| Title: | * * Computer Music, MIDI, and Related Topics * * |
| Notice: | Conference has been write-locked. Use new version. |
| Moderator: | DYPSS1::SCHAFER |
| Created: | Thu Feb 20 1986 |
| Last Modified: | Mon Aug 29 1994 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2852 |
| Total number of notes: | 33157 |
Hi. If the MIDI baud rate is about 31K, that's 3.1K bytes/second. If I use running status, I can send 1550 note events per second. That's 775 whole note-on / note-on cycles per second. Is there any synthesizer that can respond to a rate anywhere near that? My program AlgoRhythms now send about 300 notes/second (that's on/off pairs, or 600 note events), using running status. The TX gave MIDI Buffer Full messages a long time ago at lower rates. The MKS20 doesn't report one way or the other, so I'm not sure how far behind it is. . Scoping the output of the MKS may not pay; if notes hit together, they won't be distinguishable events. Anyway, does any synth maker specify the max rate of played events? Tom
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2768.1 | Don't Try This At Home, Kids | RGB::ROST | All American Alien Boy | Tue Nov 12 1991 11:45 | 13 |
I have never seen a spec for the maximum rate of MIDI events. Keyboard
magazine has had a habit for awhile of running "torture tests" which
attempt to foul up an SGU with complex multipart data. They seem to
usually make *everything* choke eventually. Welcome to the computer
age. Of course, such tests are pretty subjective.
Examining tempo vs notes per second, at the dreaded disco tempo of 120
bpm, that's only two beats per second. 300 notes per second would be
like what, 128th notes? With some more reasonable 32nd notes, you'd be
cranking close to 10 beats per second or 600 bpm. Even Bad Brains
doesn't play that fast 8^) 8^)
Brian
| |||||
| 2768.2 | Synthetic virtuouso | PIANST::JANZEN | Love looks not with the eyes | Tue Nov 12 1991 12:47 | 22 |
There's this 30-year old concept of the "virtuousity" of computer music, playing hundreds or thousands of separate events per second. To what purpose? For one, it makes it possible to explore very special textures that are hard to make any other way. You have to hear it. Over the weekend I played 300 notes/sec in a diatonic scale, then while it was playing, transposed it up a major second. You could definitely hear it rise. It's kind of cool. Even Excellent cool. If you split it with a delay into stereo (right delayed, left not delayed) it sounds stereo, not like a canon from left to right (same as white noise delayed that way). For stats, AlgoRhythms plays 65 notes/sec in one voice, and 290 in 16 voices, and it seems linear with respect to number of voices playing, but I didn't carefuly check. Let's say you have 16 synths, each with a split for 32 different sounds (or alternating or something) and polyphonal. On the average AlgoRhythms could play 290/32 or 9 notes per second in each instrument, not so fast after all. I need more speed! (AlgoRhythms can assign MIDI channel to each voice any way you like.) Tom | |||||
| 2768.3 | It's 1940 ! | SALSA::MOELLER | Karl has...left the building | Wed Nov 13 1991 11:16 | 13 |
Splitting the MIDI data stream and adding SGU's is one tactic. The
MOTU Midi Time Piece uses RS422 in a high speed mode (I seem to
remember like 500Kb/sec) between it and the MAC. Then each SGU has its
own MIDI cable. So one could just stack up the SGU's, each with its
own unshared MIDI cable.
Or one could back off from the direct SGU control issue. *IF* one
could find slash create a reasonable MIDI stream - to - piano roll
punch unit (they MUST exist) then you could soup up a player piano
and have at it. The data rate would only be dependent on the response
of the piano's action - and you'd have 88-note polyphony.
karl
| |||||