Frank Barknecht wrote:
On Tue, May 18, 2010 at 03:33:38AM +0200, PF wrote:
If the file you are playing is too big for a table and you are playing with readsf~ or so, then you could still register locations. Alternatively just record chopped sections with writesf~ and name files with numbers counting up
it will be quite large, so readsf~ might be the better way out. thanks, frank.
or so. Creating tables dynamically is a bit of a pain that I would avoid here.
even if it's not perfect for this, i wouldnt mind if someone can point me to some examples of dynamic tables.
You create them with the "usual" dynamic patching approach: Send an "obj"-creation-message to a subpatch-receiver like this:
[obj 10 20 table tablename( | [s pd-$0-subpatch]
[pd $0-subpatch]
Replace "tablename" with for example "$1-tablename" and feed it with a number to create more tables, probably you will also want to change the y-coordinate accordingly to something like like 20*<tablenumber>.
Another approach could be to simply append samples in one big table and save the indexes. (but this depends a lot on how many chopped samples and how long they are)
I've used a similar approach for a piece where I wanted to trigger certain samples: in that case the 'chunks' were small files already existed on disk, but I wouldn't have known how many files I would have finally, so using a fixed number of tables wouldn't have worked. I can point you to the piece and share the patches if you are interested.
Kind regards, Lorenzo.
Ciao