This is gavl, a library for handling uncompressed audio- and video data.

Read the file INSTALL for detailed installation instructions.

Gavl defines generic types for audio and video formats,
which are applicable to a wide range of multimedia applications.
In addition, it provides conversion functions from all possible formats
to all other formats.

If you want to write a multimedia application, gavl will help you a lot:
You don't have to mess around with the large varieties of pixelformats,
interleaving modes, audio sample formats and so on.

Current features:

- Support for the following pixelformats:
  15 bpp RGB, 15 bpp BGR, 16 bpp RGB, 16 bpp BGR, 24 bpp RGB, 24 bpp BGR
  32 bpp RGB, 32 bpp BGR, 32 bpp RGBA, YUV 422 (YUY2), YUV 420 Planar,
  YUV 422 Planar (mpeg and jpeg), YUV 444 Planar (mpeg and jpeg)
  Conversions from all pixelformat to all others (NO multistep conversions,
  each conversion has it's own function)

- Audio samples can be 8 bit and 16 PCM (signed and unsigned),
  32 bit signed and floating point.

- High quality samplerate conversion using the Secret Rabbit Code

- Video frames will allocated with memory aligned scanlines for faster
  processing.

See the header file include/gavl.h and the sample files
src/colorspace_test.c and src/colorspace_time.c for
information and how to use this thing.

For testing the library, the src directory contains
some test programs.

colorspace_test:
Makes your current directory full of pngs, one test file
for each colorspace and one for each conversion.
Ideally, all these pictures should look the same.

colorspace_time:
Makes all defined conversions 20 times and prints, how long
it needed. Good for making C <-> MMX speed comparisons.

Every programmer, who agrees to this kind of API is welcome to
integrate gavl into his/her project so we can save some
energy while writing/searching the best and fastest
conversion routines.