VI freakout in Forte

[William]

I am using Forte VST rack triggered externally to play VI and have had some problems with erratic playback - in an extremely complex, huge orchestral piece, it will play back o.k. one time, then the next start a weird popping off of notes, almost like losing polyphony, but it will do it on single lines so it cannot be too many notes. Though maybe it could be too much data, since I am using a lot of controllers simultaneously. When I stop and play the section again, it plays back fine. Does this sound like a problem with too big an instrument set up for the CPU, or is this something with the Forte rack system? If anyone has any ideas I would appreciate it.

[dpcon]

I have the same problem in Plogue Bidule William (crackles, hitching, sputtering.) Also when I isolate the passage to reduce polyphony it still occurs. Buffer settings don't seem to change things and turning off release samples seems to help but not cure. I'm wondering if it's the VI itself (which is why I'm posting here, so I hope that's okay.) Maybe VSL can investigate or poll it's users as to how pervasive this problem might be.

[William]

I just was trying to do a track again, sustain violins layered with staccato playing four note divisi to another four note divisi. That is a lot of layers, yes, but not really that much polyphony at least compared to things I've done before, so I must have something basically wrong in my system but I have no idea what. The notes start popping out, like the old familiar lost polyphony sound. There is no way I am overusing the voices here, and I have plenty of ram reading on the VI console. I also tried adjusting the sample buffer size and number, as Dave said, and there is absolutely no effect.

[nicks]

Dear William and Dave,

Had this problem a while back (I also use Forte). I found it on a big orchestral arrangement (Wind by 2s, 4 horns, 3 trumpets, 3 trombones, 2 bass tromb, tuba + ensembles, percussion and strings). Everything was okay at the beginning, but I found that in my tutti sections, notes started chopping off, and muting the parts didn't routinely help. As you did, Dave, I switched off the release samples which made it a bit better, but did not fix it.

I came to the conclusion (and can't remember what tests I did now) that it was polyphony creating CPU spikes of 100%, from too many VI instances - even though the polyphony was not that high. Another possibility (thinking about it now) is that it was CPU spikes due to waits on disk.

My solution (you're not going to like this) was to buy an extra PC for the strings (told you that you weren't going to like it). This seemed to resolve the issue, but I suspect it would also be improved by:

1) Have a "lite" version of your Forte rack with release samples switched off
2) Record a number of tracks as audio, and remove them from your rack
3) Replace any glissandi made with each note as a sample with a single glissando sample (if possible)
4) Move any processing from the Forte machine(s) back to the sequencer machine (if different)
5) Try moving some of the collections to a different physical disk (not just a different partition).

None of these are particularly satisfactory, but if you find your CPU spiking (i.e. if it's the same problem as mine) then additional machines may be the real solution.

I hope you find a solution for this - it really is the most irritating problem.

If it is of any help, my setup is 4 forte PCs (Wind, Strings, Brass, Percussion&Other), connected to Logic (which is running on a Mac), and I'm not experiencing these drop-outs any more. If I can be of assistance troubleshooting (to my limited knowledge), please PM me.

Kind Regards,

Nick.

[William]

Thanks Nick for those suggestions.

I did whittle down the number of instances from 16 to 4 as an experiment, with only violins playing, but it still is happening. I am trying the same set up in Cubase instead of Forte to see how that compares. Also, I am going to optimize the tracks to see if that helps.

As an example of what will cause this to happen - hold down a note as a pedal point, a middle c for example, and while holding that down, start playing short two note chords with the same instrument. After about four or five of those chords, the middle C pops out. Is that the limit of polyphony? I can't believe that!!! However, the exact same popping out occurs in playback off the sequencer, ERRATICALLY, for example after a heavy legato section, and then some simple sustains. The sustains, without that much polyphony, pop out. I then start it up again, go over the same section, and they play o.k. So the erratic nature of it suggests it is not mere loss of polyphony. And the hard drives are defragged, there are two gigs of ram, 3.4 ghz processor, etc.

[JBacal]

William,


Make sure that your Gigastudio memory tweaks have been restored back to the windows default. The VSL VI does not like the Gigastudio memory tweaks. This used to cause notes to cut off for me.

Best,
Jay

[jbm]

Hey William,

I too have been having strange, erratic pops and clicks. It's not so much drop-outs, like I think you're experiencing, but it is erratic, and it does seem to be related to some general performance issue (gets worse with larger arrangements, but isn't necessarily cured by muting/disabling VIs). I've made the situation *much* better by updating my audio card to a DSP version (moving up to the RME HDSP9652 from the DIGI96/8 ), and moving from the ADAT sync I was using between machines to word clock. However, I still get the occasional problem, and have even had a few hard freezes while running the VIs! [[:|]]

I know the freezes are very strange, and I don't particularly suspect the VIs for this... in fact, the freezes (finally) made me highly suspicious of the power in my studio, so I've ordered a power conditioner today. We'll see how that goes.

I realize you're using a single-machine setup, in this case, so I should mention that I've had similar problems running the VIs in Bidule, and sequencing from Sibelius. I generally suspect Sibelius' (and Mac OS X's) extremely hungry screen redraws for this, but, like your situation, it's not at all consistent. So it's hard to be certain...

I suspect some form of timing-related issue... But for now I'm putting my hopes on improving my studio's crap power. It makes sense to me, though it probably means nothing to your situation.

cheers,

J.

[William]

I never did any memory tweaks.

Thanks for those suggestions jbm. I still have not compared the performance within an internal vst-sequencer though somehow I doubt that will make any difference.

[William]

Sorry to be repetitive, but I still am having problems with this. What is going on here -

if one holds down a sustain note, and then plays on the same instrument short, separated two note chords on different pitches, several of these chords in a row - that first sustain note cuts out. Try it, please and see if this happens on your system.

I have never had a synth or sampler do this. When polyphony is exceeded, the NEW notes pop out, not the previous one that is being sustained by being held down (as opposed to pedal sustain). What is going on there? I don't understand it and therefore think there is a basic computer problem, or programming problem.

I use that as an example of how to create this note-popping-out effect, not for fun, but because I am having major problems with VI now. The notes do not play reliably on this system when I try to master audio tracks. I have to keep starting over on a midi playback, and it does not seem to be a polyphony or multiple-instance cpu overload issue, since the SAME THING happens with scaled-down versions of the rack.

Can someone please tell me if this is a phenomenon that occurs with VST racks triggered externally, as opposed to internal VST performances within a DAW? That is, assuming anyone knows. I noticed a Mac thread with a similar situation - I wonder if this is the same thing happening.