Latency still high w/DC and slop is almost unusable

[CCash]

I'm going to try this forum now for this question. The others move so fast my post keeps biting the dust. Here's the thread from the 001 forum with Brent's reply. I realize latency is expected, but up to 27ms (before MIDI offset)? And 10ms of play between recorded notes? Any tips would very appreciated.

CCash
Member
Member # 7299
posted January 21, 2002 07:34 PM
--------------------------------------------------------------------------------
Replaced 5.1 direct connect and stream manager with cs5.1.1 (the latest "low latency" set on the downloads page).
To test it, I have a track with MIDI data routed to an aux with Soft SampleCell on an insert (direct connect).

I bus that aux to an audio track, record, and measure the offset.

The latency ranges anywhere from 17 to 27ms, regardless of direct connect and stream manager versions (2 in the download). First, that's a lot of latency. Second, that's 10ms of slop.

There are no interfaces or MIDI instruments in this equation -- just PT, SSC, stream manager and direct connect. In the Samplecell Setup menu I DO see the Sound Manager output latency stating 2.9ms and DC at 11ms. Add that to the 001 host latency and 17ms seems normal, not 27. I was getting these same numbers with the original version of Direct Connect. And what about the slop? Am I doing something wrong? Thanks for any help.


--------------------------------------------------------------------------------
| IP: Logged

DigiTechSupt
Administrator
Member # 7660
posted January 22, 2002 01:06 PM
--------------------------------------------------------------------------------
When recording the audio output of any soft synth in Pro Tools via DirectConnect, you are faced with two latency values:
1-DirectConnect
2-PT LE Hardware Buffer Size
In order to combat the latency values when recording to disk the audio output of your MIDI gear (hardware or software), you will need to employ a negative MIDI playback offset.

Brent


--------------------------------------------------------------------------------
| IP: Logged

CCash
Member
Member # 7299
posted January 22, 2002 10:49 PM
--------------------------------------------------------------------------------
I have been using a negative MIDI offset and a buffer size of 128. Could you clear this up for me though?

1.) Shouldn't I also see an improvement in the latency on the recorded audio track using version cs5.1.1? Or is the difference only in monitoring the MIDI's Aux Input? (or in Sound Manager using that option).

2.) When I record quantized MIDI notes as audio, does 10ms of play (slop) on the audio track seem normal? In this case I'm only triggerring SSC with prerecorded MIDI (it isn't live). This is pushing my total latency values up to 27ms.

I guess I can deal with the latency with MIDI offset, but I'd really like to tighten up the slop. Anyone out there using SSC/001 have some numbers so I can find out if this is normal? Any tricks for lowering the latency/slop? The manual's only tip is that having too many MIDI Input Devices selected can have a negative effect.

[Nick Batzdorf]

What MIDI interface are you using? I had problems with the serial Unitor 8 and Soft SampleCell (in a G4 Port), but they went away with that interface networked to a USB AMT 8 or with the USB MOTU MTP AV.

The obvious way to isolate the problem to DirectConnect or other software vs. the MIDI interface is to test SSC with Sound Manger.

[CCash]

Thanks for the reply Nick. Maybe my lack of MIDI understanding is coming into play, but I can't understand why my interface would have anything to do with timing *if* I'm only comparing the offset between existing MIDI notes on a track with their respective recorded audio (not the slop between the MIDI notes themselves on a grid). Does DC's timing somehow rely on the interface/OMS even when you're only playing back MIDI and recording it as audio?

At any rate, I'm surprised by what you're saying, since I went with a Studio 64XTC serial to G4port based on all the issues I read about with USB interfaces.

I'd really like to get this resolved because going in and nudging all the audio notes around is counter-productive.

[Nick Batzdorf]

No, in that case it's not the MIDI interface. The problems I was having were on input. I can guess, but I don't know why the MIDI timing would be so far off in that case.