Ok... I have a sequencer built in Pd that is sending bangs/notes to my sampler and then the audio from the sampler is coming back into Pd where I can effect it. (I know running samples directly in Pd would end up being less complicated without adding extra noise and blah blah blah, but that doesn't really factor into my question, so... moving on...)

I know how to accomplish the stutter effect using audio signals, but what method would you use to accomplish the same effect with the bangs that are being sent out to the sampler. Like a data stutter instead of an audio signal stutter. So it would grab a little chunk of the bang/note sequence and repeat that over and over instead of grabbing the little bit of audio signal. That's what I'm after.

Would this have to involve recording bangs into an array and then reading them out of that or something? I have not attempted to figure this out yet, but thought I would throw the question out there to see what anybody thought. Is there a better name for this other than "stutter"?

Maybe I could have called it a MIDI-stutter in the heading

i'm afraid, only users of your sampler will be able to help you there!?

maybe you didn't understand what i was trying to say or maybe my explanation wasn't clear.

what i want to do shouldn't involve the sampler as a device (an electribe, by the way), only the bangs being sent to it. all the sampler is doing is storing the sounds that i am triggering. maybe i should have said "drum machine" instead of sampler.

for example, i have it set up on the sampler to be note 1 = bass drum, note 2 = snare, etc. so if i send 1 1 2 1 to the sampler, it will play bass bass snare bass. i want to have a stutter-style effect that will be able to grab whatever sequence of numbers that is playing and repeat that little snippet of information. this would (in my own theory) replace the use of an audio-based stutter effect and achieve (sort of) the same effect by just repeating a short pattern of numbers going to the device (sampler, drum machine, whatever)

does that make more sense, or is this just a stupid idea to begin with? i'm thinking i could record the note numbers into some sort of array and play them back out of the array in a little loop or something. i just got home from work and i am going to start trying some things.

just wondering if anyone else has done this and what kind of results they got. thanks.

ah, ok... i didn't understand.
maybe something like this?

http://www.pdpatchrepo.info/hurleur/midi_stutter.pd

sorry, forget that, i should have read your post properly.. you don't want to repeat only one step but a whole sequence. mhhh, must think about that, maybe something with lists...

cool. i'm interested in different ways of creating and manipulating patterns.

i feel i am very much an improvisational musician (who happens to be extremely interested in computer-based music) so Pd has really gotten me excited about the possibilities of performing with a computer again. i'm still working stuff out, but i thought that this would be a way of getting into manipulating patterns on a numerical level rather than just messing with the audio (which is still really cool, by the way.)

anyway..... thanks for your interest. let me know if you think of anything. i did make a little pattern recorder that isolates each of the 9 sample "pads" on the electribe so i can record a pattern for each sample. now i just need to figure out how to repeat and loop a little chunk of the patterns. i'll keep trying.

Hi there

If your sequencer a timeline-based, "sequence stutter" is quite easy to achieve. Say your clock runs from the beginning to the end of your sequence just like a table lookup index.

Then you can play with the timecode (by looping it with [mod] and set an offset or whatever), and do some sort of granular synthesis applied to your sequences. Of course, this requiers timecode-based event triggering, and I don't know how you built your sequencer.

With this principle, you can do various local looping, modify tempo, read reverse or random, even playing several part of the sequence at the same time, and do it live. This way you modify the data that is then sent to external instruments without affecting the audio process. Is that what you were thinking about ?

(hopefully, one day I'll finish this dream sequencer I started to patch some years ago...)

That pretty much sounds exactly like what i want to do. Most of the time I am running my sequencer as a 64-step setup where i can change the number of steps from 1 to 64. This is driven by a metro like so:

[metro]
|
[ f ] x [+ 1]
|
[ % ] _ [number of steps]
|
[sel 0 1 2 3 4 5 6... etc.]

I could be slightly off on the order of those, but i think you get the idea. I was trying to type that out from memory (and that was my first attempt at representing a patch typographically) Anyway, all the outputs from the [sel ] go to toggles for each step which determines whether or note a bang occurs, which in turn sends the notes to the sampler/drum machine.

Last night I was playing around with recording and running patterns out of arrays instead, but in any case i would imagine doing your timecode thing would work either way. I just need to know how to go about doing that.

Is there a "timecode" object or something? I'll look it up. Thanks for your suggestions. p.s. - keep 'em coming.

You already have the "timecode" (it's what the [sel] object receives) and the "timebased-triggering" (the [sel] object itself).

For basic stuttering, just [pipe] with feedback the timecode with an appropriate duration and send it back to the [sel]. Don't forget to manage what you send to the [pipe] to avoid infinite looping (with a [spigot], for instance).

awesome. don't think i have ever used [pipe] before. i will look into it later tonight. thanks so much.

one quick question though since i am curious (and if you have time for it): when you say to pipe the timecode an appropriate duration, would this duration be in milliseconds? and also, when you say pipe with feedback, what do you mean by feedback?

again, i'm sure this is stuff i can find out later when i start messing with [pipe] and the help fiel, but just wondering. thanks again.

note: i understand what you mean by feedback, just not how to go about doing it. but trying it out myself would probably be a good place to start before i smother people with questions.

anybody able to give me any additional info on this? maybe elaborate on what kyro was trying to tell me when he said "just [pipe] with feedback the timecode with an appropriate duration and send it back" (???)

i was messing around with [pipe] and couldn't really understand how it was any different than using [delay] because it seems to just delay the output of what goes into it by whatever number you give it. maybe i'm missing the point.

what i would want to do is loop a short bit of info passing through. like, cap off what is going into the pipe, have it repeat for however long i let it repeat, and then open it up again to the normal stream passing through it.

anyone?

If you're just doing a looping step sequencer, you might not need [pipe], just another (slightly more complicated) [% ]. I don't know, it might actually be less convoluted with [pipe], but for some reason I didn't really feel like going there . Take a look at the attached patch. I think it does what you are looking for.

http://www.pdpatchrepo.info/hurleur/seq-stutter.mmb.pd

Yeah, I realized that using pipe was far more complicated than I expected. The difference between [delay] and [pipe] is that [pipe] delays series of inputs, not only the last one.

The only advantage is that unlike [%], a [pipe] stutter can work with any timecode resolution, but you may have to synchronize its delay time to your tempo. Sorry for the messy patch

http://www.pdpatchrepo.info/hurleur/pipe_non-audio_stutter.pd

While it's true that [pipe] doesn't reset when it receives a new message like [delay] does, I think the more important difference is that [pipe] sends out the message it receives, whereas [delay] will only send bangs.

Good work on the patch. It's not too messy .

by the way, thanks guys. i only just started examining it and seeing how it works last night, but it definitely seems like just what i was after. cool!!!

Hi

I posted my timecode generator there :

http://puredata.hurleur.com/viewtopic.php?pid=15008#p15008