Hi guys, I just recorded a click track on my Protools TDM system and
there's a 300 sample discrepancy between where the midi note originally
is and where the recorded audio clicks actually land. I had no idea it
would be this great of a lag.
The recording chain is as follows:
Quantized midi note to MOTU Pocket Express midi box to a Proteus/1 box,
direct out into a Protools analog input, recorded onto a PT track.
Is this a normal amount of delay? How do people deal with that? I
suppose if all the analog recording is done after all the midi is
finished, it wouldn't matter. But you wouldn't be able to snap any
analog notes to grid or they'd be 300 samples ahead.
I'm thinking of recording all midi parts including the click into
analog and then sliding them 300 samples earlier so that they land
right on the grid. And then recording the real analog parts to that
click.
I'm curious how y'all deal with this. 300 samples is huge! Or maybe
something's wrong with my system.
Thanks, Rick.

most of thats midi delay - basically what you do is record a click - just
1 - place it at measure one beat 1 and copy/paste to the next beat/measure
and so on.


"rickymix" wrote in message
oups.com...
Hi guys, I just recorded a click track on my Protools TDM system and
there's a 300 sample discrepancy between where the midi note originally
is and where the recorded audio clicks actually land. I had no idea it
would be this great of a lag.
The recording chain is as follows:
Quantized midi note to MOTU Pocket Express midi box to a Proteus/1 box,
direct out into a Protools analog input, recorded onto a PT track.
Is this a normal amount of delay? How do people deal with that? I
suppose if all the analog recording is done after all the midi is
finished, it wouldn't matter. But you wouldn't be able to snap any
analog notes to grid or they'd be 300 samples ahead.
I'm thinking of recording all midi parts including the click into
analog and then sliding them 300 samples earlier so that they land
right on the grid. And then recording the real analog parts to that
click.
I'm curious how y'all deal with this. 300 samples is huge! Or maybe
something's wrong with my system.
Thanks, Rick.

"rickymix" wrote in message
oups.com...
Hi guys, I just recorded a click track on my Protools TDM system and
there's a 300 sample discrepancy between where the midi note originally
is and where the recorded audio clicks actually land. I had no idea it
would be this great of a lag.
The recording chain is as follows:
Quantized midi note to MOTU Pocket Express midi box to a Proteus/1 box,
direct out into a Protools analog input, recorded onto a PT track.
Is this a normal amount of delay? How do people deal with that? I
suppose if all the analog recording is done after all the midi is
finished, it wouldn't matter. But you wouldn't be able to snap any
analog notes to grid or they'd be 300 samples ahead.
I'm thinking of recording all midi parts including the click into
analog and then sliding them 300 samples earlier so that they land
right on the grid. And then recording the real analog parts to that
click.
I'm curious how y'all deal with this. 300 samples is huge! Or maybe
something's wrong with my system.

That's well under 1/100th of a second if you're recording at 44.1k, and less
than 1/300th of a second if you're at 96k... I realize some people are more
time-sensitive than others (and I'm one of 'em, for better or for worse - I
have a hard time using VSTi's in overdub mode, at even very, very slight
degrees of latency), but do you really think that THAT small of a lag will
make any kind of perceptible difference?

Neil Henderson

rickymix wrote:
Hi guys, I just recorded a click track on my Protools TDM system and
there's a 300 sample discrepancy between where the midi note originally
is and where the recorded audio clicks actually land. I had no idea it
would be this great of a lag.
The recording chain is as follows:
Quantized midi note to MOTU Pocket Express midi box to a Proteus/1 box,
direct out into a Protools analog input, recorded onto a PT track.
Is this a normal amount of delay? How do people deal with that? I
suppose if all the analog recording is done after all the midi is
finished, it wouldn't matter. But you wouldn't be able to snap any
analog notes to grid or they'd be 300 samples ahead.
I'm thinking of recording all midi parts including the click into
analog and then sliding them 300 samples earlier so that they land
right on the grid. And then recording the real analog parts to that
click.
I'm curious how y'all deal with this. 300 samples is huge! Or maybe
something's wrong with my system.
Thanks, Rick.


This amount of MIDI delay with a Pocket Express plus a Proteus 1 is not
at all surprising. I would guess that you are recording at 44.1. Nothing
wrong with your system.

It could have a subtle effect on the feel. I don't work with combined
audio and MIDI any more, but I would slide the click back if I felt
psychologically uncomfortable with the delay, and as you say, record the
MIDI parts.

Probably good to record the MIDI parts anyway because they will always
be there on the tracks. (MIDI tracks can change if you subsequently mess
with the sound modules in some way - such as editing patches or simply
changing the volume knob.)

Also, if you keep the MIDI tracks as MIDI, you will likely double the
audible latency - once on recording and again on subsequent MIDI playback.

Here are a couple of relevant websites, btw:

http://www.soundonsound.com/sos/May0...cnotes0503.asp

http://gearslutz.com/board/archive/i...p3/t-3618.html

-Naren

Thanks y'all for confirming what I suspected to be the case; that it's
just midi latency. I'm glad it's not a fault in my system.

Neil wrote:
do you really think that THAT small of a lag will make any kind of
perceptible difference?

It's about a twentieth of a sixteenth note, and yeah, I think that's
huge! I'm not at all a stickler for things being mechanically perfect,
but I want things to play back exactly where they were performed. I
strongly feel that a crucial part of what makes a performance "magic"
is where it's placed in time relative to the rest of the track; how it
pushes and pulls relative to the groove. It's probably much more
important than slight pitch variations.
Or maybe that's just my bias after a decade of recording Motown's
Holland Dozier Holland. :)
Thanks again everyone, Rick.

At 44.1/48K rates, 300 samples is between 6 and 7 milliseconds.
Significant, but not major by any means. I've found that even highly
trained professional musicians rarely perceive delays of 15 ms or less
unless it's pointed out to them that they exist, except in cases of
highly transient percussion instruments like claves for example. I've
observed this through many sessions recorded through my MOTU 2408
which, as I usually have it configured, creates about that much
latency. Very rarely, a non-percussion player will notice, and it's
always an outstandingly gifted one when they do.

In a typical Pro Tools TDM rig, a "round trip" in and out of the system
without any plugins or extra bussing will cause about a 2 ms delay.
Thus far in my years of experience in that setting, no one has ever
noticed.

Others are probably right about the source of your delay being the
Proteus itself. Its own processor might be the culprit, or the sample
itself has a few ms of dead air at the front, or a little of each.

It does raise a related question in my mind about current midi gear
though, so I'll throw it out to the "room". I remember when USB midi
first came out (I'm still using serial midi with a Stealth port adapter
on my G4) and the latency issue was unworkable. I know things have
greatly improved since then, but has USB midi "caught up" completely?
Are there any remaining latency issues with it?

Thanks

Ted Spencer, NYC

In article .com writes:

Hi guys, I just recorded a click track on my Protools TDM system and
there's a 300 sample discrepancy between where the midi note originally
is and where the recorded audio clicks actually land. I had no idea it
would be this great of a lag.

What sample rate? Typically MIDI lags 2 to 3 ms depending on what's
making the sound (I've read reports of as much as 30 ms with
Gigastudio setups that aren't well set up). At 96 kHz, 300 samples is
about 3 ms. No big deal unless you look at the pictures.


--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me he double-m-eleven-double-zero at yahoo

wrote:
At 44.1/48K rates, 300 samples is between 6 and 7 milliseconds.


Which is about 6 to 7 feet of sound travel in air to put another
perspective on it.

Mark