I have an OSC device that you could think of like a slider--when you touch it along a line, it produces a float between 0 and 1, corresponding to the proportional distance along that line.

It's easy to feed its output into osc~ and make a sine wave audio--sounds like a slide whistle. How do I use it to index a play position in an audio file, thus enabling scratching?

That is so cool and simple. Thanks! Sorry, should have said something sooner.

Here's a more complicated case--what if I want to loop the sample, and scratch from where I'm at at that moment instead of using my slider to determine the position in the sample directly?

I understand if you don't wanna create the code, but if you could point me in a direction, that would be helpful.

Use a phasor~ or vline~ to control the index of your sample. When a message is recieved from the slider, stop playback and add or subtract the slider data. Once the data stops changing, resume normal playback.

OK... got as far as a phasor~ I can switch on and off, playing a sample. Borrowed from a youtube video. It works, well. Now I need a way to base the change in the indexing into the array on the change in the slider. So I need to track the difference between the current frame of data and the previous frame. I don't see a way to do that.

OK... found [pipe]. Which doesn't exactly do what I think I need. Pipe makes a delayed stream of numbers. My first thought was, stream the output of pipe and the raw numbers to [-] and get the differences. The trouble is, I'm measuring overlapping intervals, which will move the scratch farther than I think I want.

I think I want to measure the 'slider's' motion incrementally. Like at intervals, measure how far it's moved in that interval and return that value. I think there's a metro and a pipe in that patch, but I can't quite figure out how to pull a value out of a stream at intervals. Any ideas?

OK figured that out, I think. Next up, stopping the phasor~ and incrementing from the change value output. Any ideas would be much appreciated.

Ok.... got this far.

The attached patch loads and plays a loop. It has a toggle that starts/stops the loop. it has an OSC-based switch between playing and scratching.

The playing works fine. The scratching doesn't sound good--close but too fragmented, like it's playing bits and pieces of the sample, not in a smooth way. If anyone knows a way to make this sound better, I'd appreciate your wisdom.

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

Ok, turning my attention back to this...

Still not getting a good sound out of this patch. when I switch to scratch, and I move the slider right I get a buzzing sound, and when I slide left I get silence, with intermittent popping. I recently saw this:

which looks similar to what I'm trying to do here.

Any ideas? I've attached a version of this patch that you can interact with without an OSC input. Just pick a sample and slide the slider.

OK, I've got it... I had two problems:

there's an interval of time--a bite of the sample (50 milliseconds in the attached patch). It needs to be the same length as the time fed to vline~. Duh.

Also, modulo can return a negative number. -17 % 5 = -2. That doesn't work for computing the index into the sample.

This scratches nicely. It's set up with the length of the slider set to the same interval of time as one turn on a 33 1/3 record. Playing with th bite size is interesting--too low and scratching is just annoying. Didn't really play with values over a fifth of a second.

I'm sure there are more efficient ways to implement this. I'd be curious to hear about them from pd gurus.

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

I used some of this work as the basis for my scratching system. I've posted it in a thread here: http://puredata.hurleur.com/sujet-5355-scratching-system if you're curious.