Yeah, thats right, but put it in an external lib. I think the word
"Simple" in “Simple Directmedia Layer” is very important. Keep it as
"Simple" as it could be. Keyboard input and mouse input you need every-
day but not AUDIO Input.
Well, I do not agree at this point. As we all see, software gains more
interactivity, by giving the user different options to do his input. I could
think of AUDIO input in many situations, e.g. voice controll (as one of the
most complex), just voice over net (as a medium-comlicated) and voice-input
for (like mentioned in this list before) a players name.
This means, the AUDIO-Input functions SDL provides should be as basic, as
its video functions. And this brings me to the point, what i intended to ask
in the questions, with wich I opened this thread:
This Audio-Input could be based on a structure, such as SDL_AudioStream.
This structure keeps some basic audio-data and information about volume and
the audio-curve. The developer decides, when to open the stream and when to
close it (so no problems with duplex, etc… for sure we need a
SDL_AudioHardwareInfo here, so the developer can decide, what to do).
By giving the reading of the stream an own thread, for example, the
developer can decide about the number of audio-information per secound
(maybe the timing becomes a problem here), he himself decides about sample
rate and stuff. The same could be (in my oppinion) done with
SDL_VideoStream: The developer is enabled to catch a still-video-image from
any video input device as a sdl-surface, not more, not less.
Everything “over” those described functions was not intended with my
questions. So just enable the developer to “easy-access” audio-data, not
handle that data for him. No compression, decompression functions should be
in this, no handling at all, just the basic information a “tune” has at a
time: it’s frequency and it’s volume.
Regards,
Sascha