Csound Csound-dev Csound-tekno Search About

[Csnd] Re: lots of notes with the instrument using ftgentmp ate up the ram

Date2009-08-15 23:14
From"Partev Barr Sarkissian"
Subject[Csnd] Re: lots of notes with the instrument using ftgentmp ate up the ram
When setting up tables, I tend to go with 4096 or 8192 myself.
Less than that and I get noises and artifacts, more than that
tends to bogg down and I get drop outs. I typically run MacCsound
1.25a at 44.1-kHz or sometimes 48-kHz, and it's a 2004 Mac G4 
Power Book laptop or my G4 Desk on 10.3.9 ops system.

I especially get slow as molasses when I have a lot of notes
and/or voices, so keeping it within 4096/8192 works well. I 
guess it's all about finding your system limits and staying
within them to find that happy medium.

-Partev


=================================================================



--- electriclightheads@gmail.com wrote:

From: '2+ 
To: csound@lists.bath.ac.uk
Subject: [Csnd] lots of notes with the instrument using ftgentmp ate up the ram
Date: Sat, 15 Aug 2009 07:20:38 +0900

oh oscil3 .. i've forgot to test things with that one
but i did do test some other variations

made the ftable short
chose 8192 instead of 65536
and now it looks like this

ift_sa   ftgentmp 0, 0, 8192, 10, 10, 5, 5, 1, isa, iri, 0, iga, ima,
ipa, 0, ida, ini, 0, 0, isa * iri, 0, iri * iga, 0, iga * ima, 0, 0, 0,
ima * ipa, 0, ipa * ida, 0, 0, 0, 0, 0, ida * ini, 0, 0, 0, ini * isa,
0, 0, 0, isa * iri * iga, 0, 0, 0, 0, 0, 0, 0, iri * iga * ima, 0, 0, 0,
iga * ima * ipa, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ima * ipa * ida, 0, 0,
0, 0, 0, 0, 0, ipa * ida * ini, 0, 0, 0, 0, 0, 0, 0, ida * ini * isa, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ini * isa * iri, 0, 0, 0, 0,
0, 0, 0, isa * iri * iga * ima, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, iri * iga * ima * ipa

but the ram still maxes out at about 30 sec of the same sco
maybe it should even be smaller than 8192?

well the population of the note of the original sco was like
3000 notes * 7 instances(of oil.Sa()) / 180 sec

so i re-algo-comped to make it
500 notes * 7 instances / 180 sec
still seemed to be over populated

100 notes * 7 instances / 180 sec
could have the whole sco accomplished
and the usage of the ram was always less than 130MB

50 notes * 21 instances / 180 sec
was also okay
ram was always less than 150MB

the result goes like this:
http://sarigama.namaste.jp/loops/lotsofsilence.mp3
i like some part of it but
the silence between the hills are a bit too long
well am producing these to play with female_reading_somethang.mp3 
so then it might be okay

so .. maybe i should just use a value somewhere between these .. i think
but will take a look into the manual for oscil3

On Thu, Aug 13, 2009 at 06:12:24PM +0100, jpff wrote:
> Use of oscil3 slows it down, but I cannot perceive much difference
> ==John ff
> 
> 
-- 
'2+ es podcasting
http://sarigama.namaste.jp/podcast/rss.xml


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"




_____________________________________________________________
Netscape.  Just the Net You Need.