Midi Merge Technique?

[Magnus Danielson]

From: "Batz Goodfortune" <batzman at all-electric.com>
Subject: Re: Midi Merge Technique?
Date: Sun, 23 Jan 2000 13:07:48 +1000
> 
> This can't be too bad. Certainly not in a practical sense. And most likely
> all software would do it to a certain extent. Anecdotally at least, I've
> used CW since version 1-dos and never had a sync problem. And I use
> extensive syncing. Both internally and externally. What I did have problem
> with was one of the afor mentioned KMS30 Korg sync boxes. A very curious
> thing happens at exactly 115BPM. The sync just flakes out all over the
> place. We couldn't work out what was going on but we soon leant never to
> stripe a tape at 115PBM. 116 was fine. 114 was fine but @115 it went crazy.

Hmm... strange...

> And of course there is Jitter and there's Jitter. If you're talking about
> the jitter  caused by the UART this is another thing altogether. On the
> other hand, the most common cause of sync jitter in a wide sense is BOSH95
> it self. It is why sequencer software, until very recently has mostly been
> written using 16 bit code. The preemptive multi-tasking under BOSH95 causes
> real-time  tasks to flake badly. They have apparently fixed this problem in
> BOSH98 where software can specify it's priority. But! It's still a long way
> from the claimed real-time streaming which was promised in BOSH95 and still
> not delivered in BOSH98. No-one to my knowledge actually uses this because
> it still doesn't work. WINK might be different but I'm certainly not game
> enough to try it. From all accounts it's a useless bucket of crap.

If your timing depend on scheduling and transfer on a OS like that (there are
more of them) then you should either consider to shoot your computer to stop
the suffering or change OS. If any timing critical stuff goes into the main
CPU you must use both an OS and application designed to handle these real time
requirements.

> Anyway, it would seem that a little jitter never hurt anyone. ????

Don't say that so any telecom people hear you. For instance, do you realize
what jitter requirements that just the telephone system have? A T1 or E1 is
not happy about jitter and then there is the SDH/SONET stuff.
These guys look at how jitter builds up as you go through diffrent types of
regenerators and also how wander depend on cable temperature etc.

Interestingly enougth, when they where developing the T1/DS1 system and was
investigating the jitter accumulation they used the same device over and over
hooked to a tape recorder. They simply recorded the output signal and then
replayed that for the next run and recorded that output. This way they could
get many regenerations from one box. Naturally this was an analog tape
recorder (or else it would not work).

Jitter will translate into bit errors in digital systems. The bitshuffling
guys are pretty picky about these details.

Cheers,
Magnus