cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: libsndfile : Frequently Asked Questions. cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

libsndfile : Frequently Asked Questions.

cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q1 : Do you plan to support XYZ codec in libsndfile?

cannam@85:

cannam@85: If source code for XYZ codec is available under a suitable license (LGPL, BSD, cannam@85: MIT etc) then yes, I'd like to add it. cannam@85:

cannam@85:

cannam@85: If suitable documentation is available on how to decode and encode the format cannam@85: then maybe, depending on how much work is involved. cannam@85:

cannam@85:

cannam@85: If XYZ is some proprietary codec where no source code or documentation is cannam@85: available then no. cannam@85:

cannam@85:

cannam@85: So if you want support for XYZ codec, first find existing source code or cannam@85: documentation. cannam@85: If you can't find either then the answer is no. cannam@85:

cannam@85: cannam@85: cannam@85:

Q2 : In version 0 the SF_INFO struct had a pcmbitwidth field cannam@85: but version 1 does not. Why?

cannam@85:

cannam@85: This was dropped for a number of reasons: cannam@85:

cannam@85:

pcmbitwidth makes little sense on compressed or floating point formats cannam@85:
with the new API you really don't need to know it cannam@85:

cannam@85:

cannam@85: As documented cannam@85: here cannam@85: there is now a well defined behaviour which ensures that no matter what the cannam@85: bit width of the source file, the scaling always does something sensible. cannam@85: This makes it safe to read 8, 16, 24 and 32 bit PCM files using sf_read_short() cannam@85: and always have the optimal behaviour. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q3 : Compiling is really slow on MacOS X. Why?

cannam@85:

cannam@85: When you configure and compile libsndfile, it uses the /bin/sh shell for a number cannam@85: of tasks (ie configure script and libtool). cannam@85: Older versions of OS X (10.2?) shipped a really crappy Bourne shell as /bin/sh cannam@85: which resulted in really slow compiles. cannam@85: Newer version of OS X ship GNU Bash as /bin/sh and this answer doesn't apply in that cannam@85: case. cannam@85:

cannam@85:

cannam@85: To fix this I suggest that you install the GNU Bash shell, rename /bin/sh to cannam@85: /bin/sh.old and make a symlink from /bin/sh to the bash shell. cannam@85: Bash is designed to behave as a Bourne shell when is is called as /bin/sh. cannam@85:

cannam@85:

cannam@85: When I did this on my iBook running MacOS X, compile times dropped from 13 minutes cannam@85: to 3 minutes. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q4 : When trying to compile libsndfile on Solaris I get a "bad cannam@85: substitution" error on linking. Why?

cannam@85:

cannam@85: It seems that the Solaris Bourne shell disagrees with GNU libtool. cannam@85:

cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q5 : Why doesn't libsndfile do interleaving/de-interleaving?

cannam@85:

cannam@85: This problem is bigger than it may seem at first. cannam@85:

cannam@85:

cannam@85: For a stereo file, it is a pretty safe bet that a simple interleaving/de-interleaving cannam@85: could satisfy most users. cannam@85: However, for files with more than 2 channels this is unlikely to be the case. cannam@85: If the user has a 4 channel file and want to play that file on a stereo output cannam@85: sound card they either want the first 2 channels or they want some mixed combination cannam@85: of the 4 channels. cannam@85:

cannam@85:

cannam@85: When you add more channels, the combinations grow exponentially and it becomes cannam@85: increasingly difficult to cover even a sensible subset of the possible combinations. cannam@85: On top of that, coding any one style of interleaver/de-interleaver is trivial, while cannam@85: coding one that can cover all combinations is far from trivial. cannam@85: This means that this feature will not be added any time soon. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q6 : What's the best format for storing temporary files?

cannam@85: cannam@85:

cannam@85: When you want to store temporary data there are a number of requirements; cannam@85:

cannam@85:

A simple, easy to parse header. cannam@85:
The format must provide the fastest possible read and write rates (ie cannam@85: avoid conversions and encoding/decoding). cannam@85:
The file format must be reasonably common and playable by most players. cannam@85:
Able to store data in either endian-ness. cannam@85:

cannam@85:

cannam@85: The format which best meets these requirements is AU, which allows data to be cannam@85: stored in any one of short, int, float and double (among others) formats. cannam@85:

cannam@85:

cannam@85: For instance, if an application uses float data internally, its temporary files cannam@85: should use a format of (SF_ENDIAN_CPU | SF_FORMAT_AU | SF_FORMAT_FLOAT) which cannam@85: will store big endian float data in big endian CPUs and little endian float data cannam@85: on little endian CPUs. cannam@85: Reading and writing this format will not require any conversions or byte swapping cannam@85: regardless of the host CPU. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q7 : On Linux/Unix/MaxOS X, what's the best way of detecting the presence cannam@85: of libsndfile using autoconf?

