[RTcmix-discuss] more memory issues

[John Gibson]

Thinking about this more...

If we set the max delay time based on the first dynamic frequency value
received by COMBIT, then if you want to start with a high freq. and
gliss to a low freq -- or use makerandom to generate frequencies -- then
you're screwed.  The value received by init is not some special setup
value that is distinct from the stream of values emitted by the PField.
It's just the first value in the stream.

In other words, I can imagine situations where you can't arrange for
this value to be the min freq without sacrificing something else.
That's why I had to hard-code MAXDELTIME.  Same for DELAY.  So I
don't think your solution works, at least without making it impossible
to do certain things with those instruments.

JG


On Sat, 8 Jan 2005, Brad Garton wrote:

> Ditto what I said for COMBIT with DELAY... with MAXDELTIME #defined at 
> 20.0, it's almost 2 meg/note.  Note a big problem for applications that 
> use only one or two DELAY notes, but if you schedule a bunch (maybe some 
> bizarro-multi-delay thingy) it can add up pretty quickly.  These machines 
> can certainly handle the allocation, but it takes a fair amount of 
> scheduling-time to do it.
> 
> Plus is we go towards letting the initial delay-time setting determine the 
> delay lengths, then we're not limited to whatever we #define the maximum 
> length to be.
> 
> brad
>