bump version

It turned out that mode 1 is explicit, backward compatible signaling.
We don't have to implement it on our side.

.gitignore

@@ -0,0 +1,22 @@
+*.o
+*~
+Makefile
+Makefile.in
+aclocal.m4
+autom4te.cache
+compile
+config.guess
+config.h
+config.h.in
+config.log
+config.status
+config.sub
+configure
+depcomp
+install-sh
+missing
+missings/.deps/
+src/.deps/
+src/.dirstamp
+stamp-h1
+fdkaac

ChangeLog

@@ -1,7 +1,83 @@
-2013-10-25  nu774 <honeycomb77@gmail.com>
+2013-11-04  nu774 <honeycomb77@gmail.com>
 
   * update ChangeLog [HEAD]
 
+  * update git2changelog to accept non-ascii output
+
+  * add manpage
+
+2013-11-03  nu774 <honeycomb77@gmail.com>
+
+  * fix gcc warnings [origin/master]
+
+  * Merge pull request #7 from rbrito/misc-fixes
+
+2013-11-03  Rogério Brito <rbrito@ime.usp.br>
+
+  * gitignore: Add list of files to ignore.
+
+2013-11-03  nu774 <honeycomb77@gmail.com>
+
+  * update ChangeLog
+
+  * bump version [v0.5.0]
+
+  * add --sbr-ratio to support AACENC_SBR_RATIO appeared on libFDK 3.4.12
+
+  * support 7.1 channel mode added on FDK 3.4.12
+
+2013-10-30  nu774 <honeycomb77@gmail.com>
+
+  * update ChangeLog
+
+  * bump version [v0.4.2]
+
+  * use tell() to obtain data chunk offset
+
+  * rename aacenc_result_t -> aacenc_frame_t, simplify write_sample()
+
+  * prepend 1 sample zero padding in case of SBR and enc_delay is odd
+
+  * cleanup interface of aac_encode_frame()
+
+  * add some copyright notice
+
+2013-10-29  nu774 <honeycomb77@gmail.com>
+
+  * smart padding for better gapless playback
+
+  * fix unused variable warning
+
+  * fix warning: cast size_t as sprintf() arg to int
+
+  * fix vcxproj
+
+  * fix pcm_seek() to inline
+
+2013-10-27  nu774 <honeycomb77@gmail.com>
+
+  * bump version [v0.4.1]
+
+  * add --include-sbr-delay
+
+  * fix help message: show -I as shorthand for --ignorelength
+
+  * remove --sbr-signaling
+
+2013-10-26  nu774 <honeycomb77@gmail.com>
+
+  * re-fix #ifdef cond for lrint()
+
+  * tag mapping: add recorded date and tempo, remove performer->artist
+
+2013-10-25  nu774 <honeycomb77@gmail.com>
+
+  * fix MSVC12 build issue
+
+  * fix build issue on platform where fileno is a naive macro
+
+  * update ChangeLog
+
   * bump version [v0.4.0]
 
   * update README

Makefile.am

@@ -18,6 +18,8 @@ fdkaac_SOURCES = \
     src/progress.c             \
     src/wav_reader.c
 
+dist_man_MANS = man/fdkaac.1
+
 fdkaac_LDADD = \
     @LIBICONV@ -lfdk-aac -lm
 

README

@@ -1,122 +1,332 @@
-==========================================================================
-fdkaac - command line frontend encoder for libfdk-aac
-==========================================================================
+% FDKAAC(1)
+% nu774 <honeycomb77@gmail.com>
+% November, 2013
 
-Prerequisites
--------------
-You need libfdk-aac.
-On Posix environment, you will also need GNU gettext (for iconv.m4) and
-GNU autoconf/automake.
+NAME
+====
 
-How to build on Posix environment
----------------------------------
-First, you need to build libfdk-aac and install on your system.
-Once you have done it, the following will do the task.
-(MinGW build can be done the same way, and doesn't require gettext/iconv)
+fdkaac - command line frontend for libfdk-aac encoder  
 
-$ autoreconf -i
-$ ./configure && make && make install
+SYNOPSIS
+========
 
-How to build on MSVC
---------------------
-First you have to extract libfdk-aac source here, so that directory tree will
-look like the following:
-+- fdk-aac ---+-documentation
-|             +-libAACdec
-|             +-libAACenc
-|             :
-+- m4
-+- missings
-+- MSVC
-+- src
+**fdkaac** [OPTIONS] [FILE]
 
-MSVC solution for Visual Studio 2010 is under MSVC directory.
+DESCRIPTION
+===========
 
-Available input format
-----------------------
-WAV, RF64, CAF, RAW, upto 32bit int / 64bit float format is supported.
-Metadata in CAF info chunk can be read and copied to the resulting m4a.
-This is especially useful and works well when you pipe from ffmpeg via CAF.
-For example, you can copy tag from original "foo.flac" to "foo.m4a"
-through the following pipeline:
+**fdkaac** reads linear PCM audio in either WAV, raw PCM, or CAF format,
+and encodes it into either M4A / AAC file.
 
-$ ffmpeg -i foo.flac -f caf - | fdkaac -m3 - -o foo.m4a
+If the input file is "-", data is read from stdin. Likewise, if the
+output file is "-", data is written to stdout if one of streamable AAC
+transport formats are selected by **-f**.
 
-Since FDK AAC encoder is implemented based on fixed point integer,
-encoder itself handles 16bit input only.
-Therefore, when feeding non-integer input, be careful so that input doesn't
-exceed 0dBFS to avoid hard clips.
-You might also want to apply dither/noise shape beforehand when your input
-has higher resolution.
+When CAF input and M4A output is used, tags in CAF file are copied into
+the resulting M4A.  
 
