is there such a thing?
I think a remember a few years back someone posting this symptom but I do not recall what caused it or how it was remedied.
THe problem is this.
I am using two different drum machines while recording, one built in midi on a GNX4 (digitech) the other from a Palmtop Studio (PS-04).
Anywho I am noticing when aligning the tracks to a click track in the same BPM that they are not in sync, I can sync them up at the beginning but they are noticeably early about 40 sec in.
That is why I am calling this reverse lag, because instead of the tracks being slower than the click track they gradually get faster! 
Usually this wouldn’t matter cause I am only doing 20-30 sec parts at a time…but for anything over 40 sec it is noticeable to me.
At first I though it was the drum machines maybe not having a true BPM that is the same as the computers generated one.
(BTW when I record midi drum data from the GNX4 in PRotracks it stays with the BPM perfect) but this is a live recording issue that I am assuming has something to do with my sound card or setup in Ntrack.
I am recording in 44k using a M-Audio FIrewire 410. I tried changing the buffering speed to a higher number but that did not make a difference. Right now it is a 1024 and the device sync says it is “locked”, inside Ntrack it in preferences>>Buffering it says “Custom” where you would set up buffer speed.
If I record the drum tracks directly to the devices then transfer them into the computer and line them up they are spot on after 3 min, so I am pretty sure it is a PC issue.
Anybody come across this before?
keep shinin
jerm 
So know one has ever encountered this?
Again, the symptoms are, that everything recorded is faster than it should be, in other words if I play a drum beat in 120bpm eventually (after about 2 min) it is not precisely in line with a click track that is 120bpm.
I never noticed it with guitar or vocal tracks because I just play along with the click or midi beat and probably compensate as it goes, these are beats that are generated by two different drum machines.
keep shinin
jerm
Beats generated by two different drums machines? How were they created? If they aren’t synced then they will drift apart. The only way they won’t drift is to slave one to the other, and that includes slaving n-Tracks to the drum machine, or slaving the drum machine to n-Tracks. Use one master and many slaves.
Reverse lag just means one is a little faster than the other. Relatively speaking, that could mean one is slower, but it could mean one is faster. Which one is right? That’s the point of reference.
does this happen if one macine is used to create 2 tracks - that is create one track then the other. I think that would remove speculation that the issue is with two different drum machines.
Bax
Quote: (phoo @ Oct. 16 2009, 7:27 PM)
Beats generated by two different drums machines? How were they created? If they aren't synced then they will drift apart. The only way they won't drift is to slave one to the other, and that includes slaving n-Tracks to the drum machine, or slaving the drum machine to n-Tracks. Use one master and many slaves.
Reverse lag just means one is a little faster than the other. Relatively speaking, that could mean one is slower, but it could mean one is faster. Which one is right? That's the point of reference.
Beats generated by two different drums machines? How were they created?
Good questions.
THe one would be midi, played out of a GNX4, at 120 bpm, but recorded using the XLR out then into a firewire card in the PC.
The other is generated by a Palm Studio, which has it's own internal timer BPm set at 120 but it is pure audio that comes out of it.
THis is only happening when I send out audio from the devices and record it on the PC, if I send out midi and record it in the PC it stays at 120bpm indefinitely provided the midi file is already also set up at 120bpm before recording.
All of my timing issues in the past have been lag related, where the computer was recording things slower than they were being played....this is reverse which was why it was odd and it is only noticeable after a minute or two in, really it's not even audibly noticeable, but you can SEE it on the time line when the beats are lined up to a click track.
To clarify I am not comparing the two beats to each other, but rather the beat tracks to a click track generated by the computer and in wav. format.
They may be faster or slower than each other for all I know, but they are both faster than the click track after a while which is why I thought maybe it was a PC issue.
The next thing I want to do is export the drum tracks themselves out of the cards they are on in the devices and put them into Ntrack (this can be done by recording them to the devices themselves then reading the cards via USB) to see if they are still faster than the on board software click that way.
I was under the impression that BPMs were some kind of standard, that was why I suspected a PC issue.
keep shinin
jerm
Quote: (bax3 @ Oct. 16 2009, 7:53 PM)
does this happen if one macine is used to create 2 tracks - that is create one track then the other.
I think that would remove speculation that the issue is with two different drum machines.
Bax
YEs it does.
keep shinin
jerm
Wonder if the ppq (pulse per quarter note) setting for each machine has to be the same? Don’t really know - just a thought.
Quote:
I was under the impression that BPMs were some kind of standard, that was why I suspected a PC issue.
The standard is that BPM, meant Beats Per Minute. That's a close as it gets to being a standard. The real issue is that timing is based on the interal clock, generated in hardware by a crystal on the sound card. If all devices are not being run from the same crystal then there will be some drift.
The crystals of two sound cards may be at some arbitrary speeds. Whatever the frequency, they might be close, but not perfect.
This is going to sound like gobbledygook, but her goes anyway. This needs a movie and samples. Words just don't work well to explain.
Take one drum machine, record MIDI out one into a MIDI track and the audio out into a wave track at the same time on one machine. Plat them back on the smae machine they were recorded on any there's a good chance they will stay in sync, up to a point.
Move the MIDI and wave tracks to another machine into one song. Play the MIDI track and the wave at the same time, same BPM. They may seem in time. They may drift. It depends on how close the crystal in the other machine is to the first machine.
To force the issue, intentionally change the BMP. The MIDI track will change tempo. The wave track will not. They will get out of sync, very fast.
On the other hand, the rendering of the wave file will be slightly different on two machines, causing the pitch and playback speedf to be slightly different. Usually the difference is so small that your ears don't hear the tuning differnece. The way it shows up is slight drift our of sync over a few minutes. This is a BIG bugaboo when trading wave files. It's yet another reason multiple machines must be slaved from one clock. When trading files, this affect may be negated.
