Please Fix the midi Timing

[WernerF]

What have you guys done to test for the issue. Have you printed a pass of a midi triggered instrument and then compared the audio file to the midi notes? Midi is still far from sample accurate. Apparently it has to be midi time stamped in order for it to be sample accurate. Your ears only are not a good test. Way too subjective.

[dankin]

Pt has always had horrible midi timing with 3rd party VIs. I was sure hoping this would be fixed with 11 and AAX.

[RyanC]

Are you all testing this with phase flip, how are you going about it?

VEP timing is not swimming here, I'm able to get a complete null from a printed track. If I printback from a bus in PT, I do end up having to adjust the ADC manually.

The only place where I could get the timing to mess up at all is when it's looping, the very first note of the looped part would not null %100.

[Rail Jon Rogut]

I helped Slate's programmers in Russia fix the timing of their MIDI in SSD4 a few years ago with PT 10 when we ported our MIDI Grooves to SSD4

If your timing isn't right then

A) make sure you have an updated build of SSD4

B) In your Playback Engine settings in PT turn off Ignore Errors

Rail

[WernerF]

Is the dicy timing strictly due to the VI's limitations or is the DAW part of the issue as well? Can this fix be applied to other VI's and, if applicable, to DAW's as well. Personally I'm not a huge fan of the uber slick LA sound of the SSD4 samples, but love his other products, so this individule fix does not apply in my case.

[mykhal c]

Originally Posted by WernerF

Is the dicy timing strictly due to the VI's limitations or is the DAW part of the issue as well? Can this fix be applied to other VI's and, if applicable, to DAW's as well. Personally I'm not a huge fan of the uber slick LA sound of the SSD4 samples, but love his other products, so this individule fix does not apply in my case.

what about Rail's B step? was this unchecked all the time?

[DAWgEAR]

I have not heard any MIDI timing issues.

Can those of you noticing MIDI timing issues provide a repeatable and objective test that the rest of us can use to see what you are experiencing?

I just ran some very quick tests of my own in response to this thread:

I drew in 4 MIDI notes on an instrument track at exactly 100,000, 200,000, 300,000 and 400,000 samples (and I created markers at these time locations). Then I loaded Xpand2 with a hihat sound and routed the output to a bus and then I routed that bus to an audio track. I recorded the audio and examined the number of samples of delay between the original MIDI event and the recorded audio. It was consistently 11 samples (at 96 kHz), so that is about one tenth of a millisecond. The timing delay was constant and essentially insignificant.

I then repeated the same experiment with EZ drummer, again trigerring a hi hat. The sample delays were 28, 22, 27, and 33 samples. Again, well under half a millisecond. I assume that the delays differed because of round-robin sample triggering and that they were a bit longer than with Xpand2 because EZ drummer uses samples and they do not start right on the initial transient.

Based on my admittedly simple tests, my conclusion is that there is nothing to worry about with respect to MIDI timing with delays of well under a millisecond.

My testing methodology was very simple, but it is something anyone else can repeat and confirm/refute. As I mentioned, if you have something concrete that the rest of us can test and repeat, or if I have missed something, please share.

[WernerF]

Originally Posted by mykhal c

what about Rail's B step? was this unchecked all the time?

No, it was not checked. I've never had that box checked, ever. Anyone who does is choosing to ignore the warning that it may cause clicks and pops.

[WernerF]

Originally Posted by DAWgEAR

I have not heard any MIDI timing issues.

Can those of you noticing MIDI timing issues provide a repeatable and objective test that the rest of us can use to see what you are experiencing?

I just ran some very quick tests of my own in response to this thread:

I drew in 4 MIDI notes on an instrument track at exactly 100,000, 200,000, 300,000 and 400,000 samples (and I created markers at these time locations). Then I loaded Xpand2 with a hihat sound and routed the output to a bus and then I routed that bus to an audio track. I recorded the audio and examined the number of samples of delay between the original MIDI event and the recorded audio. It was consistently 11 samples (at 96 kHz), so that is about one tenth of a millisecond. The timing delay was constant and essentially insignificant.

I then repeated the same experiment with EZ drummer, again trigerring a hi hat. The sample delays were 28, 22, 27, and 33 samples. Again, well under half a millisecond. I assume that the delays differed because of round-robin sample triggering and that they were a bit longer than with Xpand2 because EZ drummer uses samples and they do not start right on the initial transient.

Based on my admittedly simple tests, my conclusion is that there is nothing to worry about with respect to MIDI timing with delays of well under a millisecond.

My testing methodology was very simple, but it is something anyone else can repeat and confirm/refute. As I mentioned, if you have something concrete that the rest of us can test and repeat, or if I have missed something, please share.

All Avid VI's are midi time stamped. This makes them sample accurate. Just try the same test with Ivory, B4 or any Arturia VI. Also, you might want to try and trigger Slate Trigger with midi and do the same test. That is even more wildly erratic.

[DAWgEAR]

Originally Posted by WernerF

All Avid VI's are midi time stamped. This makes them sample accurate. Just try the same test with Ivory, B4 or any Arturia VI. Also, you might want to try and trigger Slate Trigger with midi and do the same test. That is even more wildly erratic.

Thanks for that info. So, the Avid VIs are exempted.

I do not own any of the other third party VIs you mentioned. The only VIs I own that are AAX 64 ready are those from Toontrack. I will test further with those.

I also have VEP5 which is AAX 64 ready. Are VIs hosted in VEP5 subject to this issue? I will test that as well with Native Instruments, Spectrasonics, etc.