-Note that fdkaac doesn't automatically resample for you
-when input samplerate is not supported by AAC spec.
+OPTIONS
+=======
 
-Tagging Options
----------------
-Generic tagging options like --tag, --tag-from-file, --long-tag allows you
-to set arbitrary tags.
-Available tags and their fcc (four char code) for --tag and --tag-from-file
-can be found at http://code.google.com/p/mp4v2/wiki/iTunesMetadata
+-h, --help
+:   Show command help
 
-For tags such as Artist where first char of fcc is copyright sign,
-you can skip first char and just say like --tag="ART:Foo Bar" or
---tag-from-file=lyr:/path/to/your/lyrics.txt
+-o \<FILE\>
+:   Output filename.
 
-Currently, --tag-from-file just stores file contents into m4a without any
-character encoding / line terminater conversion.
-Therefore, only use UTF-8 (without BOM) when setting text tags by this option.
+-p, --profile \<n\>
+:   Target profile (MPEG4 audio object type, AOT)
 
-On the other hand, --tag / --long-tag (and other command line arguments) are
-converted from locale character encoding to UTF-8 on Posix environment.
-On Windows, command line arguments are always treated as Unicode.
+    2
+    :   MPEG-4 AAC LC (default)
 
-Tagging using JSON
-------------------
-With --tag-from-json, fdkaac can read JSON file and set tags from it.
-By default, tags are assumed to be in the root object(dictionary) like this:
+    5
+    :   MPEG-4 HE-AAC (SBR)
 
-{
-  "title": "No Expectations",
-  "artist": "The Rolling Stones",
-  "album": "Beggars Banquet",
-  "track": 2
-}
+    29
+    :   MPEG-4 HE-AAC v2 (SBR+PS)
 
-In this case, you can simply specify the filename like:
---tag-from-json=/path/to/json
+    23
+    :   MPEG-4 AAC LD
 
-If the object containing tags is placed somewhere else, you can optionally
-specify the path of the object with dotted notation.
+    39
+    :   MPEG-4 AAC ELD
 
-{
-  "format" : {
-    "filename" : "Middle Curse.flac",
-    "nb_streams" : 1,
-    "format_name" : "flac",
-    "format_long_name" : "raw FLAC",
-    "start_time" : "N/A",
-    "duration" : "216.146667",
-    "size" : "11851007.000000",
-    "bit_rate" : "438628.000000",
-    "tags" : {
-      "ALBUM" : "Scary World Theory",
-      "ARTIST" : "Lali Puna",
-      "DATE" : "2001",
-      "DISCID" : "9208CC0A",
-      "TITLE" : "Middle Curse",
-      "TRACKTOTAL" : "10",
-      "track" : "2"
-    }
-  }
-}
+    129
+    :   MPEG-2 AAC LC
 
-In this example, tags are placed under the object "format.tags".
-("format" is a child of the root, and "tags" is a child of the "format").
-In this case, you can say:
---tag-from-json=/path/to/json?format.tags
+    132
+    :   MPEG-2 HE-AAC (SBR)
 
