Hello, had a few questions/suggestions :

• In rnd-sample, it would be nice to have a param controlling how smooth the result is, i.e. something like the "alfa" value in vector-smooth ( between 0 and 1 ) that allows you to smooth the result to some degree. For example, if I want a rnd-sample of a list of pitches, I could control how much the resulting pitch seq moves in steps or large jumps.

• In filter-repeat, the :position t :span :pitch option works great to prevent repeating X occurrences of pitches, but would it be possible to also have an option to prevent X occurrences of pitch classes?

For example, the below chooses pitches from 3 octaves of a major scale and prevents some repetitions of exact pitches. But if it returns something like c3 c4 d3 e3 d4 f5, then a filter on pitch class would only return c3 d3 e3 f5 for example.

(filter-repeat 4 (rnd-sample 128 (make-scale 'c3 22 :alt '(2 2 1 2 2 2 1))) :position t :span :pitch)

Finally, I'm not seeing documentation in the Assistant for some functions ( e.g. rnd-round, rnd-number, binary-invert and gen-white-noise ) which are referenced in the Composer Workshop courses. I'm on the latest version of OM 4 for Mac. Can this documentation be added?

Hi there,

Regarding the documentation for functions, some of them were used during the creation of the tutorials but have since been replaced by newer versions. For instance, the functions rnd-round and rnd-number can be replaced by the new function rndn. Similarly, the function gen-white-noise can be replaced by gen-noise.

As for the function binary-invert, I’m not sure which actual function can replace it.

S.

@opmo Any idea if it would be possible to implement the first 2 items ( rnd-sample, filter-repeat )?

IMHO sampling with ˋrnd-sampleˋ should keep the current connotation that is strictly speaking a) returning draw-by-draw-independent observations from a given distribution by b) enforcing a flat-uniform probability distribution for the draw.

Requesting results as described like „smooth“ does something different.

In my view there are already functions in OM (according to the docu) which can be used (with combi of others like mapping) if the result should not be independent or comply to a non-uniform probability-distribution for the draw .

Not independence:

• gen-ma-time-series

• gen-ar-time-series

Non uniform distributions:

• One among many functions: gaussian-distribution

Or just upsample those elements which should appear over-represented before ˋrnd-sampleˋ, even with a uniform probability distribution for the draw, those will have a higher probability for being sampled after all.

Update to filter-repeat function. 4.0.30647