PT7 and midi timing?

[Eugene]

I`d like to ask ppl, who`ve already upgraded to PT7 and who had all that midi-rtas timing issues with v.6 Is it any better mow? Does it have same problems with timing which were duscussed to death here on DUC?

Thanks in advance
Eugene

[nikki-k]

Hmmm...

MIDI notes getting stuck.
This is a tough one. With USB keyboard controllers, there seems to be a higher chance of stuck notes while playing, especially during flurries of notes (64's, 128's, etc). With a keyboard via a MIDI interface (yes, a USB one- go figure), things seem ok. For USB keyboards, it seems the older M-Audio ones (such as Oxygen-8) have more of an issue (lmao... ). I use a Kurzweil K2500XS via Digi MIDI I/O (USB) on it's own internal usb hub (most motherboards have 1-3 usb "hubs" and up to about 8 ports via these "hubs"; peole often confuse port and hub). Happy with that! Oh- Absynth 3 seems to get stuck notes more than any other synth/sampler. But, it is much better than 6.9 dreamed of being!

MIDI Timing:
I have several sessions with heavy drum and synth use. Rock music, and doing compositional scratches using Kontakt2, Indigo, and often B4, and other times CS80, Minimoog, FM7, and a few others. I dont se/feel/hear any timing issues that have stopped me in my work/play. If something sounds too far out, I simply nudge the region, or offending notes. I have noticed that every once in awhile I find a passage that just does not "feel" right, and I know it should. Selective nudge, or Real Time Properties adjust for that region (or track) have solved it for me. I have not run into problems rendering to an audio track yet. Then again, I may be having better luck for some reason.

Any other MIDI issues I am not aware of. All in all, I find PT7 HD great, especially with MIDI. Easy to use, and with real time properties, and sample based MIDI, things are damn cool. IMO, the Region List needs ALOT of attention for the next release. It is simply too chaotic/crowded now, and requires serious attention.

[insanesounds]

Hey Eugene ... The timing is a little better, but it still is not what it should be, unfortunately - not like it was at 6.0 or so. I dont really use RTAS instruments, but I do send midi out to various keyboards and drum machines, as well as gigastudio and kontakt on another computer connected to protools via litepipe and there are definite timing issues. Most noticeably on percussive intruments. Recorded audio from these tracks vary randomly from the original position of the midi notes. Beat detective has become an indispensible tool. I'm curious what DIGI says about this issue. It really is to the point that I might have to invest in a separate midi program although that will make recording audio and midi cumbersome during the preproduction stage.

Otherwise extremely happy w PT 7.

[Eugene]

Thanks for replies.

That's really sad to know. I hoped for this update will fix it.
Hanging notes... Is it so hard just to add "reset midi on stop" in preference? All other sequencers have it for years.

I'm curious what DIGI says about this issue.

The last i remember was - "Midi issues? What midi issues?"

[nikki-k]

Yes, it is really suprising they dont have that "all midi off/stop" on stop option. I guess I have become so used to using workarounds for certain things that I simply forget them!

Stuck notes: as I said earlier, I am simply not experiencing this that much. Certain sessions seem more prone to it tho, which is odd to me. Odd because I cannot narrow it down to a certain combo, certain synth/sa,pler, or size of session (including edits, automation, etc amounts).

MIDI timing...wow. I guess I need to render stuff and compare across some diverse sessions then. I have heard little sections that seem "out of time," even as much of a human feel I utilize (I record real time, and only quantize in the rarest case, or for dance type music; manually moving a note here and there tho...well, it happens ). Will try some things and post back...

What would be quite interesting is to find out if it is consistent timing problems. In other words, are ALL passages out; if not, when bad timing happens, will the rendered MIDI passage show identicle results every time if rendered say, 5 times? Does the bad timing occur in the same spot every time? If 10 people load, play and render a MIDI passage to audio on 10 different systems, is the bad timing present for each? To the same degree? Same places?

I guess what I am getting at is- is there any "structure" to the bad timing, or is it simply and completely random timing issues? Is there something one can do to guarantee a timing mishap at a specific point in a song? Is there a session that uses only common items, is not that large, that will produce horrible results on any HD machine?

[insanesounds]

Unfortuantely, the MIDI issues are not consistent. Each pass of rendering will put the recorded audio at different spots in the timeline; slightly after the midi note, before it - and that will vary too! It's crazy. It would be nice not to render, but with layering kicks and snares and whatever else, the timing's gotta be near perfect - otherwise these percussive instruments are flamming (drummer's term)and flanging all over the place. Plus - I hear the timing errors in a simple playback with a bassline following a simple drum pattern, forget the layering and all that.

Anyway, I love PT but this MIDI problem is inexcusable, especially since it did work great in earlier versions! Like I said b4, I think it was from Ver. 6.4 and later that it started acting up.