-For your information, ffprobe of ffmpeg project (or avprobe of libav) can
-output media information/metadata in json format like this.
+    156
+    :   MPEG-2 HE-AAC v2 (SBR+PS)
+
+-b, --bitrate \<n\>
+:   Target bitrate (for CBR)
+
+-m, --bitrate-mode \<n\>
+:   Bitrate configuration mode. Available VBR quality value depends on
+    other parameters such as profile, sample rate, or number of
+    channels.
+
+    0
+    :   CBR (default)
+
+    1-5
+    :   VBR (higher value -\> higher bitrate)
+
+-w, --bandwith \<n\>
+:   Frequency bandwith (lowpass cut-off frequency) in Hz. Available on
+    AAC LC only.
+
+-a, --afterburner \<n\>
+:   Configure afterburner mode. When enabled, quality is increased at
+    the expense of additional computational workload.
+
+    0
+    :   Off
+
+    1
+    :   On (default)
+
+-L, --lowdelay-sbr \<n\>
+:   Configure SBR activity on AAC ELD.
+
+    -1
+    :   Use ELD SBR auto configuration
+
+    0
+    :   Disable SBR on ELD (default)
+
+    1
+    :   Enable SBR on ELD
+
+-s, --sbr-ratio \<n\>
+:   Controls activation of downsampled SBR.
+
+    0
+    :   Use lib default (default)
+
+    1
+    :   Use downsampled SBR (default for ELD+SBR)
+
+    2
+    :   Use dual-rate SBR (default for HE-AAC)
+
+    Dual-rate SBR is what is normally used for HE-AAC, where AAC is
+    encoded at half the sample rate of SBR, hence "dual rate". On the
+    other hand, downsampled SBR uses same sample rate for both of AAC
+    and SBR (single rate), therefore downsampled SBR typically consumes
+    more bitrate.
+
+    Downsampled SBR is newly introduced feature in FDK encoder library
+    version 3.4.12. When libfdk-aac in the system doesn't support this,
+    dual-rate SBR will be used. When available, dual-rate SBR is the
+    default for HE-AAC and downsampled SBR is the default for ELD+SBR.
+
+    Note that downsampled HE-AAC is not so common as dual-rate one. When
+    downsampled HE-AAC is selected, **fdkaac** is forced to choose
+    explicit hierarchical SBR signaling, which (at least) iTunes doesn't
+    accept.
+
+-f, --transport-format \<n\>
+:   Transport format. Tagging and gapless playback is only available on
+    M4A. Streaming to stdout is only available on others.
+
+    0
+    :   M4A (default)
+
+    1
+    :   ADIF
+
+    2
+    :   ADTS
+
+    6
+    :   LATM MCP=1
+
+    7
+    :   LATM MCP=0
+
+    10
+    :   LOAS/LATM (LATM within LOAS)
+
+-C, --adts-crc-check
+:   Add CRC protection on ADTS header.
+
+-h, --header-period \<n\>
+:   StreamMuxConfig/PCE repetition period in the transport layer.
+
+-G, --gapless-mode \<n\>
+:   Method to declare amount of encoder delay (and padding) in M4A
+    container. These values are mandatory for proper gapless playback on
+    player side.
+
+    0
+    :   iTunSMPB (default)
+
+    1
+    :   ISO standard (edts and sgpd)
+
+    2
+    :   Both
+
+--include-sbr-delay
+:   When specified, count SBR decoder delay in encoder delay.
+
+    This is not iTunes compatible and will lead to gapless playback
+    issue on LC only decoder, but this is the default behavior of FDK
+    library.
+
+    Whether counting SBR decoder delay in encoder delay or not result in
+    incompatibility in gapless playback. You should pick which one will
+    work for your favorite player.
+
+    However, it's better not to choose SBR at all if you want gapless
+    playback. LC doesn't have such issues.
+
+-I, --ignorelength
+:   Ignore length field of data chunk in input WAV file.
+
+-S, --silent
+:   Don't print progress messages.
+
+--moov-before-mdat
+:   Place moov box before mdat box in M4A container. This option might
+    be important for some hardware players, that are known to refuse
+    moov box placed after mdat box.
+
+-R, --raw
+:   Regard input as raw PCM.
+
+--raw-channels \<n\>
+:   Specify number of channels of raw input (default: 2)
+
+--raw-rate \<n\>
+:   Specify sample rate of raw input (default: 44100)
+
+--raw-format \<spec\>
+:   Specify sample format of raw input (default: "S16L"). **Spec** is as
+    the following (case insensitive):
+
+    1st char -- type of sample
+    :   **S** (igned) | **U** (nsigned) | **F** (loat)
+
+    2nd part (in digits)
+    :   bits per channel
+
+    Last char -- endianness (can be omitted)
+    :   **L** (ittle, default) | **B** (ig)
+
+--title \<string\>
+:   Set title tag.
+
+--artist \<string\>
+:   Set artist tag.
+
+--album \<string\>
+:   Set album tag.
+
+--genre \<string\>
+:   Set genre tag.
+
+--date \<string\>
+:   Set date tag.
+
+--composer \<string\>
+:   Set composer tag.
+
+--grouping \<string\>
+:   Set grouping tag.
+
+--comment \<string\>
+:   Set comment tag.
+
+--album-artist \<string\>
+:   Set album artist tag.
+
+--track \<number[/total]\>
+:   Set track tag, with or without number of total tracks.
+
+--disk \<number[/total]\>
+:   Set disk tag, with or without number of total discs.
+
+--tempo \<n\>
+:   Set tempo (BPM) tag.
+
+--tag \<fcc\>:\<value\>
+:   Set iTunes predefined tag with explicit fourcc key and value. See
+    [https://code.google.com/p/mp4v2/wiki/iTunesMetadata](https://code.google.com/p/mp4v2/wiki/iTunesMetadata)
+    for known predefined keys. You can omit first char of **fcc** when
+    it is the copyright sign.
+
+--tag-from-file \<fcc\>:\<filename\>
+:   Same as --tag, but set content of file as tag value.
+
+--long-tag \<name\>:\<value\>
+:   Set arbitrary tag as iTunes custom metadata. Stored in
+    com.apple.iTunes field.
+
+--tag-from-json \<filename[?dot\_notation]\>
+:   Read tags from JSON. By default, tags are assumed to be direct
+    children of the root object in JSON. Optionally you can specify
+    arbitrary dot notation to locate the object containing tags.
+ 
+
+EXAMPLES
+========
+
+Encode WAV file into a M4A file. MPEG4 AAC LC, VBR quality 3:
+
+    fdkaac -m3 foo.wav
+
+Encode WAV file into a M4A file. MPEG4 HE-AAC, bitrate 64kbps:
+
+    fdkaac -p5 -b64 foo.wav
+
+Piping from **ffmpeg** (you need version supporting CAF output):
+
+    ffmpeg -i foo.flac -f caf - | fdkaac -b128 - -o foo.m4a
+
+Import tags via json:
+
+    ffprobe -v 0 -of json -show_format foo.flac >foo.json
+
+    flac -dc foo.flac | \
+    fdkaac - -ox.m4a -m2 --import-tag-from-json=foo.json?format.tags
+
+NOTES
+=====
+
+Upto 32bit integer or 64bit floating point format is supported as input.
+However, FDK library is implemented based on fixed point math and only
+supports 16bit integer PCM. Therefore, be wary of clipping. You might
+want to dither/noise shape beforehand when your input has higher
+resolution.
+
+Following channel layouts are supported by the encoder.
+
+1ch
+:   C
+
+2ch
+:   L R
+
+3ch
+:   C L R
+
+4ch
+:   C L R Cs
+
+5ch
+:   C L R Ls Rs
+
+5.1ch
+:   C L R Ls Rs LFE
+
+7.1ch (front)
+:   C Lc Rc L R Ls Rs LFE
+
+7.1ch (rear)
+:   C L R Ls Rs Rls Rrs LFE
 
-Note that not all tags can be read/written this way.

git2changelog.py

@@ -39,6 +39,6 @@ with Popen(GITLOG_CMD, shell=False, stdout=PIPE).stdout as pipe:
     commits = parse_gitlog(pipe)
     commits_by_date_author = groupby(commits, key=lambda x: (x.date, x.author))
     for (date, author), commits in commits_by_date_author:
-        output('{0}  {1}\n\n'.format(date, author))
+        output(u'{0}  {1}\n\n'.format(date, author).encode('utf-8'))
         for c in commits:
-            output('  * {0}{1}\n\n'.format(c.subject, c.ref))
+            output(u'  * {0}{1}\n\n'.format(c.subject, c.ref).encode('utf-8'))