cannam@85: cannam@85:

cannam@85: libsndfile uses the pkg-config (man pkg-config) method of registering itself with the cannam@85: host system. cannam@85: The best way of detecting its presence is using something like this in configure.ac cannam@85: (or configure.in): cannam@85:

cannam@85:

cannam@85:         PKG_CHECK_MODULES(SNDFILE, sndfile >= 1.0.2, ac_cv_sndfile=1, ac_cv_sndfile=0)
cannam@85: 
cannam@85:         AC_DEFINE_UNQUOTED([HAVE_SNDFILE],${ac_cv_sndfile},
cannam@85: 			[Set to 1 if you have libsndfile.])
cannam@85: 
cannam@85:         AC_SUBST(SNDFILE_CFLAGS)
cannam@85:         AC_SUBST(SNDFILE_LIBS)
cannam@85:

cannam@85:

cannam@85: This will automatically set the SNDFILE_CFLAGS and SNDFILE_LIBS cannam@85: variables which can be used in Makefile.am like this: cannam@85:

cannam@85:

cannam@85:         SNDFILE_CFLAGS = @SNDFILE_CFLAGS@
cannam@85:         SNDFILE_LIBS = @SNDFILE_LIBS@
cannam@85:

cannam@85:

cannam@85: If you install libsndfile from source, you will probably need to set the cannam@85: PKG_CONFIG_PATH environment variable as suggested at the end of the cannam@85: libsndfile configure process. For instance on my system I get this: cannam@85:

cannam@85:

cannam@85:         -=-=-=-=-=-=-=-=-=-= Configuration Complete =-=-=-=-=-=-=-=-=-=-
cannam@85: 
cannam@85:           Configuration summary :
cannam@85: 
cannam@85:             Version : ..................... 1.0.5
cannam@85:             Experimental code : ........... no
cannam@85: 
cannam@85:           Tools :
cannam@85: 
cannam@85:             Compiler is GCC : ............. yes
cannam@85:             GCC major version : ........... 3
cannam@85: 
cannam@85:           Installation directories :
cannam@85: 
cannam@85:             Library directory : ........... /usr/local/lib
cannam@85:             Program directory : ........... /usr/local/bin
cannam@85:             Pkgconfig directory : ......... /usr/local/lib/pkgconfig
cannam@85: 
cannam@85:         Compiling some other packages against libsndfile may require
cannam@85:         the addition of "/usr/local/lib/pkgconfig" to the
cannam@85:         PKG_CONFIG_PATH environment variable.
cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q8 : But I just want a simple Makefile! What do I do?

cannam@85: cannam@85:

cannam@85: The pkg-config program makes finding the correct compiler flag values and cannam@85: library location far easier. cannam@85: During the installation of libsndfile, a file named sndfile.pc is installed cannam@85: in the directory ${libdir}/pkgconfig (ie if libsndfile is installed in cannam@85: /usr/local/lib, sndfile.pc will be installed in cannam@85: /usr/local/lib/pkgconfig/). cannam@85:

cannam@85:

cannam@85: In order for pkg-config to find sndfile.pc it may be necessary to point the cannam@85: environment variable PKG_CONFIG_PATH in the right direction. cannam@85:

cannam@85:

cannam@85:         export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig
cannam@85:

cannam@85: cannam@85:

cannam@85: Then, to compile a C file into an object file, the command would be: cannam@85:

cannam@85:

cannam@85:         gcc `pkg-config --cflags sndfile` -c somefile.c
cannam@85:

cannam@85:

cannam@85: and to link a number of objects into an executable that links against libsndfile, cannam@85: the command would be: cannam@85:

cannam@85:

cannam@85:         gcc `pkg-config --libs sndfile` obj1.o obj2.o -o program
cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q9 : How about adding the ability to write/read sound files to/from cannam@85: memory buffers?

cannam@85: cannam@85:

cannam@85: This has been added for version 1.0.13. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q10 : Reading a 16 bit PCM file as normalised floats and then cannam@85: writing them back changes some sample values. Why?

