karl krach

i have a problem which is pretty tough to get one's head around (or maybe it's just me?) - and it's hard to explain too.

have a look at the small patch below. this is a video delay (based on a patch by andrew benson) where i can control the delay by providing the absolute number of frames to delay. now i want this delay to work in waves, like smoothly going to slowmo, then getting fast again, then slowmo etc.. to achieve this i currently base the number of frames to delay on a scaled sine curve.

to control the output i can tweak two things:

- the length of one cycle: presence to presence (= the speed of the wave)

- and the maximum frames to go back (= amplitude of the wave).

my problem now is, that depending on these variables, the output sometimes goes back in time (which i do not want). or it does not let me get back in time very far. or put more generally: i want more control over this process.

- length of one cycle: presence to presence (= the speed of the wave)

to achieve this, i guess this sine wave would need to change into some other function where the positive and negative amplitudes would have different heights and also different speeds...?

many thanks for any help with this!! i'm afraid my head will explode soon...

hi,
i have a problem which is pretty tough to get one's head around (or maybe it's just me?) - and it's hard to explain too.

have a look at the small patch below. this is a video delay (based on a patch by andrew benson) where i can control the delay by providing the absolute number of frames to delay. now i want this delay to work in waves, like smoothly going to slowmo, then getting fast again, then slowmo etc.. to achieve this i currently base the number of frames to delay on a scaled sine curve.
to control the output i can tweak two things:
- the length of one cycle: presence to presence (= the speed of the wave)
- and the maximum frames to go back (= amplitude of the wave).

my problem now is, that depending on these variables, the output sometimes goes back in time (which i do not want). or it does not let me get back in time very far. or put more generally: i want more control over this process.
ideally i would have these controls:
- length of one cycle: presence to presence (= the speed of the wave)
- slowest speed (in fps or multiples)
- and fastest speed.

to achieve this, i guess this sine wave would need to change into some other function where the positive and negative amplitudes would have different heights and also different speeds...?
how would i go on about this???

many thanks for any help with this!! i'm afraid my head will explode soon...
karl


translating-absolute-frame-numbers-to-playspeed

i also tried another approch where i used two counters and based the speed control on the speed of bangs - one counter for writing to the matrixset, one for reading it out. this approach gives me "playspeed control" but unfortunately it turned out to be almost impossible to keep these two counters properly synced...

isn't there any other way to do this? i mean, the base problem really sounds pretty simple: have a video buffer and then oscillate smoothly between slowmo and highspeed playback in a _controlled_ way...

still stuck.
i also tried another approch where i used two counters and based the speed control on the speed of bangs - one counter for writing to the matrixset, one for reading it out. this approach gives me "playspeed control" but unfortunately it turned out to be almost impossible to keep these two counters properly synced...

thank you very much for helping me with this!

i am afraid i do not understand your patch though - if i disable "save data with patcher" on the itable, i can't recreate this curve already drawn in the itable... also, i tried to put the logic into my patcher above but did not manage to figure out how to control anything with that... how would i use that?

but ok, maybe i should take two steps back and move only one foward at a time. also, i remembered i ran into the basic problem (convert absolute framenumbers to playspeed) before - back then i solved this with an external, the f0 ultimate counter - basically just a counter but which can do floats and also vary the increments to count. unfortunately this external does not exist on win64bit, so i need to recreate this somehow. have a look at the patch below: how would i get this delay to play back at a constant 0.5x or 3x speed?

hi mzed,
thank you very much for helping me with this!
i am afraid i do not understand your patch though - if i disable "save data with patcher" on the itable, i can't recreate this curve already drawn in the itable... also, i tried to put the logic into my patcher above but did not manage to figure out how to control anything with that... how would i use that?

p.s.: the second step would then be to oscillate the playspeeds smoothly up and down...

however, if someone as another idea how to do this (eg. not involving counters but with a function or whatever): please, please step forward!!

p.s.: the second step would then be to oscillate the playspeeds smoothly up and down...
however, if someone as another idea how to do this (eg. not involving counters but with a function or whatever): please, please step forward!!
k


Just send a bang into the Uzi on my patch and it redraws the curve.

strange, i tried this and just checked it again but it does not do on my side. banging the uzi results in 240x an output of "10.000" at the "+ 0." and no curve.

anyway, did i now manage to relate my problem better? do you understand what i am after?

strange, i tried this and just checked it again but it does not do on my side. banging the uzi results in 240x an output of "10.000" at the "+ 0." and no curve.
anyway, did i now manage to relate my problem better? do you understand what i am after?
k


I'm sorry, but what you're asking for doesn't really make any sense. You can't use a delay to playback at consistently slower or faster speed. Eventually you'll run out of delayed frames or room to store them. For a delay to change the playback speed, it either needs to be getting longer or shorter.

I tuned up my patch a bit, and used matrixset. Maybe this helps?

this version works perfectly fine! and i do think it will solve a big part of my problem if i play around with it a bit... i'll try as soon as i'm back at my office - thank you so very much!!!

regarding the "constant": you are totally right, if i get slower, i have to get faster at some point if i do not want to run over my buffer limits. i just thought maybe i try splitting up my problem and first try to get some sort of general (constant) playspeed control before is start messing around with oscillation. as said, i had a somewhat similar problem some time ago and there i used this float-counter for a "frame to playspeed conversion" and could then do pretty much faster/slower as much i liked.

however this curve-approach could be much more elegant...

i'll fiddle around with your patch and report back, again: thank you so much for helping me with this!!!

hello again,
this version works perfectly fine! and i do think it will solve a big part of my problem if i play around with it a bit... i'll try as soon as i'm back at my office - thank you so very much!!!

regarding the "constant": you are totally right, if i get slower, i have to get faster at some point if i do not want to run over my buffer limits. i just thought maybe i try splitting up my problem and first try to get some sort of general (constant) playspeed control before is start messing around with oscillation. as said, i had a somewhat similar problem some time ago and there i used this float-counter for a "frame to playspeed conversion" and could then do pretty much faster/slower as much i liked.
however this curve-approach could be much more elegant...
i'll fiddle around with your patch and report back, again: thank you so much for helping me with this!!!
k


translating absolute frame numbers to playspeed...