code module timer (partially working)

infuzion · 13 Jan 2006

Couldn't find the answer, but did find the following: 1) Prob could help in the 2nd example you don't connect the sample rate to monoin run 😉 2) Tip: you don't need the Int Switcher any more; there is a Not primitive. 3) I got a crash after closing after a bit of editing... anyone else? 4) {edit} 1st one seems to stop at 6533?

NothanUmber · 13 Jan 2006

Thanks for having a look!

infuzion wrote: Couldn't find the answer, but did find the following: 1) Prob could help in the 2nd example you don't connect the sample rate to monoin run 😉

Hm, indeed... 😊 is still not working but it's certainly a step into the right direction 🙂

infuzion wrote: 2) Tip: you don't need the Int Switcher any more; there is a Not primitive.

it was late 😉

infuzion wrote: 3) I got a crash after closing after a bit of editing... anyone else?

yes, had also some crashs, seems somehow to be related to the combination of the tthread component with the code/audio out component. (see below)

infuzion wrote: 4) {edit} 1st one seems to stop at 6533?

not reproducable here, doesn't stop

Here a new version of the qual. timer (the one I'm actually using now in my schematic, having a working sample precise timer would of course be even nicer 😉 ). This new version doesn't require the code component and the audio out and thus is much more stable (has a slightly other scaling than the old one: it's much faster for lower delay values and far slower for bigger ones.)

Greetings, NothanUmber

NothanUmber · 13 Jan 2006

The new variant is also sometimes crashing SM (when closing the schematic) so perhaps just one tthread is enough for unwanted effects... (the exported vst effect that contains the same timer seems to work flawlessly though up to now so it's not that bad.)

Greetings, NothanUmber

infuzion · 14 Jan 2006

Cool, on v3: 1) The 1st & 3rd timers give way different values; 3rd much more higher. 2) The TThread in 3rd isn't needed; I think primitives that have a trigger out & in have a more selective anti-feedback looping system (I think fixed 1-3 versions ago). 3) You should really try not to use bending links in your schematics; they (at this time) get saved w/the VST, plus the new version really saves memory on straight vs curved links. 4) When running, or run then stop this schematic creates a huge amount of page faults (prob not your fault, but just documenting it) 5) Now will crash automaticly before closing.

NothanUmber · 14 Jan 2006

infuzion wrote: Cool, on v3: 1) The 1st & 3rd timers give way different values; 3rd much more higher.

for smaller delay values the new version is faster, for bigger it's slower. As both timers are only qualitative that's probably something one can live with.

infuzion wrote: 2) The TThread in 3rd isn't needed; I think primitives that have a trigger out & in have a more selective anti-feedback looping system (I think fixed 1-3 versions ago).

Cool! Haven't been able to get this to work though. (If I connect the BSwitchTrig output to the start input of the loop in this schematic (removed the other timers) then SM crashes immediately, the loop doesn't seem to be particularly fond of being used in a direct feedback line)

infuzion wrote: 3) You should really try not to use bending links in your schematics; they (at this time) get saved w/the VST, plus the new version really saves memory on straight vs curved links.

Thanks for the tip, wasn't aware of that. (Will straighten the links at least when releasing things, looks kind of messy though)

infuzion wrote: 4) When running, or run then stop this schematic creates a huge amount of page faults (prob not your fault, but just documenting it)

Ugh, not nice 😕 (Don't have the slightest clue what's wrong here (at least when it's also the case without the tthread component), that's a case for the devs)

infuzion wrote: 5) Now will crash automaticly before closing.

Will even crash while still connecting the wires 🙂 (see above)

Greetings, NothanUmber

NothanUmber · 14 Jan 2006

Did further experiments and made yet another qual. timer that's not as unneccessarily complicated as the old one (no loop required, some thumb double negations eliminated etc.). The page fault problem as well as the crash when closing the schematic while the timer is still running are notheless an issue (get ~ 7000 p-faults per second in comparison to ~ 60 when the timer isn't running) so there seems to be something wrong with the tthread. (Eliminating the tthread in that schematic isn't possible I guess as the counter component is increased only once in a direct feedback chain.)

Greetings, NothanUmber