Topic: Controlling Drift between 2 JB3's

[PrayForSurf]

I recently came upon a broken JB3 which just needed the input re-soldered. I now have two JB3's and was hoping to record using mics on one and take a sbd feed with the other, or just work on the source that came out better. I talked to a seasoned taper about this and he told me that "matrixing" them would be hard because of drift. He also said that there might be a way of controlling that drift but had to leave and I don't know how to contact him. So yea any help would be useful.

[stantheman1976]

Since they are the same exact device then drift shouldn't happen.  However, in the real world it doesn't turn out that way most of the time.  Even though you're recording the same even the two different devices may write information at slightly different speeds.  Digital devices work at constant speeds though so any difference will be constant and can be fixed easily.  It's not really that hard but can be intimidating if it's your first time.  I use Sony Vegas which allows adjusting by holding CTRL and grabbing the end of the track and moving it.  I believe Wavelab does this also.  You can use pretty much any audio editor but with others you have to figure out the difference by making some calculations and entering your value to stretch/shrink by.

What programs do you have or have access to? 

[heyitsmejess]

i just did this with 2 irivers (and 2 jb3s)

with the iriver sources (CA-14 omnis on 1, sbd on the other), i cut the SBD into roughly 30 min chunks, and matched those up to the audience in reaper.

the 2 jb3s matched up perfectly.

best advice i can give is this....zoom in as CLOSE as you can, as it makes it easier to sync.

[Roving Sign]

my jb3s never match - but I've heard others - (spark-e maybe) say they have 2 that are pretty close. So YMMV on that one.

Only way its really worked for me(without any stretching tricks) is using two external A/Ds with their clocks synced via a spdif.

[PrayForSurf]

I am currently using Audition. I guess I'll try my luck with some other editors as well, but cutting up into 30min chunks also seems like a good idea, shall try it out.

Thanks

[stantheman1976]

There's really no need to cut a recording.  It just means you have to sync more than one segment.  The difference between the two will be constant.  With a visual tool like Wavelab or Vegas you just go to the end of the segment, zoom in and drag it until points match up.

[Brian Skalinder]

The difference between the two will be constant.

YMMV -- This has not been my experience, but I know of others whose experience matches the above statement.  Only way to find out for sure is to try it, see what happens, and then take the appropriate approach once you know the situation.

[stantheman1976]

I've never seen a difference that wasn't constant when talking about digital devices.  I've recorded on PC, laptop, irivers, MD, and MiniDV.  I know analog is not constant and the only time I ever tried to sync an analog with a digital I gave up.

[heyitsmejess]

I am currently using Audition. I guess I'll try my luck with some other editors as well, but cutting up into 30min chunks also seems like a good idea, shall try it out.

Thanks

if you do the 30 min chunks, remember to leave one of the sources whole, and then synch up the 30 min chunks from the other source.

its a bit more work, but excellent if the source is only slightly off.

if they are WAY off, then you get into streching one to meet the other, and that is something that frustrates me to no end.

good luck!

[Yane]

I used to matrix using 2 JB3's, and their clocks were different.  The good news is the different rates were constant.  If I could determine the relative rates by locating noticeable features in the waveform at the beginning and end of the tracks, it was easy to compute the different rates by simple division; even better, this process also takes care of any delays in the board feed versus the speakers.  Putting your audio editor's time scale in samples rather than minutes makes the math easier. I applied correction  per song (about 3-7 minutes usually) taking a point in the middle of the song as reference. Rather than trying to apply time stretching -- didn't have any plug-ins that sounded decent -- I just cropped each song and moved it as a block, and any misalignment was not noticeable.