken together, combining the windows
they would generate up to the specified \f[I]N\f[R].
Specifying subdivide_tukey(3) entails a tukey, a partial_tukey(2), a
partial_tukey(3) and a punchout_tukey(3); specifying subdivide_tukey(5)
will on top of that add a partial_tukey(4), a punchout_tukey(4), a
partial_tukey(5) and a punchout_tukey(5) - but all with tapering chosen
to facilitate the re-use of computation.
Thus the \f[I]P\f[R] parameter (defaulting to 5e-1) is applied for the
smallest used window: For example, subdivide_tukey(2/5e-1) results in
the same taper as that of tukey(25e-2) and subdivide_tukey(5) in the
same taper as of tukey(1e-1).
.SH SEE ALSO
.PP
\f[B]metaflac(1)\f[R]
.SH AUTHOR
.PP
This manual page was initially written by Matt Zimmerman
<mdz\[at]debian.org> for the Debian GNU/Linux system (but may be used by
others).
It has been kept up-to-date by the Xiph.org Foundation.