man/fdkaac.1

@@ -0,0 +1,497 @@
+.TH FDKAAC 1 "November, 2013" 
+.SH NAME
+.PP
+fdkaac - command line frontend for libfdk-aac encoder \ 
+.SH SYNOPSIS
+.PP
+\f[B]fdkaac\f[] [OPTIONS] [FILE]
+.SH DESCRIPTION
+.PP
+\f[B]fdkaac\f[] reads linear PCM audio in either WAV, raw PCM, or CAF
+format, and encodes it into either M4A / AAC file.
+.PP
+If the input file is "-", data is read from stdin.
+Likewise, if the output file is "-", data is written to stdout if one of
+streamable AAC transport formats is selected by \f[B]-f\f[].
+.PP
+When CAF input and M4A output is used, tags in CAF file are copied into
+the resulting M4A.
+\ 
+.SH OPTIONS
+.TP
+.B -h, --help
+Show command help
+.RS
+.RE
+.TP
+.B -o <FILE>
+Output filename.
+.RS
+.RE
+.TP
+.B -p, --profile <n>
+Target profile (MPEG4 audio object type, AOT)
+.RS
+.TP
+.B 2
+MPEG-4 AAC LC (default)
+.RS
+.RE
+.TP
+.B 5
+MPEG-4 HE-AAC (SBR)
+.RS
+.RE
+.TP
+.B 29
+MPEG-4 HE-AAC v2 (SBR+PS)
+.RS
+.RE
+.TP
+.B 23
+MPEG-4 AAC LD
+.RS
+.RE
+.TP
+.B 39
+MPEG-4 AAC ELD
+.RS
+.RE
+.TP
+.B 129
+MPEG-2 AAC LC
+.RS
+.RE
+.TP
+.B 132
+MPEG-2 HE-AAC (SBR)
+.RS
+.RE
+.TP
+.B 156
+MPEG-2 HE-AAC v2 (SBR+PS)
+.RS
+.RE
+.RE
+.TP
+.B -b, --bitrate <n>
+Target bitrate (for CBR)
+.RS
+.RE
+.TP
+.B -m, --bitrate-mode <n>
+Bitrate configuration mode.
+Available VBR quality value depends on other parameters such as profile,
+sample rate, or number of channels.
+.RS
+.TP
+.B 0
+CBR (default)
+.RS
+.RE
+.TP
+.B 1-5
+VBR (higher value -> higher bitrate)
+.RS
+.RE
+.RE
+.TP
+.B -w, --bandwith <n>
+Frequency bandwith (lowpass cut-off frequency) in Hz.
+Available on AAC LC only.
+.RS
+.RE
+.TP
+.B -a, --afterburner <n>
+Configure afterburner mode.
+When enabled, quality is increased at the expense of additional
+computational workload.
+.RS
+.TP
+.B 0
+Off
+.RS
+.RE
+.TP
+.B 1
+On (default)
+.RS
+.RE
+.RE
+.TP
+.B -L, --lowdelay-sbr <n>
+Configure SBR activity on AAC ELD.
+.RS
+.TP
+.B -1
+Use ELD SBR auto configuration
+.RS
+.RE
+.TP
+.B 0
+Disable SBR on ELD (default)
+.RS
+.RE
+.TP
+.B 1
+Enable SBR on ELD
+.RS
+.RE
+.RE
+.TP
+.B -s, --sbr-ratio <n>
+Controls activation of downsampled SBR.
+.RS
+.TP
+.B 0
+Use lib default (default)
+.RS
+.RE
+.TP
+.B 1
+Use downsampled SBR (default for ELD+SBR)
+.RS
+.RE
+.TP
+.B 2
+Use dual-rate SBR (default for HE-AAC)
+.RS
+.RE
+.PP
+Dual-rate SBR is what is normally used for HE-AAC, where AAC is encoded
+at half the sample rate of SBR, hence "dual rate".
+On the other hand, downsampled SBR uses same sample rate for both of AAC
+and SBR (single rate), therefore downsampled SBR typically consumes more
+bitrate.
+.PP
+Downsampled SBR is newly introduced feature in FDK encoder library
+version 3.4.12.
+When libfdk-aac in the system doesn\[aq]t support this, dual-rate SBR
+will be used.
+When available, dual-rate SBR is the default for HE-AAC and downsampled
+SBR is the default for ELD+SBR.
+.PP
+Note that downsampled HE-AAC is not so common as dual-rate one.
+When downsampled HE-AAC is selected, \f[B]fdkaac\f[] is forced to choose
+explicit hierarchical SBR signaling, which (at least) iTunes doesn\[aq]t
+accept.
+.RE
+.TP
+.B -f, --transport-format <n>
+Transport format.
+Tagging and gapless playback is only available on M4A.
+Streaming to stdout is only available on others.
+.RS
+.TP
+.B 0
+M4A (default)
+.RS
+.RE
+.TP
+.B 1
+ADIF
+.RS
+.RE
+.TP
+.B 2
+ADTS
+.RS
+.RE
+.TP
+.B 6
+LATM MCP=1
+.RS
+.RE
+.TP
+.B 7
+LATM MCP=0
+.RS
+.RE
+.TP
+.B 10
+LOAS/LATM (LATM within LOAS)
+.RS
+.RE
+.RE
+.TP
+.B -C, --adts-crc-check
+Add CRC protection on ADTS header.
+.RS
+.RE
+.TP
+.B -h, --header-period <n>
+StreamMuxConfig/PCE repetition period in transport layer.
+.RS
+.RE
+.TP
+.B -G, --gapless-mode <n>
+Method to declare amount of encoder delay (and padding) in M4A
+container.
+These values are mandatory for proper gapless playback on player side.
+.RS
+.TP
+.B 0
+iTunSMPB (default)
+.RS
+.RE
+.TP
+.B 1
+ISO standard (edts and sgpd)
+.RS
+.RE
+.TP
+.B 2
+Both
+.RS
+.RE
+.RE
+.TP
+.B --include-sbr-delay
+When specified, count SBR decoder delay in encoder delay.
+.RS
+.PP
+This is not iTunes compatible and will lead to gapless playback issue on
+LC only decoder, but this is the default behavior of FDK library.
+.PP
