[Csnd] PMAX question
Date | 2009-06-09 09:04 |
From | Schumacher Marlon |
Subject | [Csnd] PMAX question |
Hello List, I wonder if there might be a way to change/override the 'PMAX' variable (max number of p-fields in a score statement) without a re- compile of csound? afaik PMAX is currently set to 1000. (csound5.10 double, MacOS10.5.6 on intel) best, Marlon |
Date | 2009-06-09 10:01 |
From | Panos Katergiathis |
Subject | [Csnd] How can i find ALL my previous posts to this list? |
Hello all I am trying to locate an older post of mine to this list, one that was planning to be a tutorial for building Csound on Ubuntu (from source). Is there a way to locate and collect all my previous posts to this list? Panos |
Date | 2009-06-09 10:31 |
From | jpff@cs.bath.ac.uk |
Subject | [Csnd] Re: PMAX question |
> Hello List, > > I wonder if there might be a way to change/override the 'PMAX' > variable (max number of p-fields in a score statement) without a re- > compile of csound? > afaik PMAX is currently set to 1000. (csound5.10 double, MacOS10.5.6 > on intel) > > best, > Marlon The simple answer is no. At present the structure for a note event is pre-allocated and this parameter determines the size. Making it a variable would take quite a lot of thought and recoding; at present i cannot see how to do this without a lot of additional runtime checking values of n for Pn to see if it exists. Actually I am bemused that 1000 is insufficient. It used to be 50 in theoriginal csound ==John ff |
Date | 2009-06-09 14:20 |
From | Hector Centeno |
Subject | [Csnd] Re: How can i find ALL my previous posts to this list? |
Hello Panos, Did you try searching at the Csound list archive in Nabble? http://www.nabble.com/Csound---General-f483.html ... or also accessible from here: http://csounds.com/community Cheers, Hector On Tue, Jun 9, 2009 at 5:01 AM, Panos Katergiathis |
Date | 2009-06-09 16:37 |
From | Marco Stroppa |
Subject | [Csnd] AW: Re: PMAX question |
Hello! I must aknowledge that I stumbled on this limitation several times, to the point that I had to fid ways to reduce the size of data when generating them automatically. This is particularly true, for instance, in case of GEN07, where a breakpoint function is generated from analysis data (of course, if one wants to manually write 1000 p-fields this can be quite tedious...!). In the original versions of csound this was not possible (not only because of PMAX). The amount of GEN functions used to be limited, and is not dynamically allocated. There may still be other (undocumented) limits within the orchestra. As far as PMAX is concerned, I did not look at the code carefully, but perhaps one could imagine to set it to 10000 (or 100000...??), in order to minimize this kind of problems. Memory is no longer so expensive nowadays. Best, Marco ----- Ursprüngliche Mail ---- > Von: "jpff@cs.bath.ac.uk" |
Date | 2009-06-09 16:57 |
From | Richard Dobson |
Subject | [Csnd] Re: AW: Re: PMAX question |
Marco Stroppa wrote: > Hello! > > I must aknowledge that I stumbled on this limitation several times, > to the point that I had to fid ways to reduce the size of data when > generating them automatically. This is particularly true, for > instance, in case of GEN07, where a breakpoint function is generated > from analysis data (of course, if one wants to manually write 1000 > p-fields this can be quite tedious...!). In the original versions of > csound this was not possible (not only because of PMAX). The amount > of GEN functions used to be limited, and is not dynamically > allocated. There may still be other (undocumented) limits within the > orchestra. As far as PMAX is concerned, I did not look at the code > carefully, but perhaps one could imagine to set it to 10000 (or > 100000...??), in order to minimize this kind of problems. Memory is > no longer so expensive nowadays. > My instinct for that size of data would be to write it to an ftable, and index the entries in the orch. Could even be a raw f/p quasi-soundfile (and the GEN fucntion would skip auto-normalization). One issue with writing parameters as text is that the precision/accuracy of the numbers is determined by the number of decimal places to write (up to 17 to represent a full-blown double). All of which make the line even longer, a right pain for viewing in text editors, etc. Richard Dobson |