Reason, RME, PTle and midi timing problems

[bennals]

I'm currently routing Reason's output via an RME 96/8 PAD lightpipe into the 001 and PTle. I've got the asio buffer set at max (2048) to avoid digital artefacts. I'm trying to eliminate (or at least reduce sufficiently) the latency between the position of the midi notes and the recorded audio by measuring the space in samples between the two. The first attempt yielded a latency of about 2800 samples. After adjusting the global midi offset in PTle preferences accordingly, recorded audio can still be out by as much as 1200 samples! And, both in front and behind! This amount of mistiming is definately audible, and really annoying to have to go in and manually edit. Using this same method with the Tokyo synths, I was able to get the timing to within around 2 ms (more than satisfactory for my needs), but that was routing audio via Direct Connect. I'm thinking then that the inconsistent timing is related to the RME card (changing latency amounts from note to note). Any suggestions, help, or even similar stories out there? Thanks in advance.

[Chompers]

I have never monitored any midi latency with my RME card. But i seems to be around tiny, so small that I dont even need to nudge after a recorded pass.
Im not clear on where the latency is occurring. Are you playing midi notes in REason and routing the audio to Pto tools. Or are you recording midi in pro tools and routing the midi to reason and then back to Pro tools. OR.... Are you recording midi in resaon and routing midi data to Pto tools. There are so many options that clarification would be alot of help

[bennals]

Ok Chompers, I hope you have the patience to sift through this lengthy post. To clarify, I think its the latency that exists between Reason being triggered from a PTle midi track, and the resultant audio data being routed via the RME 96/8 card into the 001 and a PTle audio track.

I'm sending midi data from the sequencer in PTle via OMS (2.3.8 I think) which triggers the Reason devices, the audio of which then gets routed back to the 001 via adat lightpipe to a combination of mono and stereo audio tracks (eight in total). I figure whats the point of having the RME card if you can't get eight channels of audio from Reason into PTle, but thats by the by.

Since my first post I've done some tests on the timing accuracy of the audio coming from Reason via the RME card into PTle.

With the RME buffer set to 256 samples in the ASIO section of the DIGI 96 settings control panel, the accuracy is pretty good. With a global midi offset of -400 samples in PTle, I measured the recorded audio between 2 samples and 230 samples later than the corresponding midi notes position (in a PTle midi track) over quite a few passes. So whats that, something like 1 to 6 ms variance. More than adequate for my needs.

But the problem with the 256 sample buffer when several of the Reason devices are generating audio is the audible clicks and pops which you referred to in a previous post. So, crank the buffer up to 2048 samples, and the noises dissappear, but the recorded audio becomes wildly mistimed with the corresponding midi notes. I was using a global midi offset of -2200 samples, and the recorded audio ranged from being 577 samples early (before the midi note) and 1254 samples late (after the midi note). This amounts to over 1800 samples of inaccuracy, or 41 milliseconds (hope my calculations are correct) which is really audible for a percussion track with sharp attack.

I would love to be able to compose whole songs using only the PTle midi sequencer to trigger the various Reason devices in real time, then when satisfied with a rough arrangement, record the various Reason parts (eight at a time of course) onto PTle audio tracks for mix eq, dynamics, effects etc, but this is impossible due to the digital artefacts caused by the low asio buffer setting. I'd also love to do this without having to manually correct the recorded audio's timing, but this seems impossible at a high asio buffer setting. Can you see the dilemma? I hope I haven't confused the issue more, and would really appreciate a response Chompers. You seem to have your head around this type of configuration. Cheers.

Bill.

[bennals]

Chompers, you out there?

[bennals]

ok, one last bump. Hope the long posts haven't put you off replying Chompers.

[Chompers]

Hmmmm. I never use Pro tools to trigger reason instruments, so Im not the best guy to trouble shoot. However, I am about to do exactly what you are talking about, cuae I have a mix that has midi data in pro tools. So Ill soon see what the deal is. Ill post later tonight (Jan, 20 ,2002) with the results.
I always recod midi data directly in Reason, and route to pro tools. Works great. WHat is the advantage of doing it in pro tools?

[tommydee]

yeah, why use PTLE's midi??? Reason's seems so much tighter and time accurate....

i'd do what chompers does, and as soon as i can afford one of those RME things, then i'm all there!

peace,
t.

[Chompers]

I mixed a song that had Pro tools midi data sending midi to Reason. I didnt experience any latency out of the ordinary. THe Pro tools midi tracks triggered reason and reason sent back the audio all within 3 or 4 miliseconds.
Im not sure what your issues are. Perhaps If I was using many midi tracks the latency would increase. Im not anything close to a midi expert. I use it sparingly.

[bennals]

Thanks for the replys guys. I guess the main reason for using PT le to send midi data is so I can keep all my audio and midi track information visually contained on the one edit screen in PT. I do use a lot of midi tracks, and a lot of automated controller data too, so maybe it is asking to much of PT le, but my main issue (above) was the dramatically increased inaccuracy of the recorded audio from Reason (in relation to the position of the midi notes in PT le) when using the highest buffer setting (2048 samples) on the RME card. I'll try using the internal sequencer in Reason to see if there is an improvement.

Do you ever run all eight lightpipe tracks from Reason into PT le at the same time Chompers? And with that track you mixed with midi data in it, which buffer size on the RME card were you using?

Cheers, Bill.

[Chompers]

I often run all 8 channels from reason to PT.
I use the largest buffer , 2300 or whatever it is.
The latency is barely a factor. I run pro tools and reason at the same time, and at the end of the session record the Reason tracks into Pro tools. I never nudge any thing. ALl the audio and midi just falls right into place. If I examine the latency it is tiny ( 2 or 3 miliseconds) I havent set any buffers to anything other then the way they come pre set.
Im amazed at the RME card and how it performs.