+Whether counting SBR decoder delay in encoder delay or not result in
+incompatibility in gapless playback.
+You should pick which one will work for your favorite player.
+.PP
+However, it\[aq]s better not to choose SBR at all if you want gapless
+playback.
+LC doesn\[aq]t have such issues.
+.RE
+.TP
+.B -I, --ignorelength
+Ignore length field of data chunk in input WAV file.
+.RS
+.RE
+.TP
+.B -S, --silent
+Don\[aq]t print progress messages.
+.RS
+.RE
+.TP
+.B --moov-before-mdat
+Place moov box before mdat box in M4A container.
+This option might be important for some hardware players, that are known
+to refuse moov box placed after mdat box.
+.RS
+.RE
+.TP
+.B -R, --raw
+Regard input as raw PCM.
+.RS
+.RE
+.TP
+.B --raw-channels <n>
+Specify number of channels of raw input (default: 2)
+.RS
+.RE
+.TP
+.B --raw-rate <n>
+Specify sample rate of raw input (default: 44100)
+.RS
+.RE
+.TP
+.B --raw-format <spec>
+Specify sample format of raw input (default: "S16L").
+\f[B]Spec\f[] is as the following (case insensitive):
+.RS
+.TP
+.B 1st char -- type of sample
+\f[B]S\f[] (igned) | \f[B]U\f[] (nsigned) | \f[B]F\f[] (loat)
+.RS
+.RE
+.TP
+.B 2nd part (in digits)
+bits per channel
+.RS
+.RE
+.TP
+.B Last char -- endianness (can be ommited)
+\f[B]L\f[] (ittle, default) | \f[B]B\f[] (ig)
+.RS
+.RE
+.RE
+.TP
+.B --title <string>
+Set title tag.
+.RS
+.RE
+.TP
+.B --artist <string>
+Set artist tag.
+.RS
+.RE
+.TP
+.B --album <string>
+Set album tag.
+.RS
+.RE
+.TP
+.B --genre <string>
+Set genre tag.
+.RS
+.RE
+.TP
+.B --date <string>
+Set date tag.
+.RS
+.RE
+.TP
+.B --composer <string>
+Set composer tag.
+.RS
+.RE
+.TP
+.B --grouping <string>
+Set grouping tag.
+.RS
+.RE
+.TP
+.B --comment <string>
+Set comment tag.
+.RS
+.RE
+.TP
+.B --album-artist <string>
+Set album artist tag.
+.RS
+.RE
+.TP
+.B --track <number[/total]>
+Set track tag, with or without number of total tracks.
+.RS
+.RE
+.TP
+.B --disk <number[/total]>
+Set disk tag, with or without number of total discs.
+.RS
+.RE
+.TP
+.B --tempo <n>
+Set tempo (BPM) tag.
+.RS
+.RE
+.TP
+.B --tag <fcc>:<value>
+Set iTunes predefined tag with explicit fourcc key and value.
+See
+https://code.google.com/p/mp4v2/wiki/iTunesMetadata (https://code.google.com/p/mp4v2/wiki/iTunesMetadata)
+for known predefined keys.
+You can omit first char of \f[B]fcc\f[] when it is the copyright sign.
+.RS
+.RE
+.TP
+.B --tag-from-file <fcc>:<filename>
+Same as --tag, but set content of file as tag value.
+.RS
+.RE
+.TP
+.B --long-tag <name>:<value>
+Set arbitrary tag as iTunes custom metadata.
+Stored in com.apple.iTunes field.
+.RS
+.RE
+.TP
+.B --tag-from-json <filename[?dot_notation]>
+Read tags from JSON.
+By default, tags are assumed to be direct children of the root object in
+JSON.
+Optionary you can speficy arbitrary dot notation to locate the object
+containing tags.
+.RS
+.RE
+.PP
+\ 
+.SH EXAMPLES
+.PP
+Encode WAV file into a M4A file.
+MPEG4 AAC LC, VBR quality 3:
+.IP
+.nf
+\f[C]
+fdkaac\ -m3\ foo.wav
+\f[]
+.fi
+.PP
+Encode WAV file into a M4A file.
+MPEG4 HE-AAC, bitrate 64kbps:
+.IP
+.nf
+\f[C]
+fdkaac\ -p5\ -b64\ foo.wav
+\f[]
+.fi
+.PP
+Piping from \f[B]ffmpeg\f[] (you need version supporting CAF output):
+.IP
+.nf
+\f[C]
+ffmpeg\ -i\ foo.flac\ -f\ caf\ -\ |\ fdkaac\ -b128\ -\ -o\ foo.m4a
+\f[]
+.fi
+.PP
+Import tags via json:
+.IP
+.nf
+\f[C]
+ffprobe\ -v\ 0\ -of\ json\ -show_format\ foo.flac\ >foo.json
+
+flac\ -dc\ foo.flac\ |\ \\
+fdkaac\ -\ -ox.m4a\ -m2\ --import-tag-from-json=foo.json?format.tags
+\f[]
+.fi
+.SH NOTES
+.PP
+Upto 32bit integer or 64bit floating point format is supported as input.
+However, FDK library is implemented based on fixed point math and only
+supports 16bit integer PCM.
+Therefore, be careful of clipping.
+You might want to dither/noise shape beforehand when your input has
+higher resolution.
+.PP
+Following channel layouts are supported by the encoder.
+.TP
+.B 1ch
+C
+.RS
+.RE
+.TP
+.B 2ch
+L R
+.RS
+.RE
+.TP
+.B 3ch
+C L R
+.RS
+.RE
+.TP
+.B 4ch
+C L R Cs
+.RS
+.RE
+.TP
+.B 5ch
+C L R Ls Rs
+.RS
+.RE
+.TP
+.B 5.1ch
+C L R Ls Rs LFE
+.RS
+.RE
+.TP
+.B 7.1ch (front)
+C Lc Rc L R Ls Rs LFE
+.RS
+.RE
+.TP
+.B 7.1ch (rear)
+C L R Ls Rs Rls Rrs LFE
+.RS
+.RE
+.SH AUTHORS
+nu774 <honeycomb77@gmail.com>.

