Quantized notes sound un-quantized

[paulkeeley]

Originally Posted by Totty

Hi Paul,

I had a listen. I can hear from your audio what you are saying. I'm more logic based and I set up a similar test using your patch. With just those 4 chords and changing oscillators I have to say it sounded more accurate than your file... Was this played originally in VC or via midi?

thanks for giving this a go Totty - i haven't tried the test yet in Logic - i assumed (possibly incorrectly) that the MIDI timing between Ableton and Logic would be the same - i'm predominantly a Logic user myself.

i recorded the notes in realtime and then quantized them to 16th notes within Ableton. i'm running an iMac C2D 2.16ghz so horsepower shouldn't be an issue there.

i'll give it a shot tonight within Logic and see if that makes any difference whatsoever.

cheers,
Paul.

[paulkeeley]

ok, i tried the exact same MIDI clip in Logic 8, and i'm still getting the same note drift.

here's the Logic project if anyone's interested in trying it out further:
http://www.mediafire.com/?sharekey=9...db6fb9a8902bda

cheers,
Paul.

[paulkeeley]

aaaaaaand since i clearly haven't posted enough today on this topic, i've made a bit of a breakthrough.

in VC mode, everything is rock solid. it's only in MIDI standalone mode where the timing goes to hell and back. again, no problems with my other synths whatsoever using the same MIDI interface(s), cables, etc.

so, it must simply be the way the TI is handling MIDI input info. or at least that seems to be the case with mine.

hope this helps in any way, however insignificant.

cheers,
Paul.

[Totty]

Originally Posted by paulkeeley

aaaaaaand since i clearly haven't posted enough today on this topic, i've made a bit of a breakthrough.

in VC mode, everything is rock solid. it's only in MIDI standalone mode where the timing goes to hell and back. again, no problems with my other synths whatsoever using the same MIDI interface(s), cables, etc.

so, it must simply be the way the TI is handling MIDI input info. or at least that seems to be the case with mine.

hope this helps in any way, however insignificant.

cheers,
Paul.

Hi Paul, sounds like you have made a breakthrough I'm downloading your Logic file and will have a go here too. From what I know the VC timestamps midi events so the timing should be accurate on VC, However as you say, your other midi synths don't exhibit this problem.

I will also try a midi connection to my TI, may take me a few days as I'm set up very minimally at the moment, so I'll have to dust off my AMT8 and Motu in the attic

Best regards

Michael

[oblivion]

Very interesting thread!

I hope Access is reading this.

[Totty]

Yes it is! I'm conscious that there may be 2 separate issues are here.

1: The Midi Timing of the Virus TI as a standalone midi Synth.
2: The issue I mentioned regarding heavy patches and fast modulation/envelopes tripping up the sound.

I'm inclined to think they may be one of the same. I'm speculating here but if there are DSP resource issues as a standard midi synth this could be alleviated to an extent by loosening the sound reaction time so that lots of notes don't start exactly together. I know midi is a serial protocol and no 2 notes can sound exactly together, but by reducing reaction time would give the hardware more processing time (not good!).

With Virus Control the data has to appear out sample accurately apparently. This would disallow the theory above, but since I have experienced these modulation and envelope discrepancies, that could also be as a result of DSP resources lacking.

I'm happy to be proven wrong - in fact I hope I'm wrong, because as it stands I don't feel I can make a complex sound and have it play back correctly unless I'm using only one channel.

[paulkeeley]

hey all - once again, another development on my end. with the patch in question, running in MIDI standalone mode, i noticed that if i adjust the amplifier envelope release time to 20, where there is no note overlap, the timing is relatively tight. not as tight as VC mode, but much tighter than before. however when i bump it up to 36, where there is slight note overlap, the timing goes awry. yet another clue that it may be the way the virus is handling note information and note stealing.

i've sent this latest development to Access support. hopefully they can shed some light on the subject. if anyone has a chance to try the same out on their setup it would be greatly appreciated. if this isn't just an issue with my virus, then this is something hopefully Access can fix.

cheers,
Paul.

[oblivion]

Have you experienced any problem with ordinary patches = Bass, Trance sounds, strings and pads?

Hope there is only problem with complex patches.

[synthfiend]

As a matter of interest Paul, when you are using the Virus in VC mode are you using the Virus as your sound card or are you using the sound card on your PC ?

Out of all the different ways you can route audio and midi with the TI I am starting to suspect from what I read and have experienced that you get the best performance (latency wise etc) when using VC with the TI as your sound card. Maybe wrong though

[seantyas]

Originally Posted by paulkeeley

aaaaaaand since i clearly haven't posted enough today on this topic, i've made a bit of a breakthrough.

in VC mode, everything is rock solid. it's only in MIDI standalone mode where the timing goes to hell and back. again, no problems with my other synths whatsoever using the same MIDI interface(s), cables, etc.

so, it must simply be the way the TI is handling MIDI input info. or at least that seems to be the case with mine.

hope this helps in any way, however insignificant.

cheers,
Paul.

odd that ur problem only occurs in midi mode.. mine is both present in VC AND midi modes, also present whether i use TI as my interface OR my m-audio, ALSO when i use analog outs (all outs).

BUT here is one interesting tidbit... DOESNT happen for me in ableton. plays PERFECT there. Something is up with the connecttion btwn logic to leopard or logic to VC or both.