I had a nice day and now I'm going to compile some Linux MIDI stuff I need, but this misinformation that's common in the Linux community still makes me sad, yes sad, I'm not angry, e.g.
"Integer representation [snip] - Ignores other feasible representations." computers only use bits, a bit can be set or cleared, the interpretation of the bits can be an integer or a floating point or a char or what ever, the key are the number of bits. For 'normal' MIDI data there are 128 steps or 128 * 128 = 16384 steps, if those steps aren't enough, there are other possibilities, e.g. the usage of controller combinations or very short SysEx commands. For example, it's common to use Pitch Bend over a range of 2 semi tones, for this you have got 16384 steps just by the normal MIDI standard, without using SysEx or any other trick. Time code is in 1/16 notes for the time code, not for the quantisation, the quantisation can fit to the baud rate and that resolution is better than any Linux application seems to be able to do internal the computer, because of the jitter that Linux has got, even on better machines than my machine. MIDI is a bad, slow, limited standard, but I don't think any audio Linux application really reaches to this standard in common practise. MIDI is not as bad as the rumours say. It's much over the standard of what Linux is able to do. It would be a big step to clear out problems that Linux has got, instead of insist on wrong knowledge. So, I won't add anything to this issue, instead I will work on my Linux, what maybe will help other Linux users too. :), Ralf _______________________________________________ 64studio-users mailing list [email protected] http://lists.64studio.com/mailman/listinfo/64studio-users
