Krzysztof Foltman wrote:

> >can't help code them, since I have zero programming ability. I still think
> an
> >orc/sco that emulates the original preset is the most educational and
> flexible
> >option,
>
> What about macro (well, a sort of) libraries ? Instead of hardwired opcode,
> one could use a bunch of Csound-language-based presets, for example
> something
> like that:
>
>         macro sinelfo(iamp,ifreq) returns kout
>
> bsine   gen10 1024,1 ;'b' prefix could mean 'global ftable' (shared among
> instances)
>
> kout    oscili iamp,ifreq,bsine ;'bsine' substituted by number of the ftable
>
>         endmac
>
> or anything similar. Of course, that would be slower than optimized C
> language,
> but much more flexible (anyone could add his/her favourite building blocks),
> and sharing such building blocks could create a kind of source of
> inspiration for many. Or...?
>
> And for the multisampling-related stuff... what about the "conditional
> loscil" opcode ? one which could output 0's (and not play/calculate at all)
> when some kind of condition is not met ? Much of multisampling/multilayering
> related work can be simplified using such unit. Another optimization can
> use "range-based select" opcode, like this:
>
> aout    select
> kctrl,asig1,kmin1,asig2,kmin2,asig3,kmin3,...,asigN,kminN,kmaxN
>
> (or even quantize/select pair, first converting a continuous value to
> discrete
> one - for example, pitch to sample zone number - and another outputting only
> Nth signal)
>
> ...kf