as i said previously, loopstart/end have not been implemented.
I think the problem is to implement this 'as is' would be inefficient...
the table lengths are powers of 2, so that we can determine positions by bit shifting, setting arbitrary end points breaks this.
(working on the assumption (for being useful) is that pos = 0 , loop start and pos = 64 (100%) = loop end)
what we could do is:
- allow arbitrary start points. (this is just used as a simple offset)
- change loop end, to loop length, where powers of 2 are offered.
this would not allow for arbitrary sample lengths (just 2^x) , but would allow for handling processing many samples from the same wave file.
I will also add loop start/end are currently parameters, it would be more efficient to have these as attributes (compile time change) , it would be nice to have a variant with inlets, but the issue then is how to 'insist' the length is 2^x.
(I think its these oddities, that mean table parameters are usually expressed in % (0-64) rather than int indexes)
