latency on VI's vs. buffer settings

[Judson]

question: is -playback- latency of virtual instruments affected by the RTAS buffer setting? like if I change the buffer setting from 256 to 1024 (with delay compensation on), will all my VI's play a touch later? Thanks much-

[Phil Buckle]

I never thought so but there was a recent thread started by 25 Ghosts who claimed that the midi "slop" increased with different buffer settings on playback.
That is to say that the timing wasn't any later or earlier, it was just "not tight" and "different".
Others agreed.
Worth noting .
I think it was also stated in the thread that the other major DAWS exhibited the same behavior.
So it's not exclusive to PT.

[Judson]

I've always done some post work in Pro Tools but reserved my production and writing for Logic...now looking at PT as a possibility for a composition tool also. Anyway, i started doing some testing with VI's and it seems that when I do raise the buffer setting the delay increases on VI playback. I'm recording a simple quantized backbeat back into PT and at 256 buffer setting the sample hits 144 samples after the MIDI note. if I increase the buffer, this delay increases but it doesn't seem to be by double. more like 50%. the delay seems to be consistent...but this is an empty session...maybe the slop starts as more data is being thrown around. albeit 144 samples is only like 4 ms at 44.1k i think which is basically nothing, but its unsettling that I'm messing with the timing of my VI tracks by changing buffer settings. in contrast, no matter where I have the buffer in Logic, VI's are essentially right on the grid along with the MIDI note when you bounce as audio. It seems like "ADC" should make the VI's playback sample accurate no matter what the buffer setting is.

Can anyone confirm this behavior?

[Phil Buckle]

I think 25 Ghosts did some comparisons with Logic.
Hey 25.......

[Naptalene]

Im on 8.4 cs2.

I run as low as possible on buffers and lay audio when glitching begins.

VI's seem very variable in terms of performance. Delay comp.really affects the handling of units especially PLAY by East West. I've gone insane trying to lock down a specific cause for it. Silly things like an arp. being triggered on Expand real time will be completely out (with no referance to the players "out of timeness") until you quantize exactly to the beat.

It also sux in terms of system utilization of VI's compared Logic. My HD Accell 2 is a lumbering pig (on a 16 core Mactel with 16 GB Ram) compared to Logic. I'm hoping 9HD helps out with this.

You need to be REALLY careful of routing when laying stuff in with DC. I often have to have a "Music Master" within the session and the VO's etc running direct to my system Master. If you have DComp. active and lay a vocal that isn't routed through the Music master, and the master has a plug based latency, you're new vocal will be the latency's amount ahead..........it's fooked but true.

Bit off topic but it's only when you run a post AND music track in one that these issues really arise due to the "client wants another VO take" factor and Judson seems to be in a similar environment.

[dtmprod]

I always shut ADC off on the instrument strips when using VI's. Seems to have accurate timing in that mode. Any reason why you'd wanna have ADC on when tracking VI's?

[Naptalene]

So you kill it per channel, not on the entire session? Dammit.....must have missed that option and here's me shooting off about them. I'll check it out thanks.

Going back to VI's with ADC comes up when you've had to go back and re-lay because of client changes. I often end up with a final post mix and music session in one as no-one will make a call on either, and they insist on being able to tweak both at the same time. Kill ADC and some of your sync goes to pot.... any other ideas on doing it?

[Naptalene]

I can't belive I missed that stupid bypass ADC...... THANK YOU