You're right.
The "smoothness" was meant to create less "jagged" waveforms that still contain randomization. I used the [line] object to make the random additions/subtractions function gradually as opposed to value-by-value, but I was off in terms of how long the ramps should happen in milliseconds. The newer edit adds a scale to make the [line] make sense within the "speed" parameter.
I also altered some math and added [min] and [max] to make the waveforms always within -1 to 1. This adds some "clipping" effects, but the patch is based on creating interesting waveforms in the first place so I think this is an ok addition.
I think overall the version 2 patch works better, and I hope others can use it! Please offer whatever feedback you might have.
~A
On Thu, Apr 14, 2016 at 12:17 AM, S.E.P. < dreamoftheshoreofanotherworld@gmail.com> wrote:
Hello AP,
Thanks for sharing your patch. I like how, by rewriting the wave-form gradually, it allows for a gradual transition from a purer, sine-like sound to a grittier, more metallic sound. I wonder if it's necessary to have both a "randomization" and a "smoothness" controller, as they seem to be affecting the same parameters, i.e. putting both controllers up high and putting both down low produces similar results. Making the wave-form more random seems to make it less smooth at the same time and vice-versa.
You also need some kind of volume-scaler. When producing more randomized waves, the amplitude quickly became too loud and I had to lower the [~*.5]s to [~*.1]. This probably has something to do with the randomized wave-forms falling well outside of your arrays' boundaries (of -1 to 1).
Best, Elliot
On Thu, Apr 14, 2016 at 3:14 AM, AP Vague apvague@gmail.com wrote:
I wrote a patch to make interesting, pseudo-randomized waveforms in real time as a way to add complexity and variation to drones. The patch duplicates a table with randomization and uses [tabosc4~] as the final oscillators to dac. This version cascades the duplication process; table b is a copy of a, c is a copy of b, d is a copy of c. The results are pretty interesting and I think the signals could work well with other manipulations like filters.
I'm afraid that this might have been the extreme longhand way to do this, ha ha ha. It works fine for what I wanted but I thought there might be some ideas on how to accomplish it in an easier way. I also just kind of wanted to share it.
Cheers.
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list