Sound Programming in Linux

1. Definitions
2. Audio Subsystem
3. OSS/ALSA kernel drivers
4. Sound Libraries
5. File formats
6. Sampling an Effect

1. Definitions

ADC	Analog to Digital Converter. Takes voltage from microphone and converts it into PCM data
DAC	Digital to Analog Converter. Takes PCM data and converts the values into a voltage
PCM	Pulse Code Modulation. A way of describing audio as a set of integer values corresponding to points on a sound wave
SDL	Simple DirectMedia Layer. Graphics and sound library

2. Audio Subsystem

Playing an audio file

Recording a sound

3. OSS/ALSA kernel drivers

Opensound by 4Front Technologies

Already in kernel sources
Commercial
Well documented API (compared to that of ALSA)

Advanced Linux Sound Architecture

Installation is outside kernel
FREE
Support for OSS applications
Documentation could use some work

4. Sound Libraries

YIFF Soundserver
Skysound (Example)
SDL (Example)
libMikMod (Example)

5. File Formats
- WAV
  - SDL,libMikMod,YIFF
  - Easy to manipulate
  - Computationally insignificant
  - filesize can be rather large (10-20MB background music) (1-20k effect)
- MOD,S3M...
  - libMikMod
  - Small file size (300-700kb background music)
  - Computationally insignificant
  - MP3
    - SDL (with SMPEG), SkySound
  - Some overhead
  - Small file size (3-4MB background music) (1-20k effect)
  6. Recording an Effect
  
  When recording an effect it's best to set the software to record at 44100 samples per second and 16 bit resolution. You can always convert it down later.
  
  WAV files are the best to use because.... that is what is used by linux recording software.
  
  During looped playback of a WAV file an annoying click can be heard at the end of each repetition. This is due to the fact that the first and last value for the WAV file are not zero. Make sure those values are zero.