Comments

midigod wrote on 7/15/2003, 1:56 PM
We'll need a bit more detail.....
pwppch wrote on 7/15/2003, 5:11 PM
This is a known issue and will be addressed in the next update to Vegas 4.0.

Peter
CDM wrote on 7/15/2003, 11:33 PM
I'm sorry... what's the issue?
pwppch wrote on 7/16/2003, 6:46 AM
If I understand correctly, when recording audio the new recording can be delayed on the time line relative to exist tracks when recording using using Wave Classic Drivers (or the Direct Sound Surround mapper).

ASIO defines the offset between input and output so the issue does not manifest itself, though depending on the ASIO driver there can be recording offsets.

Also, the manual record offset is not persisted correctly and can be ignored in some scenarios.

Peter
akg wrote on 7/16/2003, 10:12 AM
It`s big bug in V4 !!
I think Vegas will add more oil in audio !!!
Not only on video !!!
CDM wrote on 7/16/2003, 10:44 AM
thanks Peter
Cold wrote on 7/16/2003, 12:31 PM
Is there a work around for this that is effective? I'm running a motu 2408 that is not happy with the asio drivers and the track delay makes vv4 currently unusable. I have time today to experiment with alternate (pci 424) motu drivers, but just wondering if anyone else has found a usable work around.
Thanks
Steve S
Geoff_Wood wrote on 7/19/2003, 4:20 AM
I wonder what they broke ? The PCI324 ASIO drivers seem fine ...


geoff
Jacose wrote on 7/19/2003, 11:48 AM
i am sure i should get a new soundcard, but admittedly, this makes vegas 4 literally unusable for most projects on my system....
pwppch wrote on 7/19/2003, 2:14 PM
This is a Wave Classic/Mapper problem only. ASIO defines exactly the input to output latency/offset (though some drivers are better than others in providing completely accurate times.)

With ASIO you should never see more than a few tenths of a millisecond offset.

Peter
bgc wrote on 7/21/2003, 1:07 AM
Hi Peter,
By a "few tenths of a millisecond" offset do you literally mean .1 msec or .2 msec offset?
I'm seeing 2-4 msec offsets with ASIO on my Card Deluxe (yeah, that damn thing). Totally usable and what I was seeing with the Classic Wave Driver setting before the recent V4 issue (which I know is being addressed in the next rev.)
Thanks,
bgc
pwppch wrote on 7/21/2003, 9:52 AM
2-4 ms with an ASIO driver it not a good thing. They should (can) be spot on. Will have to give the CardDeluxe a try and see what is up.

Yes, I did mean .1 ms.

When the update is released we have permited even ASIO drivers to be "tweaked" so that you can get rid of any record offsets. It will still be at a granularity of ms vs 10ths/100ths of ms.

Peter
Cold wrote on 7/21/2003, 1:16 PM
Has anyone else run into issues using the motu pci 324 + ASIO driver with VV4? Or is this my own system issue? I've had this across two totally different machines now.
Steve S.
bgc wrote on 7/21/2003, 5:53 PM
Hmmm, let me know how it goes with the CardDeluxe Peter.
bgc wrote on 7/21/2003, 6:04 PM
I found an interesting paper on latency here:

http://gigue.peabody.jhu.edu/~mdboom/latency-icmc2001.pdf

It's interesting that NONE of the tested systems provided < 1mes (.001 seconds) of latency, even with ASIO.

Brett
pwppch wrote on 7/21/2003, 9:24 PM
This is a different type of latency than what is happening here. This article covers input to output monitoring. In general I would say that the paper is accurate, though a bit outdated.

The issue that this thread is about is how accurately can input be aligned with output. The ideal test is to record the output of the DAW back into the DAW.
Output sample frame sizes should not be a consideration here if the driver model permits accurate measurement of the input to output offsets.

Wave does not have a mechanism to do this, so the host must attempt to time stamp input buffers against output buffers. The problem with this method is that the time it takes to do the measurement is difficult to account for. This results in a best average and inconsistencies in the results from recording to recording.

ASIO drivers by definintion permit the driver to tell the host the exact offset between the input buffer with relation to an exact sample in the output buffer. Depending on the driver, this can result in either absolute positioning or "very close" positioning. The best part is that even in the "very close" case it is always exactly the same and predictable.

Peter
bgc wrote on 7/21/2003, 11:52 PM
Hi Peter,
The test you site, record the output of the DAW back into the input is the test I've been doing and getting 2-4msec latency with the Card Deluxe in ASIO mode.
Brett
pwppch wrote on 7/22/2003, 8:52 AM
Are you saying that for a single sample frame size setting with the ASIO driver the record latency offset is variable from 2-4 ms?

Peter
bgc wrote on 7/22/2003, 3:07 PM
Hi Peter,
The Card Deluxe has a minimum setting of 1ms for ASIO. Regardless of this, I can't use settings of less than 25 msec without sputtering and stuttering and skipping.
Brett
pwppch wrote on 7/22/2003, 9:08 PM
1 ms is very tough to handle. 25 ms is a bit high. I can run the CardDeluxe at 4-8 ms on an average project.

I did a simple test on the recording the output to the input with the CardDeluxe. I was seeing ~1 ms record offset regardless of the sample frame size I set for the driver. This is consitent with the behavior of most ASIO drivers.

Peter
bgc wrote on 7/23/2003, 1:17 AM
Thanks for doing a check on the Card Deluxe Peter, I really appreciate it! To be honest, for me 2-4 ms offset is perfectly fine for recording - heck MIDI delay is longer than that. I will be interested in trying the new offset compensation for ASIO in the next rev though.
Geoff_Wood wrote on 7/23/2003, 2:49 AM
Recording offset on the timeline is not quite the same thing as 'latency' though related..

geoff
bgc wrote on 7/23/2003, 12:54 PM
Geoff -
Yes, I'm constantly confused by that and have more than once confused the two.
At least it'll be nice to have a slider to play with to see if it makes any bit of difference with my system :)
bgc