Hello, Is it possible or necessary to quantize simultaneously outgoing MIDI signals of 'n' [ctlout X X] objects. Question aroused from the simple sort of problems: My patch is a medium between audio (-rt) and "windows" (OS) VJ software. Generated movements of sliders are smooth and constantly respond to sound, that means always traffic through MIDI output. Eight simultaneous signals causes a VJ software to slow down. I added abstraction that in a speed of 3ms sends data of only one [ctlout], what solved a problem for a while. Rising a quantity of [ctlout]'s anyway causes me to see steps in a movements of sliders on "pd" patch and VJ software. Is there a solution I'm not yet familiar with, to use midi protocol, or it might be the problem of "windows", VJ software or MIDI protocol?
pd-vanilla. Ubuntu 9.10
Thank you Gintaras
hi.
Gintaras Lau. wrote:
Hello, Is it possible or necessary to quantize simultaneously outgoing MIDI signals of 'n' [ctlout X X] objects.
hmm, i'm not sure whether i understand your problem.
anyhow, three generic hints:
at [speedlim] (part of iemlib(2) and most likely cyclone)
(which is what comes to my mind when i hear "quantize"), do something like y=(int(x*k))/k for normalized values
fdmasr IOhannes
Yes, seems like [change] is one of the supper heroes here and will appear in my patch immediately. I have tried [speedlim] a little ago and it's not so useful in my case this time. Could you write a short description on your third suggestion with a formula for normalized values. I know a little how [expr] and mathematics works in 'pd', but how to use that mathematical expression to [ctlout] data?
Thanks Gintaras
2010/1/25 IOhannes m zmoelnig zmoelnig@iem.at:
hi.
Gintaras Lau. wrote:
Hello, Is it possible or necessary to quantize simultaneously outgoing MIDI signals of 'n' [ctlout X X] objects.
hmm, i'm not sure whether i understand your problem.
anyhow, three generic hints:
- if you want to avoid sending duplicate values, have a look at [change]
- if you want to control the maximum data rate of messages, have a look
at [speedlim] (part of iemlib(2) and most likely cyclone)
- if you want to limit the different values that can possibly occur
(which is what comes to my mind when i hear "quantize"), do something like y=(int(x*k))/k for normalized values
fdmasr IOhannes