From the original SynthMaker forum (2004–2014), with its members’ names and dates.
infuzion ·
test procedure: 1) download exonerate's SPL: [old link] 2) copy a 2nd player looper & connect it so both play at the same time. 3) load samples in both, set looping on, & set loop points as you please 4) play a note; CPU goes up to about 95%, extra memory gets used at the rete of 0.7k/second, & 1000s of page faults are made/sec.
Doesn't seem to happen w/SM's sample players. Bummer, I like the module
infuzion ·
If you delete the Time bar inside the sample looper, the issue goes away. Well, aside from the 5 pagefaults/sec, but that is normal for SM.
exonerate ·
yeah i've just tested it with one sample player and the cpu goes to 40% on my computer, i changed the ticker 100 for a tick 25 and it went down to 20%, which is obviously still really high, so the time bar is definetly going to have to go and stay gone this time
i can't believe i never noticed it, thanks for pointing this out infuzion
cheers
infuzion ·
welcome The CPU is one thing, but all that memory being eaten & the page faults makes me believe something is really wrong. I'm suspecting it is the Area S&H or Area Union.
Honestly, for what I want to use your module for (short one-shot samples) I don't care much about the cursor. But for longer looping programs, esp glitch, in-depth playback synth, etc, I would say the cursor (once working) is needed. Ableton Live has no problem with the cursor.
sambean ·
There doesn't seem to be that much going on inside the time bar module, it shouldn't take that much horsepower. Try replacing the mono to float with a float constant and see if it still eats memory..
exonerate ·
Ok i've done a bit of testing, and it seems to be a problem with redraws, i've added another two redraw limiters before the redraw, and replaced the ticker 100 with a ticker 25, and now the CPU seems to be OK but theres still a memory issue, even when i disconnect the mono to float and replace it with a float constant.
heres the looper with the extra redraw limiters.... http://www.box.net/shared/4vfnuufoq2
i've tested with short and long samples and the memory keeps on increasing on both although on longer samples the memory doesn't increase in as big jumps as short ones. page faults are really bad as well, worse than firefox!
also i've just done a little test with just the area redraw without the area union and sample and hold, and the memory still increases but not as much and it seems to go up a bit and then down a bit but gradually gets higher.
cheers
infuzion ·
Pagefaults are always bad in SM; it is something I've complained about for the last 1.5 years atleast. You can leave SM open with no schematics open & it genterates about 6 faults/sec. Firefox used to be almost as bad, but the released 2.0.x seems to be much better.
Wrote something here? To have a post removed or your name hidden, write to [email protected].
SynthMaker