Record a wave at 44100 sample rate - send it to a friend.
Friend plays is back at 44100 - records a new track at 44100 while playing along -- sends the new file back to you.
You open the new file into you song at 44100 and it plays in perfect sync.
Bring you machine to your friends house - play his and your songs at the same time. They won't be in time unless you sync one machine to the other.
44100 is just what we call it. It can be slightly different on different machine, because it's that crystal hardware clock. It could be bassed on 192000.00005hz or 47999.0005hz. Or something like that. Whatever math is used to convert to 44100, or close. Every machihne is different to be a major pain in the butt. Software has to go to great lengths to account for this and try to get it close.
Record a wave at 44101 sample rate - send it to a friend.
Friend plays is back at 44099 - records a new track at 44099 while playing along -- sends the new file back to you.
You open the new file into you song at 44101 and it plays in perfect sync.
That's because recording and playback happen at the same time from one clock on each machine.
When playing indpendently, the 44101 machine will play both wave tracks at a slightly higher pitch and faster tempo - in sync and in pitch. The 44099 machine will play both tracks slower and lower - in sync an in pitch. The pitch of a MIDI track will stay the same on both machines (assuming external MIDI module - VSTi and soptware synths adds yet another variable), but may be slightly off tempo from one machine to the other but just a litle.
UGH...I know how bad and confusing this reads.
http://www.agm.me.uk/blog/2007/09/soundcard-clock-accuracy.php
There’s a link to a utility to see how accurate the sound card clock is. I haven’t tried it, but it might be useful.
Quote: (phoo @ Oct. 19 2009, 12:03 PM)
Quote:
I was under the impression that BPMs were some kind of standard, that was why I suspected a PC issue.
The standard is that BPM, meant Beats Per Minute. That's a close as it gets to being a standard. The real issue is that timing is based on the interal clock, generated in hardware by a crystal on the sound card. If all devices are not being run from the same crystal then there will be some drift.
The crystals of two sound cards may be at some arbitrary speeds. Whatever the frequency, they might be close, but not perfect.
This is going to sound like gobbledygook, but her goes anyway. This needs a movie and samples. Words just don't work well to explain.
Take one drum machine, record MIDI out one into a MIDI track and the audio out into a wave track at the same time on one machine. Plat them back on the smae machine they were recorded on any there's a good chance they will stay in sync, up to a point.
Move the MIDI and wave tracks to another machine into one song. Play the MIDI track and the wave at the same time, same BPM. They may seem in time. They may drift. It depends on how close the crystal in the other machine is to the first machine.
To force the issue, intentionally change the BMP. The MIDI track will change tempo. The wave track will not. They will get out of sync, very fast.
On the other hand, the rendering of the wave file will be slightly different on two machines, causing the pitch and playback speedf to be slightly different. Usually the difference is so small that your ears don't hear the tuning differnece. The way it shows up is slight drift our of sync over a few minutes. This is a BIG bugaboo when trading wave files. It's yet another reason multiple machines must be slaved from one clock. When trading files, this affect may be negated.
Record a wave at 44100 sample rate - send it to a friend.
Friend plays is back at 44100 - records a new track at 44100 while playing along -- sends the new file back to you.
You open the new file into you song at 44100 and it plays in perfect sync.
Bring you machine to your friends house - play his and your songs at the same time. They won't be in time unless you sync one machine to the other.
44100 is just what we call it. It can be slightly different on different machine, because it's that crystal hardware clock. It could be bassed on 192000.00005hz or 47999.0005hz. Or something like that. Whatever math is used to convert to 44100, or close. Every machihne is different to be a major pain in the butt. Software has to go to great lengths to account for this and try to get it close.
Record a wave at 44101 sample rate - send it to a friend.
Friend plays is back at 44099 - records a new track at 44099 while playing along -- sends the new file back to you.
You open the new file into you song at 44101 and it plays in perfect sync.
That's because recording and playback happen at the same time from one clock on each machine.
When playing indpendently, the 44101 machine will play both wave tracks at a slightly higher pitch and faster tempo - in sync and in pitch. The 44099 machine will play both tracks slower and lower - in sync an in pitch. The pitch of a MIDI track will stay the same on both machines (assuming external MIDI module - VSTi and soptware synths adds yet another variable), but may be slightly off tempo from one machine to the other but just a litle.
UGH...I know how bad and confusing this reads.
Makes perfect sense to me!
I have done this same maneuver with countless tracks and other peoples computers over the past decade, very rarely was there ever a problem, in 48 or 44.1.
And the problems were usually other peoples tracks (since we have multiple players and machines involved and the only tracks off were specific ones) but lag, not this.
I do understand the calculated bit rate the computers are trying to pull of and have witnessed slight pitch differences in the past.
Like I would burn a file and put it on the net, then download it to my wifes new Compaq on Vista and try to play along with it on a guitar that was already in tune with it previously and it was a bit off pitch wise.
I am trying to find out what "chip" is involved in my GNX4, and my Palm Studio, and how that differs from a computer.
I tried playing a track at 119.9 BPM out of the palm studio to see if it would equate to the 120bpm generated in Audacity and rendered to wav, but it was too slow so the difference is less than 1/10th a beat.
Will read the articles when I have a chance.
Thank you for the replies, and am leaning toward calling this a hardware issue, one I need to solves since I am doing a great deal of syncing for live shows and need to be able to maintain a proper BPM that for at least 5 min that is going to line up to the backing tracks.
Plus I am using the GNX4 to capture dry tracks when I do recording,(while the PC captures the amp sims/FX) and if they are not going to sync up later because I am not starting with an identical BMP I could be wasting time there as well.
keep shinin
jerm