cannam@85: cannam@85:

cannam@85: This is caused by the fact that the conversion from 16 bit short to float is cannam@85: done by dividing by 32768 (0x8000 in hexadecimal) while the conversion from cannam@85: float to 16 bit short is done by multiplying by 32767 (0x7FFF in hex). cannam@85: So for instance, a value in a 16 bit PCM file of 20000 gets read as a floating cannam@85: point number of 0.6103515625 (20000.0 / 0x8000). cannam@85: Converting that back to a 16 bit short results in a value of 19999.3896484375 cannam@85: (0.6103515625 * 0x7FFF) which then gets rounded down to 19999. cannam@85:

cannam@85:

cannam@85: You will notice that for this particular case, the error is 1 in 20000 or cannam@85: 0.005%. cannam@85: Interestingly, for values of less than 16369, dividing by 0x8000 followed cannam@85: by multiplying by 0x7FFF and then rounding the result, gives back the cannam@85: original value. cannam@85: It turns out that as long as the host operating system supplies the 1999 ISO cannam@85: C Standard functions lrintf and lrint (or a replacement has cannam@85: been supplied) then the maximum possible error is 1 in 16369 or about 0.006%. cannam@85:

cannam@85:

cannam@85: Regardless of the size of the error, the reason why this is done is rather cannam@85: subtle. cannam@85:

cannam@85:

cannam@85: In a file containing 16 bit PCM samples, the values are restricted to the range cannam@85: [-32768, 32767] while we want floating point values in the range [-1.0, 1.0]. cannam@85: The only way to do this conversion is to do a floating point division by a value cannam@85: of 0x8000. cannam@85: Converting the other way, the only way to ensure that floating point values in cannam@85: the range [-1.0, 1.0] are within the valid range allowed by a 16 bit short is cannam@85: to multiply by 0x7FFF. cannam@85:

cannam@85:

cannam@85: Some people would say that this is a severe short-coming of libsndfile. cannam@85: I would counter that anybody who is constantly converting back and forth cannam@85: between 16 bit shorts and normalised floats is going to suffer other losses cannam@85: in audio quality that they should also be concerned about. cannam@85:

cannam@85:

cannam@85: Since this problem only occurs when converting between integer data on disk and cannam@85: normalized floats in the application, it can be avoided by using something cannam@85: other than normalized floats in the application. cannam@85: Alternatives to normalized floats are the short and int data cannam@85: types (ie using sf_read_short or sf_read_int) or using un-normalized floats cannam@85: (see cannam@85: cannam@85: SFC_SET_NORM_FLOAT). cannam@85:

cannam@85:

cannam@85: Another way to deal with this problem is to consider 16 bit short data as a cannam@85: final destination format only, not as an intermediate storage format. cannam@85: All intermediate data (ie which is going to be processed further) should be cannam@85: stored in floating point format which is supported by all of the most common cannam@85: file formats. cannam@85: If floating point files are considered too large (2 times the size of a 16 bit cannam@85: PCM file), it would also be possible to use 24 bit PCM as an intermediate cannam@85: storage format (and which is also supported by most common file types). cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q11 : I'm having problems with u-law encoded WAV files generated by cannam@85: libsndfile in Winamp. Why? cannam@85:

cannam@85: cannam@85:

cannam@85: This is actually a Winamp problem. cannam@85: The official Microsoft spec suggests that the 'fmt ' chunk should be 18 bytes. cannam@85: Unfortunately at least one of Microsoft's own applications (Sound Recorder on cannam@85: Win98 I believe) did not accept 18 bytes 'fmt ' chunks. cannam@85:

cannam@85:

cannam@85: Michael Lee did some experimenting and found that: cannam@85:

cannam@85:

cannam@85:     I have checked that Windows Media Player 9, QuickTime Player 6.4,
cannam@85:     RealOne Player 2.0 and GoldWave 5.06 can all play u-law files with
cannam@85:     16-byte or 18-byte 'fmt ' chunk. Only Winamp (2.91) and foobar2000
cannam@85:     are unable to play u-law files with 16-byte 'fmt ' chunk.
cannam@85:

cannam@85: cannam@85:

cannam@85: Even this is a very small sampling of all the players out there. cannam@85: For that reason it is probably not a good idea to change this now because there cannam@85: is the risk of breaking something that currently works. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