src/aacenc.c

@@ -13,6 +13,17 @@
 #include <string.h>
 #include "aacenc.h"
 
+int aacenc_is_sbr_ratio_available()
+{
+#if AACENCODER_LIB_VL0 < 3 || (AACENCODER_LIB_VL0==3 && AACENCODER_LIB_VL1<4)
+    return 0;
+#else
+    LIB_INFO lib_info;
+    aacenc_get_lib_info(&lib_info);
+    return lib_info.version > 0x03040000;
+#endif
+}
+
 int aacenc_is_sbr_active(const aacenc_param_t *params)
 {
     switch (params->profile) {
@@ -26,66 +37,31 @@ int aacenc_is_sbr_active(const aacenc_param_t *params)
     return 0;
 }
 
-static const unsigned aacenc_sampling_freq_tab[] = {
-    96000, 88200, 64000, 48000, 44100, 32000, 24000, 22050, 
-    16000, 12000, 11025, 8000, 7350, 0, 0, 0
-};
-
-static
-unsigned sampling_freq_index(unsigned rate)
+int aacenc_is_dual_rate_sbr(const aacenc_param_t *params)
 {
-    unsigned i;
-    for (i = 0; aacenc_sampling_freq_tab[i]; ++i)
-        if (aacenc_sampling_freq_tab[i] == rate)
-            return i;
-    return 0xf;
+    if (params->profile == AOT_PS || params->profile == AOT_MP2_PS)
+        return 1;
+    else if (params->profile == AOT_SBR || params->profile == AOT_MP2_SBR)
+        return params->sbr_ratio == 0 || params->sbr_ratio == 2;
+    else if (params->profile == AOT_ER_AAC_ELD && params->lowdelay_sbr)
+        return params->sbr_ratio == 2;
+    return 0;
 }
 
-/*
- * Append backward compatible SBR/PS signaling to implicit signaling ASC,
- * if SBR/PS is present.
- */
-int aacenc_mp4asc(const aacenc_param_t *params,
-                  const uint8_t *asc, uint32_t ascsize,
-                  uint8_t *outasc, uint32_t *outsize)
+void aacenc_get_lib_info(LIB_INFO *info)
 {
-    unsigned asc_sfreq = aacenc_sampling_freq_tab[(asc[0]&0x7)<<1 |asc[1]>>7];
-
-    switch (params->profile) {
-    case AOT_SBR:
-    case AOT_PS:
-        if (*outsize < ascsize + 3)
-            return -1;
-        memcpy(outasc, asc, ascsize);
-        /* syncExtensionType:11 (value:0x2b7) */
-        outasc[ascsize+0] = 0x2b << 1;
-        outasc[ascsize+1] = 0x7 << 5;
-        /* extensionAudioObjectType:5 (value:5)*/
-        outasc[ascsize+1] |= 5;
-        /* sbrPresentFlag:1 (value:1) */
-        outasc[ascsize+2] = 0x80;
-        /* extensionSamplingFrequencyIndex:4 */
-        outasc[ascsize+2] |= sampling_freq_index(asc_sfreq << 1) << 3;
-        if (params->profile == AOT_SBR) {
-            *outsize = ascsize + 3;
-            break;
+    LIB_INFO *lib_info = 0;
+    lib_info = calloc(FDK_MODULE_LAST, sizeof(LIB_INFO));
+    if (aacEncGetLibInfo(lib_info) == AACENC_OK) {
+        int i;
+        for (i = 0; i < FDK_MODULE_LAST; ++i) {
+            if (lib_info[i].module_id == FDK_AACENC) {
+                memcpy(info, &lib_info[i], sizeof(LIB_INFO));
+                break;
+            }
         }
-        if (*outsize < ascsize + 5)
-            return -1;
-        /* syncExtensionType:11 (value:0x548) */
-        outasc[ascsize+2] |= 0x5;
-        outasc[ascsize+3] = 0x48;
-        /* psPresentFlag:1 (value:1) */
-        outasc[ascsize+4] = 0x80;
-        *outsize = ascsize + 5;
-        break;
-    default:
-        if (*outsize < ascsize)
-            return -1;
-        memcpy(outasc, asc, ascsize);
-        *outsize = ascsize;
     }
-    return 0;
+    free(lib_info);
 }
 
 static
@@ -93,10 +69,10 @@ int aacenc_channel_mode(const pcm_sample_description_t *format)
 {
     uint32_t chanmask = format->channel_mask;
 
-    if (format->channels_per_frame > 6)
+    if (format->channels_per_frame > 8)
         return 0;
     if (!chanmask) {
-        static uint32_t defaults[] = { 0x4, 0x3, 0x7, 0, 0x37, 0x3f };
+        static uint32_t defaults[] = { 0x4, 0x3, 0x7, 0, 0x37, 0x3f, 0, 0x63f };
         chanmask = defaults[format->channels_per_frame - 1];
     }
     switch (chanmask) {
@@ -108,6 +84,10 @@ int aacenc_channel_mode(const pcm_sample_description_t *format)
     case 0x107: return MODE_1_2_1;
     case 0x607: return MODE_1_2_2;
     case 0x60f: return MODE_1_2_2_1;
+#if AACENCODER_LIB_VL0 > 3 || (AACENCODER_LIB_VL0==3 && AACENCODER_LIB_VL1>=4)
+    case 0xff:  return MODE_1_2_2_2_1;
+    case 0x63f: return MODE_7_1_REAR_SURROUND;
+#endif
     }
     return 0;
 }
@@ -118,8 +98,11 @@ int aacenc_init(HANDLE_AACENCODER *encoder, const aacenc_param_t *params,
 {
     int channel_mode;
     int aot;
+    LIB_INFO lib_info;
 
