bicsf(4) bicsf(4)
BICSF, IRCAM - Berkeley/IRCAM/CARL Sound File Format
#include <dmedia/audiofile.h>
The Audio File Library currently supports 13 of the commonly found audio
file formats, i.e., is able to recognize, read, and write sample data and
header information to and from files in these formats. It is important
not to confuse sample or audio data formats with file formats. The
former refers to the bit-wise organization of the sound samples in the
file, i.e., whether the format is 8-bit integer or 16-bit unsigned, etc.
Audio file format refers to the structure of the audio file header, the
chunk of on-disk data which preceeds the samples and which provides
information about the file to the audio program. A single audio file
format may support a large variety of sample formats.
The Berkeley/IRCAM/CARL Sound File Format (bicsf) was developed during
the 1980's and represented the merging of several earlier header formats.
It is used extensively in academic computer music centers, and is the
default file format for several sound synthesis packages such as MIT's
Csound. It consists of a fixed-length header of 1024 bytes. The first
17 bytes are reserved for sample format information; the remainder is
used for optional "sfcodes" which provide additional information such as
maximum amplitude, creation date, or a text comment.
DATA FORMAT SPECIFICATIONS [Toc] [Back] Sample Formats:
Two's complement integer and 32-bit single-precision floating point
only.
Sample Widths:
For integer: 16 bit only; floating points are always 32 bit.
Byte Orders:
Only bigendian currently supported for writing; littleendian files,
generated on DEC Alpha machines and some other platforms, are
writable and readable.
Channel Counts:
1, 2, and 4 channels only.
Compression Formats:
None supported, and none in common usage.
FILE FORMAT SPECIFICATIONS [Toc] [Back] BICSF files can contain additional chunks of data. Most of these are
parsed by the AF and made available:
Page 1
bicsf(4) bicsf(4)
Instrument Configurations:
None allowed.
Markers:
None allowed.
Miscellaneous Chunks:
AF_MISC_IRCAM_PEAKAMP peak amplitude sfcode (see CAVEATS)
AF_MISC_COMMENT text comment
The miscellaneous chunk AF_MISC_IRCAM_PEAKAMP is always 36 bytes and
contains a fixed "sfcode" structure:
typedef struct _sfmaxamp {
float value[4];
int samploc[4];
int timetag;
} SfMaxamp;
Care must be taken when using afWriteMisc(3dm) to provide the correct
number of bytes of data, or the result will be unreadable.
The miscellaneous chunk AF_MISC_COMMENT has a maximum size of 512 bytes,
and will always be allocated out to 256 bytes. This does not affect the
size of the BICSF file header. This chunk was formerly known as
AF_MISC_IRCAM_COMMENT and can still be referred to as this, for backwards
compatibility.
afInitFileFormat(3dm), afGetFileFormat(3dm), afIntro(3dm)
bicsf(4) bicsf(4)
BICSF, IRCAM - Berkeley/IRCAM/CARL Sound File Format
#include <dmedia/audiofile.h>
The Audio File Library currently supports 13 of the commonly found audio
file formats, i.e., is able to recognize, read, and write sample data and
header information to and from files in these formats. It is important
not to confuse sample or audio data formats with file formats. The
former refers to the bit-wise organization of the sound samples in the
file, i.e., whether the format is 8-bit integer or 16-bit unsigned, etc.
Audio file format refers to the structure of the audio file header, the
chunk of on-disk data which preceeds the samples and which provides
information about the file to the audio program. A single audio file
format may support a large variety of sample formats.
The Berkeley/IRCAM/CARL Sound File Format (bicsf) was developed during
the 1980's and represented the merging of several earlier header formats.
It is used extensively in academic computer music centers, and is the
default file format for several sound synthesis packages such as MIT's
Csound. It consists of a fixed-length header of 1024 bytes. The first
17 bytes are reserved for sample format information; the remainder is
used for optional "sfcodes" which provide additional information such as
maximum amplitude, creation date, or a text comment.
DATA FORMAT SPECIFICATIONS [Toc] [Back] Sample Formats:
Two's complement integer and 32-bit single-precision floating point
only.
Sample Widths:
For integer: 16 bit only; floating points are always 32 bit.
Byte Orders:
Only bigendian currently supported for writing; littleendian files,
generated on DEC Alpha machines and some other platforms, are
writable and readable.
Channel Counts:
1, 2, and 4 channels only.
Compression Formats:
None supported, and none in common usage.
FILE FORMAT SPECIFICATIONS [Toc] [Back] BICSF files can contain additional chunks of data. Most of these are
parsed by the AF and made available:
Page 1
bicsf(4) bicsf(4)
Instrument Configurations:
None allowed.
Markers:
None allowed.
Miscellaneous Chunks:
AF_MISC_IRCAM_PEAKAMP peak amplitude sfcode (see CAVEATS)
AF_MISC_COMMENT text comment
The miscellaneous chunk AF_MISC_IRCAM_PEAKAMP is always 36 bytes and
contains a fixed "sfcode" structure:
typedef struct _sfmaxamp {
float value[4];
int samploc[4];
int timetag;
} SfMaxamp;
Care must be taken when using afWriteMisc(3dm) to provide the correct
number of bytes of data, or the result will be unreadable.
The miscellaneous chunk AF_MISC_COMMENT has a maximum size of 512 bytes,
and will always be allocated out to 256 bytes. This does not affect the
size of the BICSF file header. This chunk was formerly known as
AF_MISC_IRCAM_COMMENT and can still be referred to as this, for backwards
compatibility.
afInitFileFormat(3dm), afGetFileFormat(3dm), afIntro(3dm)
PPPPaaaaggggeeee 2222 [ Back ]
|