DIGI, if you're listening, it would be nice if you at least acknowledged that this is an issue and that a fix is in the works. Thanks for the love.

re: Nikki's post... Test it out if you get a second. Trigger a kick - a measure or two - and then record the audio. Put the midi track right next to the audio track and look at the discrepancies in all of their glory. Then do the same pass again and examine the different differences! It would be great to have someone corroborate my findings. Thanks...

[DigiTechSupt]

insanesounds - what you're describing is normal, as long as we're talking about differences of a few milliseconds either way.

MIDI is a serial protocol and the timing can vary depending a number of factors - how many notes are being played simultaneously, the response time of the synth, the type of MIDI interface being used, etc.

One thing you mentioned - that sometimes the audio note will show up on the timeline BEFORE the MIDI note that triggered the sound. If that is happening, we'd be very interested to see it because that would signify something is definitely wrong.

The fact that it 'worked' in 6.4 is not a conclusive comparison unless you are using the exact same gear, same computer, same OS, etc. as all of those can have an effect on MIDI timing.

[HotRats]

Unfortuantely, the MIDI issues are not consistent. Each pass of rendering will put the recorded audio at different spots in the timeline; slightly after the midi note, before it - and that will vary too!

I'm an LE user (7.0) and I've got the same issue since ever (pt 6.4).
Every time I render midi tracks I have to zoom in and nudge every single track.

My Old post on this forum
http://duc.digidesign.com/showflat.p...fpart=all&vc=1

I paste here a part of that post:


*************************************

Last night I've made a simple test.

I've recorded 3 audio tracks from 3 aux tracks using 3 different versions
of the same insrument:

1 aux: kompakt RTAS
2 aux: kompakt VST > RATS
3 aux: kompakt VST > REWIRE.

on each kompakt i've loaded the same drum set.
each kompakt was triggered by the same midi track (1/4 snare hits).

I've recorder the output of the three aux at the same time using
different H/W buffer settings (68 or 1024), with or with out
global midi offset.

the result is:
kompakt RTAS have always about 1000 samples latency without any realtionship
whith H/W buffer size

kompart VST > RATS and VST > REWIRE are "latency free"
(1 sample using a 68 samples H/B buffer size)

if I use (and I need to do it) a mix of RTAS and wrapped istruments, global midi offset is useless

conclusions?
1. do always test performance of RTAS instuments before useing them!;
2. never use a mix of instruments formats;
3. VST converted instruments are much better than RTAS ones.

[nikki-k]

OK- I went ahead and played around with RTAS and TDM instruments, and used percussive type patches (perc organ, toms, cowbells, etc.). Not using any huge sessions (10-30 tracks), and not ny other RTAS. 44.1/24, HW buffer @ 128, 2 processor (HT enaled P4), DAE lev 2. Nothing out of the ordinary. Tried stereo and 5.1 sessions.

Routing out from stereo RTAS fueled Instrument track bussed to Audio track, Delay Comp off:
All is well. Or so it seemed... Upon using Tab to Transients, and swapping from Samples to Bar/Beat timelines, it becomes apparent very quickly that random offsets are occuring. The printed audio has the transient start occurring anywhere from a few samples late, to about 100 sample slae with the same instrument. ith ADC of, I would expect a latency, and a consistent one. However, this is NOT the case! It is a random amount, and it seems as though it is far, far less latency than the plugin itself should produce (260 samples in the example I will cite below). here is this randoness coming from?

Turn ADC on, and I am getting random timing issues also...or so it seemed.
The RTAS I will use as an example incurs a reported 260 samples of delay. With 2 samples out, and 2 samples back in, that leaves 256 samples by the plugin itself.
Record a pass, and then position the cursor at the first recorded "pulse." It is dead-on, and the second was the same. But- the rest of the beats were 246 samples EARLY. No variations, no randomness...246 samples every time! (BTW- PT7, no RTP engaged, either region or track based)

If someone has a session that has no audio, and uses a common RTAS plug (any NI, Arturia, or Lounge Lizard for example), and no audio and consistently has MIDI timing problems, I would love to load it up and see what happens here. There is a problem, but I am just not experiencing a drastic problem here- at least not if I turn ADC on, and then render and nudge after. With ADC on, at least it is a consistent issue.

Yes- there is a problem...or so it seems here with just a casual set of tests.

[nikki-k]

OK- way back when the VST-RTAS wrapper came out, I had a forum going, and had a thread dedicated to it. I (and others) reported this problem, and I reported the latency of RTAS vs Wrapped.

With PT7, it is easy to compensate:
Using real time properties, you can offset a MIDI/Instrument track appropriately, without affecting anything else. Further, you can even do so on a region-by-region basis. Very slick, and works quite well.

Irregardless of the MIDI instrument- virtual or hardware- there is a latency. Figure it, and use RTP to compensate. hen done composing, you can always render to audio, and then nudge as needed. With HD, and ADC, the valueseems more consistent at this time, and therefore compensation after a render is quite a bit easier, if one is being exact.