     *encoder = 0;
+    aacenc_get_lib_info(&lib_info);
+
     if ((channel_mode = aacenc_channel_mode(format)) == 0) {
         fprintf(stderr, "ERROR: unsupported channel layout\n");
         goto FAIL;
@@ -145,13 +128,21 @@ int aacenc_init(HANDLE_AACENCODER *encoder, const aacenc_param_t *params,
         fprintf(stderr, "ERROR: unsupported sample rate\n");
         goto FAIL;
     }
-    aacEncoder_SetParam(*encoder, AACENC_CHANNELMODE, channel_mode);
+    if (aacEncoder_SetParam(*encoder, AACENC_CHANNELMODE,
+                            channel_mode) != AACENC_OK) {
+        fprintf(stderr, "ERROR: unsupported channel mode\n");
+        goto FAIL;
+    }
     aacEncoder_SetParam(*encoder, AACENC_BANDWIDTH, params->bandwidth);
     aacEncoder_SetParam(*encoder, AACENC_CHANNELORDER, 1);
     aacEncoder_SetParam(*encoder, AACENC_AFTERBURNER, !!params->afterburner);
 
-    if (aot == AOT_ER_AAC_ELD && params->lowdelay_sbr)
-        aacEncoder_SetParam(*encoder, AACENC_SBR_MODE, 1);
+    aacEncoder_SetParam(*encoder, AACENC_SBR_MODE, params->lowdelay_sbr);
+
+#if AACENCODER_LIB_VL0 > 3 || (AACENCODER_LIB_VL0==3 && AACENCODER_LIB_VL1>=4)
+    if (lib_info.version > 0x03040000)
+        aacEncoder_SetParam(*encoder, AACENC_SBR_RATIO, params->sbr_ratio);
+#endif
 
     if (aacEncoder_SetParam(*encoder, AACENC_TRANSMUX,
                             params->transport_format) != AACENC_OK) {

src/aacenc.h

@@ -15,6 +15,7 @@
     unsigned bandwidth; \
     unsigned afterburner; \
     unsigned lowdelay_sbr; \
+    unsigned sbr_ratio; \
     unsigned sbr_signaling; \
     unsigned transport_format; \
     unsigned adts_crc_check; \
@@ -29,11 +30,13 @@ typedef struct aacenc_frame_t {
     uint32_t size, capacity;
 } aacenc_frame_t;
 
+int aacenc_is_sbr_ratio_available();
+
 int aacenc_is_sbr_active(const aacenc_param_t *params);
 
-int aacenc_mp4asc(const aacenc_param_t *params,
-                  const uint8_t *asc, uint32_t ascsize,
-                  uint8_t *outasc, uint32_t *outsize);
+int aacenc_is_dual_rate_sbr(const aacenc_param_t *params);
+
+void aacenc_get_lib_info(LIB_INFO *info);
 
 int aacenc_init(HANDLE_AACENCODER *encoder, const aacenc_param_t *params,
                 const pcm_sample_description_t *format,

src/main.c

@@ -54,7 +54,8 @@ static void handle_signals(void)
 {
     int i, sigs[] = { SIGINT, SIGHUP, SIGTERM };
     for (i = 0; i < sizeof(sigs)/sizeof(sigs[0]); ++i) {
-        struct sigaction sa = { 0 };
+        struct sigaction sa;
+        memset(&sa, 0, sizeof sa);
         sa.sa_handler = signal_handler;
         sa.sa_flags |= SA_RESTART;
         sigaction(sigs[i], &sa, 0);
@@ -132,7 +133,14 @@ PROGNAME " %s\n"
 " -a, --afterburner <n>         Afterburner\n"
 "                                 0: Off\n"
 "                                 1: On(default)\n"
-" -L, --lowdelay-sbr            Enable ELD-SBR (AAC ELD only)\n"
+" -L, --lowdelay-sbr <-1|0|1>   Configure SBR activity on AAC ELD\n"
+"                                -1: Use ELD SBR auto configurator\n"
+"                                 0: Disable SBR on ELD (default)\n"
+"                                 1: Enable SBR on ELD\n"
+" -s, --sbr-ratio <0|1|2>       Controls activation of downsampled SBR\n"
+"                                 0: Use lib default (default)\n"
+"                                 1: downsampled SBR (default for ELD+SBR)\n"
+"                                 2: dual-rate SBR (default for HE-AAC)\n"
 " -f, --transport-format <n>    Transport format\n"
 "                                 0: RAW (default, muxed into M4A)\n"
 "                                 1: ADIF\n"
@@ -228,7 +236,7 @@ static
 int parse_options(int argc, char **argv, aacenc_param_ex_t *params)
 {
     int ch;
-    unsigned n;
+    int n;
 
 #define OPT_INCLUDE_SBR_DELAY    M4AF_FOURCC('s','d','l','y')
 #define OPT_MOOV_BEFORE_MDAT     M4AF_FOURCC('m','o','o','v')
@@ -247,7 +255,8 @@ int parse_options(int argc, char **argv, aacenc_param_ex_t *params)
         { "bitrate-mode",     required_argument, 0, 'm' },
         { "bandwidth",        required_argument, 0, 'w' },
         { "afterburner",      required_argument, 0, 'a' },
-        { "lowdelay-sbr",     no_argument,       0, 'L' },
+        { "lowdelay-sbr",     required_argument, 0, 'L' },
+        { "sbr-ratio",        required_argument, 0, 's' },
         { "transport-format", required_argument, 0, 'f' },
         { "adts-crc-check",   no_argument,       0, 'C' },
         { "header-period",    required_argument, 0, 'P' },
@@ -325,7 +334,18 @@ int parse_options(int argc, char **argv, aacenc_param_ex_t *params)
             params->afterburner = n;
             break;
         case 'L':
-            params->lowdelay_sbr = 1;
+            if (sscanf(optarg, "%d", &n) != 1 || n < -1 || n > 1) {
+                fprintf(stderr, "invalid arg for lowdelay-sbr\n");
+                return -1;
+            }
+            params->lowdelay_sbr = n;
+            break;
+        case 's':
+            if (sscanf(optarg, "%u", &n) != 1 || n > 2) {
+                fprintf(stderr, "invalid arg for sbr-ratio\n");
+                return -1;
+            }
+            params->sbr_ratio = n;
             break;
         case 'f':
             if (sscanf(optarg, "%u", &n) != 1) {
@@ -567,19 +587,11 @@ void put_tool_tag(m4af_ctx_t *m4af, const aacenc_param_ex_t *params,
 {
     char tool_info[256];
     char *p = tool_info;
-    LIB_INFO *lib_info = 0;
+    LIB_INFO lib_info;
 