**Q12 : I'm looking at sf_read*. What are items? What are frames? cannam@85:**

cannam@85: cannam@85:

cannam@85: An itemtt> is a single sample of the data type you are reading; ie a cannam@85: single short value for sf_read_short or a single float cannam@85: for sf_read_float. cannam@85:

cannam@85: cannam@85: For a sound file with only one channel, a frame is the same as a item (ie a cannam@85: single sample) while for multi channel sound files, a single frame contains a cannam@85: single item for each channel. cannam@85:

cannam@85: cannam@85:

cannam@85: Here are two simple, correct examples, both of which are assumed to be working cannam@85: on a stereo file, first using items: cannam@85:

cannam@85: cannam@85:

cannam@85:         #define CHANNELS 2
cannam@85:         short data [CHANNELS * 100] ;
cannam@85:         sf_count items_read = sf_read_short (file, data, 200) ;
cannam@85:         assert (items_read == 200) ;
cannam@85:

cannam@85: cannam@85:

cannam@85: and now readng the exact same amount of data using frames: cannam@85:

cannam@85: cannam@85:

cannam@85:         #define CHANNELS 2
cannam@85:         short data [CHANNELS * 100] ;
cannam@85:         sf_count frames_read = sf_readf_short (file, data, 100) ;
cannam@85:         assert (frames_read == 100) ;
cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q13 : Why can't libsndfile open this Sound Designer II (SD2) file? cannam@85:

cannam@85: cannam@85:

cannam@85: This is somewhat complicated. cannam@85: First some background. cannam@85:

cannam@85: cannam@85:

cannam@85: SD2 files are native to the Apple Macintosh platform and use features of cannam@85: the Mac filesystem (file resource forks) to store the file's sample rate, cannam@85: number of channels, sample width and more. cannam@85: When you look at a file and its resource fork on Mac OS X it looks like cannam@85: this: cannam@85:

cannam@85: cannam@85:

cannam@85:         -rw-r--r--  1 erikd erikd   46512 Oct 18 22:57 file.sd2
cannam@85:         -rw-r--r--  1 erikd erikd     538 Oct 18 22:57 file.sd2/rsrc
cannam@85:

cannam@85: cannam@85:

cannam@85: Notice how the file itself looks like a directory containing a single file cannam@85: named rsrc. cannam@85: When libsndfile is compiled for MacOS X, it should open (for write and read) cannam@85: SD2 file with resource forks like this without any problems. cannam@85: It will also handle files with the resource fork in a separate file as cannam@85: described below. cannam@85:

cannam@85: cannam@85:

cannam@85: When SD2 files are moved to other platforms, the resource fork of the file cannam@85: can sometimes be dropped altogether. cannam@85: All that remains is the raw audio data and no information about the number cannam@85: of channels, sample rate or bit width which makes it a little difficult for cannam@85: libsndfile to open the file. cannam@85:

cannam@85: cannam@85:

cannam@85: However, it is possible to safely move an SD2 file to a Linux or Windows cannam@85: machine. cannam@85: For instance, when an SD2 file is copied from inside MacOS X to a windows cannam@85: shared directory or a Samba share (ie Linux), MacOS X is clever enough to cannam@85: store the resource fork of the file in a separate hidden file in the cannam@85: same directory like this: cannam@85:

cannam@85:

cannam@85:         -rw-r--r--  1 erikd erikd     538 Oct 18 22:57 ._file.sd2
cannam@85:         -rw-r--r--  1 erikd erikd   46512 Oct 18 22:57 file.sd2
cannam@85:

cannam@85: cannam@85:

cannam@85: Regardless of what platform it is running on, when libsndfile is asked to cannam@85: open a file named "foo" and it can't recognize the file type from cannam@85: the data in the file, it will attempt to open the resource fork and if cannam@85: that fails, it then tries to open a file named "._foo" to see if cannam@85: the file has a valid resource fork. cannam@85: This is the same regardless of whether the file is being opened for read cannam@85: or write. cannam@85:

cannam@85: cannam@85:

cannam@85: In short, libsndfile should open SD2 files with a valid resource fork on cannam@85: all of the platforms that libsndfile supports. cannam@85: If a file has lost its resource fork, the only option is the open the file cannam@85: using the SF_FORMAT_RAW option and guessing its sample rate, channel count cannam@85: and bit width. cannam@85:

cannam@85: cannam@85:

cannam@85: Occasionally, when SD2 files are moved to other systems, the file is cannam@85: BinHexed cannam@85: which wraps the resource fork and the data fork together. cannam@85: For these files, it would be possible to write a BinHex parser but cannam@85: there is not a lot to gain considering how rare these BinHexed SD2 cannam@85: files are. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q14 : I'd like to statically link libsndfile to my closed source cannam@85: application. Can I buy a license so that this is possible? cannam@85:

cannam@85: cannam@85:

cannam@85: Unfortunately no. cannam@85: libsndfile contains code written by other people who have agreed that their cannam@85: code be used under the GNU LGPL but no more. cannam@85: Even if they were to agree, there would be significant difficulties in cannam@85: dividing up the payments fairly. cannam@85:

cannam@85: cannam@85:

cannam@85: The only way you can legally use libsndfile as a statically linked cannam@85: library is if your application is released under the GNU GPL or LGPL. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q15 : My program is crashing during a call to a function in libsndfile. cannam@85: Is this a bug in libsndfile? cannam@85:

cannam@85: cannam@85:

cannam@85: libsndfile is being used by large numbers of people all over the world cannam@85: without any problems like this. That means that it is much more likely cannam@85: that your code has a bug than libsndfile. However, it is still possible cannam@85: that there is a bug in libsndfile. cannam@85:

cannam@85:

cannam@85: To figure out whether it is your code or libsndfile you should do the cannam@85: following: cannam@85:

cannam@85:

Make sure you are compiling your code with warnings switched on and cannam@85: that you fix as many warnings as possible. cannam@85: With the GNU compiler (gcc) I would recommend at least cannam@85: -W -Wall -Werror which will force you to fix all warnings cannam@85: before you can run the code. cannam@85:
Try using a memory debugger. cannam@85: Valgrind on x86 Linux is excellent. cannam@85: Purify also cannam@85: has a good reputation. cannam@85:
If the code is clean after the above two steps and you still get cannam@85: a crash in libsndfile, then send me a small snippet of code (no cannam@85: more than 30-40 lines) which includes the call to sf_open() and cannam@85: also shows how all variables passed to/returned from sf_open() cannam@85: are defined. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q16 : Will you accept a fix for compiling libsndfile with compiler X? cannam@85:

cannam@85: cannam@85:

cannam@85: If compiler X is a C++ compiler then no. cannam@85: C and C++ are different enough to make writing code that compiles as valid C cannam@85: and valid C++ too difficult. cannam@85: I would rather spend my time fixing bugs and adding features. cannam@85:

cannam@85: cannam@85:

cannam@85: If compiler X is a C compiler then I will do what I can as long as that does cannam@85: not hamper the correctness, portability and maintainability of the existing cannam@85: code. cannam@85: It should be noted however that libsndfile uses features specified by the 1999 cannam@85: ISO C Standard. cannam@85: This can make compiling libsndfile with some older compilers difficult. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q17 : Can libsndfile read/write files from/to UNIX pipes? cannam@85:

cannam@85: cannam@85:

cannam@85: Yes, libsndfile can read files from pipes. cannam@85: Unfortunately, the write case is much more complicated. cannam@85:

cannam@85: cannam@85:

cannam@85: File formats like AIFF and WAV have information at the start of the file (the cannam@85: file header) which states the length of the file, the number of sample frames cannam@85: etc. cannam@85: This information must be filled in correctly when the file header is written, cannam@85: but this information is not reliably known until the file is closed. cannam@85: This means that libsndfile cannot write AIFF, WAV and many other file types cannam@85: to a pipe. cannam@85:

cannam@85: cannam@85:

cannam@85: However, there is at least one file format (AU) which is specifically designed cannam@85: to be written to a pipe. cannam@85: Like AIFF and WAV, AU has a header with a sample frames field, but it is cannam@85: specifically allowable to set that frames field to 0x7FFFFFFF if the file cannam@85: length is not known when the header is written. cannam@85: The AU file format can also hold data in many of the standard formats (ie cannam@85: SF_FORMAT_PCM_16, SF_FORMAT_PCM_24, SF_FORMAT_FLOAT etc) as well as allowing cannam@85: data in both big and little endian format. cannam@85:

cannam@85: cannam@85:

cannam@85: See also FAQ Q6. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q18 : Is it possible to build a Universal Binary on Mac OS X? cannam@85:

cannam@85: cannam@85:

cannam@85: Yes, but you must do two separate configure/build/test runs; one on PowerPC cannam@85: and one on Intel. cannam@85: It is then possible to merge the binaries into a single universal binary using cannam@85: one of the programs in the Apple tool chain. cannam@85:

cannam@85: cannam@85:

cannam@85: It is not possible to build a working universal binary via a single cannam@85: compile/build run on a single CPU. cannam@85:

cannam@85: cannam@85:

cannam@85: The problem is that the libsndfile build process detects features of the CPU its cannam@85: being built for during the configure process and when building a universal binary, cannam@85: configure is only run once and that data is then used for both CPUs. cannam@85: That configure data will be wrong for one of those CPUs. cannam@85: You will still be able to compile libsndfile, and the test suite will pass on cannam@85: the machine you compiled it on. cannam@85: However, if you take the universal binary test suite programs compiled on one cannam@85: CPU and run them on the other, the test suite will fail. cannam@85:

cannam@85: cannam@85:

cannam@85: Part of the problem is the the CPU endian-ness is detected at configure time. cannam@85: Yes, I know the Apple compiler defines one of the macros __LITTLE_ENDIAN__ cannam@85: and __BIG_ENDIAN__, but those macros are not part of the 1999 ISO C Standard cannam@85: and they are not portable. cannam@85:

cannam@85: cannam@85:

cannam@85: Endian issues are not the only reason why the cross compiled binary will fail. cannam@85: The configure script also detects other CPU specific idiosyncrasies to provide cannam@85: more optimized code. cannam@85:

cannam@85: cannam@85:

cannam@85: Finally, the real show stopper problem with universal binaries is the problem cannam@85: with the test suite. cannam@85: libsndfile contains a huge, comprehensive test suite. cannam@85: When you compile a universal binary and run the test suite, you only test the cannam@85: native compile. cannam@85: The cross compiled binary (the one with the much higher chance of having cannam@85: problems) cannot be tested. cannam@85:

cannam@85: cannam@85:

cannam@85: Now, if you have read this far you're probably thinking there must be a way cannam@85: to fix this and there probably is. cannam@85: The problem is that its a hell of a lot of work and would require significant cannam@85: changes to the configure process, the internal code and the test suite. cannam@85: In addition, these changes must not break compilation on any of the platforms cannam@85: libsndfile is currently working on. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q19 : I have project files for Visual Studio / XCode / Whatever. Why cannam@85: don't you distribute them with libsndfile? cannam@85:

cannam@85: cannam@85:

cannam@85: There's a very good reason for this. cannam@85: I will only distribute things that I actually have an ability to test and cannam@85: maintain. cannam@85: Project files for a bunch of different compilers and Integrated Development cannam@85: Environments are simply too difficult to maintain. cannam@85:

cannam@85: cannam@85:

cannam@85: The problem is that every time I add a new file to libsndfile or rename an cannam@85: existing file I would have to modify all the project files and then test that cannam@85: libsndfile still built with all the different compilers. cannam@85:

cannam@85: cannam@85:

cannam@85: Maintaining these project files is also rather difficult if I don't have access cannam@85: to the required compiler/IDE. cannam@85: If I just edit the project files without testing them I will almost certainly cannam@85: get it wrong. cannam@85: If I release a version of libsndfile with broken project files, I'll get a bunch cannam@85: of emails from people complaining about it not building and have no way of cannam@85: fixing or even testing it. cannam@85:

cannam@85: cannam@85:

cannam@85: I currently release sources that I personally test on Win32, Linux and cannam@85: MacOS X (PowerPC) using the compiler I trust (GNU GCC). cannam@85: Supporting one compiler on three (actually much more because GCC is available cannam@85: almost everywhere) platforms is doable without too much pain. cannam@85: I also release binaries for Win32 with instructions on how to use those cannam@85: binaries with Visual Studio. cannam@85: As a guy who is mainly interested in Linux, I'm not to keen to jump through cannam@85: a bunch of hoops to support compilers and operating systems I don't use. cannam@85:

cannam@85: cannam@85:

cannam@85: So, I hear you want to volunteer to maintain the project files for Some Crappy cannam@85: Compiler 2007? cannam@85: Well sorry, that won't work either. cannam@85: I have had numerous people over the years offer to maintaining the project cannam@85: files for Microsoft's Visual Studio. cannam@85: Every single time that happened, they maintained it for a release or two and cannam@85: then disappeared off the face of the earth. cannam@85: Hence, I'm not willing to enter into an arrangement like that again. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q20 : Why doesn't libsndfile support MP3? Lots of other Open Source cannam@85: projects support it! cannam@85:

cannam@85: cannam@85:

cannam@85: MP3 is not supported for one very good reason; doing so requires the payment cannam@85: of licensing fees. cannam@85: As can be seen from cannam@85: cannam@85: mp3licensing.com cannam@85: the required royalty payments are not cheap. cannam@85:

cannam@85: cannam@85:

cannam@85: Yes, I know other libraries ignore the licensing requirements, but their legal cannam@85: status is extremely dubious. cannam@85: At any time, the body selling the licenses could go after the authors of those cannam@85: libraries. cannam@85: Some of those authors may be students and hence wouldn't be worth pursuing. cannam@85:

cannam@85: cannam@85:

cannam@85: However, libsndfile is released under the name of a company, Mega Nerd Pty Ltd; cannam@85: a company which has income from from libsamplerate licensing, libsndfile based cannam@85: consulting income and other unrelated consulting income. cannam@85: Adding MP3 support to libsndfile could place that income would be under legal cannam@85: threat. cannam@85:

cannam@85: cannam@85:

cannam@85: Fortunately, Ogg Vorbis exists as an alternative to MP3. cannam@85: Support for Ogg Vorbis was added to libsndfile (mostly due to the efforts of cannam@85: John ffitch of the Csound project) in version 1.0.18. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85: cannam@85:

Q21 : How do I use libsndfile in a closed source or commercial program cannam@85: and comply with the license? cannam@85:

cannam@85: cannam@85:

cannam@85: Here is a checklist of things you need to do to make sure your use of libsndfile cannam@85: in a closed source or commercial project complies with the license libsndfile is cannam@85: released under, the GNU Lesser General Public License (LGPL): cannam@85:

cannam@85: cannam@85:

Make sure you are linking to libsndfile as a shared library (Linux and Unix cannam@85: systems), Dynamic Link Library (Microsoft Windows) or dynlib (Mac OS X). cannam@85: If you are using some other operating system that doesn't allow dynamically cannam@85: linked libraries, you will not be able to use libsndfile unless you release cannam@85: the source code to your program. cannam@85:
In the licensing documentation for your program, add a statement that your cannam@85: software depends on libsndfile and that libsndfile is released under the GNU cannam@85: Lesser General Public License, either cannam@85: version 2.1 cannam@85: or optionally cannam@85: version 3. cannam@85:
Include the text for both versions of the license, possibly as separate cannam@85: files named libsndfile_lgpl_v2_1.txt and libsndfile_lgpl_v3.txt. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q22 : What versions of Windows does libsndfile work on? cannam@85:

cannam@85: cannam@85:

cannam@85: Currently the precompiled windows binaries are thoroughly tested on Windows XP. cannam@85: As such, they should also work on Win2k and Windows Vista. cannam@85: They may also work on earlier versions of Windows. cannam@85:

cannam@85: cannam@85:

cannam@85: Since version 0.1.18 I have also been releasing precompiled binaries for Win64, cannam@85: the 64 bit version of Windows. cannam@85: These binaries have received much less testing than the 32 bit versions, but cannam@85: should work as expected. cannam@85: I'd be very interested in receiving feedback on these binaries. cannam@85:

cannam@85: cannam@85: cannam@85: cannam@85:

Q23 : I'm cross compiling libsndfile for another platform. How can I cannam@85: run the test suite? cannam@85:

cannam@85: cannam@85:

cannam@85:

cannam@85: cannam@85:

cannam@85: Since version 1.0.21 the top level Makefile has an extra make target, cannam@85: 'test-tarball'. cannam@85: Building this target creates a tarball called called: cannam@85:

cannam@85: cannam@85: cannam@85: libsndfile-testsuite-${host_triplet}-${version}.tar.gz cannam@85: cannam@85: cannam@85:

cannam@85: in the top level directory. cannam@85: This tarball can then be copied to the target platform. cannam@85: Once untarred and test script test_wrapper.sh can be run from cannam@85: the top level of the extracted tarball. cannam@85:

cannam@85: cannam@85: cannam@85:

cannam@85:

cannam@85: The libsndfile home page is here : cannam@85: cannam@85: http://www.mega-nerd.com/libsndfile/. cannam@85:
cannam@85: Version : 1.0.25 cannam@85:

cannam@85: cannam@85: cannam@85: