8.6 KiB
Mediaextract
Extract media files that are embedded within other files.
Setup
make builddir
make BUILD_TYPE=release
sudo make install BUILD_TYPE=release PREFIX=/usr
The default BUILD_TYPE is debug.
Cross compile for Windows (uses i686-w64-mingw32-gcc):
make TARGET=win32 builddir
make TARGET=win32
Or (uses x86_64-w64-mingw32-gcc):
make TARGET=win64 builddir
make TARGET=win64
Warning: This program only works correctly on platforms that allow unaligned memory access (e.g. x86 and ARM, although it might be quite slow on the latter).
Note: 32bit binaries can only process 2 GB of a file at once. The rest of
bigger files will be ignored. You need to run this program several times with
different --offset values to process such a file whole.
This also means that using a 32bit binary extracted files can never be larger than 2 GB.
This is because mediaextract uses mmap to read files, which maps files to
memory. On 32bit platforms the address space of the main memory is simply not
big enough. 64bit binaries can read up to 8 EB (8 Exabytes) at once.
Usage
mediaextract [option...] <filename> [<filename> ...]
Examples
Extract .wav, .aif and .ogg (might actually be .ogg, .opus or .ogm) files from
the file data.bin and store them in the ~/Music directory.
mediaextract -f riff,aiff,ogg -o ~/Music data.bin
This will then write files like such into ~/Music:
data.bin_00000000.ogg
data.bin_00FFB2E3.wav
data.bin_01F3CD45.aif
The hexadecimal number in the written file names give the offset where in the data file the audio file was found.
Extract .mp3, .mp2 and .mp1 files (with or without ID3v2 tags). The mpg123
option yields a lot of false positives because there is no nice way to
unambigiously detect MPEG files. These false positives are however usually very
small, so using the --min-size option one can hopefully extract only real MPEG
files.
mediaextract -f id3v2,mpg123 --min-size=100k -o ~/Music data.bin
Options
-h, --help Print this help message.
-v, --version Print program version.
-q, --quiet Do not print status messages.
-s, --simulate Don't write any output files.
-o, --output=DIR Directory where extracted files should be written. (default: ".")
-a, --filename=FORMAT Format string for the file names.
(default: "{filename}_{offset}.{ext}")
Supported variables:
filename Filename of the extracted archive.
offset Offset within the archive in hexadecimal.
index 0-based index of the extracted file in decimal.
size Size of the extracted file in decimal.
ext Extension associated with the filetype of the
extracted file.
-i, --offset=OFFSET Start processing at byte OFFSET. (default: 0)
-n, --length=LENGTH Only process LENGTH bytes.
(default and maximum: 8 EB)
-m, --min-size=SIZE Minumum size of extracted files (skip smaller). (default: 0)
-x, --max-size=SIZE Maximum size of extracted files (skip larger).
(default and maximum: 16 EB)
The last character of OFFSET, LENGTH and SIZE may be one of the
following:
B (or none) for Bytes
k for Kilobytes (units of 1024 Bytes)
M for Megabytes (units of 1024 Kilobytes)
G for Gigabytes (units of 1024 Megabytes)
T for Terabytes (units of 1024 Gigabytes)
P for Petabytes (units of 1024 Terabytes)
E for Exabytes (units of 1024 Petabytes)
The special value "max" selects the maximum alowed value.
-f, --formats=FORMATS Comma separated list of formats (file magics) to extract.
Supported formats:
all all supported formats
default the default set of formats (AIFF, ASF, AU, AVIF, BINK,
BMP, GIF, HEIF, ID3v2, IT, JPEG, MPEG 1, MPEG PS,
MIDI, MP4, Ogg, PNG, RIFF, S3M, SMK, XM, XMIDI)
audio all audio files (AIFF, ASF, AU, ID3v2, IT, MIDI, MP4,
Ogg, RIFF, S3M, XM, XMIDI)
text all text files (ASCII, UTF-8, UTF-16LE, UTF-16BE,
UTF-32LE, UTF-32BE)
image all image files (BMP, PNG, JPEG, GIF, AVIF, HEIF)
mpeg all safe mpeg files (MPEG 1, MPEG PS, ID3v2)
tracker all tracker files (MOD, S3M, IT, XM)
video all video files (ASF, BINK, MP4, RIFF, SMK)
avif AVIF image files
aiff big-endian (Apple) wave files
ascii 7-bit ASCII files (only printable characters)
asf Advanced Systems Format files (also WMA and WMV)
au Sun Microsystems audio file format (.au or .snd)
bink BINK files
bmp Windows Bitmap files
gif Graphics Interchange Format files
heif HEIF images files
id3v2 MPEG layer 1/2/3 files with ID3v2 tags
it ImpulseTracker files
jpeg JPEG Interchange Format files
midi MIDI files
mod Noisetracker/Soundtracker/Protracker Module files
mpg123 MPEG layer 1/2/3 files (MP1, MP2, MP3)
mpeg1 MPEG 1 System Streams
mpegps MPEG 2 Program Streams
mpegts MPEG 2 Transport Streams
mp4 MP4 files (M4A, M4V, 3GPP etc.)
ogg Ogg files (Vorbis, Opus, Theora, etc.)
png Portable Network Graphics files
riff Resource Interchange File Format files (ANI, AVI, MMM,
PAL, RDI, RMI, SGT, STY, WAV, WEBP and more)
s3m ScreamTracker III files
smk Smaker files
utf-8 7-bit ASCII and UTF-8 files (only printable code points)
utf-16be big-endian UTF-16 files (only printable code points)
utf-16le little-endian UTF-16 files (only printable code points)
utf-32be big-endian UTF-32 files (only printable code points)
utf-32le little-endian UTF-32 files (only printable code points)
xm Extended Module files
xmidi XMIDI files
WARNING: Because MP1/2/3 files do not have a nice file magic, using
the 'mpg123' format may cause *a lot* of false positives. Nowadays
MP3 files usually have an ID3v2 tag at the start, so using the
'id3v2' format is the better option anyway.
The detection accuracy of MOD files is not much better and of MPEG TS
it is even worse and thus the 'mpg123', 'mpegts' and 'mod' formats
are per default disabled.
NOTE: When using only the 'mpg123' format but not 'id3v2' any ID3v2
tag will be stripped. ID3v1 tags will still be kept.
NOTE: The 'text' format might detect too much bogus text in UTF-16 or
UTF-32 encodings. I recommend to use 'utf-8' or 'ascii' instead, if
you can.
If '-' is written before a format name the format will be
removed from the set of formats to extract. E.g. extract
everything except tracker files:
mediaextract --formats=all,-tracker data.bin