     p += sprintf(p, PROGNAME " %s, ", fdkaac_version);
-
-    lib_info = calloc(FDK_MODULE_LAST, sizeof(LIB_INFO));
-    if (aacEncGetLibInfo(lib_info) == AACENC_OK) {
-        int i;
-        for (i = 0; i < FDK_MODULE_LAST; ++i)
-            if (lib_info[i].module_id == FDK_AACENC)
-                break;
-        p += sprintf(p, "libfdk-aac %s, ", lib_info[i].versionStr);
-    }
-    free(lib_info);
+    aacenc_get_lib_info(&lib_info);
+    p += sprintf(p, "libfdk-aac %s, ", lib_info.versionStr);
     if (params->bitrate_mode)
         sprintf(p, "VBR mode %d", params->bitrate_mode);
     else
@@ -758,9 +770,9 @@ int main(int argc, char **argv)
     AACENC_InfoStruct aacinfo = { 0 };
     m4af_ctx_t *m4af = 0;
     const pcm_sample_description_t *sample_format;
-    int downsampled_timescale = 0;
     int frame_count = 0;
     int sbr_mode = 0;
+    unsigned scale_shift = 0;
 
     setlocale(LC_CTYPE, "");
     setbuf(stderr, 0);
@@ -773,16 +785,18 @@ int main(int argc, char **argv)
 
     sample_format = pcm_get_format(reader);
 
-    /*
-     * We use explicit/hierarchical signaling for LOAS.
-     * Other than that, we request implicit signaling to FDK library, then 
-     * append explicit/backward-compatible signaling to ASC in case of MP4FF.
-     *
-     * Explicit/backward-compatible signaling of SBR is the most recommended
-     * way in MPEG4 part3 spec, and seems the only way supported by iTunes.
-     * Since FDK library does not support it, we have to do it on our side.
-     */
-    params.sbr_signaling = (params.transport_format == TT_MP4_LOAS) ? 2 : 0;
+    sbr_mode = aacenc_is_sbr_active((aacenc_param_t*)&params);
+    if (sbr_mode && !aacenc_is_sbr_ratio_available()) {
+        fprintf(stderr, "WARNING: Only dual-rate SBR is available "
+                        "for this version\n");
+        params.sbr_ratio = 2;
+    }
+    scale_shift = aacenc_is_dual_rate_sbr((aacenc_param_t*)&params);
+    params.sbr_signaling =
+        (params.transport_format == TT_MP4_LOAS) ? 2 :
+        (params.transport_format == TT_MP4_RAW)  ? 1 : 0;
+    if (sbr_mode && !scale_shift)
+        params.sbr_signaling = 2;
 
     if (aacenc_init(&encoder, (aacenc_param_t*)&params, sample_format,
                     &aacinfo) < 0)
@@ -800,28 +814,23 @@ int main(int argc, char **argv)
         goto END;
     }
     handle_signals();
-    sbr_mode = aacenc_is_sbr_active((aacenc_param_t*)&params);
+
     if (!params.transport_format) {
         uint32_t scale;
-        uint8_t mp4asc[32];
-        uint32_t ascsize = sizeof(mp4asc);
         unsigned framelen = aacinfo.frameLength;
-        if (sbr_mode)
-            downsampled_timescale = 1;
-        scale = sample_format->sample_rate >> downsampled_timescale;
+        scale = sample_format->sample_rate >> scale_shift;
         if ((m4af = m4af_create(M4AF_CODEC_MP4A, scale, &m4af_io,
                                 params.output_fp)) < 0)
             goto END;
-        aacenc_mp4asc((aacenc_param_t*)&params, aacinfo.confBuf,
-                      aacinfo.confSize, mp4asc, &ascsize);
-        m4af_set_decoder_specific_info(m4af, 0, mp4asc, ascsize);
+        m4af_set_decoder_specific_info(m4af, 0,
+                                       aacinfo.confBuf, aacinfo.confSize);
         m4af_set_fixed_frame_duration(m4af, 0,
-                                      framelen >> downsampled_timescale);
+                                      framelen >> scale_shift);
         m4af_set_vbr_mode(m4af, 0, params.bitrate_mode);
         m4af_set_priming_mode(m4af, params.gapless_mode + 1);
         m4af_begin_write(m4af);
     }
-    if (sbr_mode && (aacinfo.encoderDelay & 1)) {
+    if (scale_shift && (aacinfo.encoderDelay & 1)) {
         /*
          * Since odd delay cannot be exactly expressed in downsampled scale,
          * we push one zero frame to the encoder here, to make delay even
@@ -841,12 +850,11 @@ int main(int argc, char **argv)
 
         if (sbr_mode && params.profile != AOT_ER_AAC_ELD &&
             !params.include_sbr_delay)
-            delay -= 481 << 1;
-        if (sbr_mode && (delay & 1))
+            delay -= 481 << scale_shift;
+        if (scale_shift && (delay & 1))
             ++delay;
         padding = frame_count * aacinfo.frameLength - frames_read - delay;
-        m4af_set_priming(m4af, 0, delay >> downsampled_timescale,
-                         padding >> downsampled_timescale);
+        m4af_set_priming(m4af, 0, delay >> scale_shift, padding >> scale_shift);
         if (finalize_m4a(m4af, &params, encoder) < 0)
             goto END;
     }

version.h

@@ -1,4 +1,4 @@
 #ifndef VERSION_H
 #define VERSION_H
-const char *fdkaac_version = "0.4.2";
+const char *fdkaac_version = "0.5.1";
 #endif