Hi everyone!
I found this thesis from 2008 online and was reading it, when it occurred to me that the author's ideas about quasi-periodic sounds being more pleasing could be tested and mocked up in pd. Here's the outcome! Because of the thesis, this had to be licensed under Creative Commons 4.0 BY-NC-SA. Sorry!
[rosc~] takes two creation arguments, and every cycle of the sine wave it generates a random frequency in a specific range to use for the next cycle. I played with it a bit and found that a range of 2 sounded the best to me, but it does require some calibration. Sorry if I'm not too clear right now, it's really late, and the help file should clear up most confusions.
Hope someone enjoys it!

http://www.pdpatchrepo.info/hurleur/rosc~.zip

I'm not on my conputer, so can't check the .zip, but random osc can be easily done like this:

Use a phasor~ for the oscillator, a noise~ object for the randomisation, and a samhold~ to sample the noise value every time the phasor~ starts a new cycle.

Then, use a s~ / r~ pair to route the samphold~ output back to the pitch input of the phasor~, and scale accordingly with simple arithmetic.

Finally, to make a sinewave oscillator, take the output of phasor~ through a cos~ object.