Accuracy of CD players.

farss wrote on 11/26/2005, 2:56 PM
Here's a trap!
Just started work on event we shot last night, a dance performance. Got a feed from the desk and used the R4 to record things like tap mics that weren't being fed to house mix. For some odd reason we could only get a mono feed from the desk but no sweat, they gave us the CDs of the dance music.
So I started syncing that up to our recordings but shock horror it's out enough to cause problems. Well, duh, I finally realise, the venues CD players have speed control and I'd bet even on 'lock' they're not that accurate.

But here's a question for a ways down the track. Say we're running all our cameras with genlock. I know one can get sync generators that also generate audio clocks but can one get things like CD players that'll lock onto those clocks as well?

It does explain why someone I bumped into was doing wht they were doing. They'd dubbed all their audio to a DB deck for playback at the venue. At the time I thought they were nuts tying up all that expensive kit to do what a CD player could do, now I understand the logic of it.

One last thing, a simple Vegas feature request. Markers that can be locked to tracks, not the T/L. That'd make sample accurate syncing so much easier.

Bob.

Comments

PeterWright wrote on 11/26/2005, 4:30 PM
Yes, there are variations - I recorded some school kids singing to a backing CD last week, and the player they used was slightly slower than the audio I extracted from the same CD.

Fortunately Vegas makes it very easy to stretch or squash to fix this.
Spot|DSE wrote on 11/26/2005, 4:35 PM
In a DJ situation, it may well be the DJ sped up or slowed down the CD to match beats from other cuts. This is pretty standard fare in a dance environment.
farss wrote on 11/26/2005, 5:54 PM
Thankfully that didn't happen, hadn't even thought of that one, yipes!
Fortunately just ballet and tap. Only real issue I'm having is extracting the 'taps', the stage has some sort of urethane finish so the 'taps' didn't much and the mics picked up a lot of the room i.e. the music.
I'd had some wild notion about cancelling the music as I have the music, even a copy as it was fed into the room however as the tap track is a mixdown from three shotguns I think this'll be way, way too hard given the effect of room acoustics, if I had separate tracks from each mic than I might have a shot.
So what I'll do is use some massive amount of compression and Eq to lift the taps out of the track.
One thing I learnt from this gig, it pays to have everything you think you'll need and everything you didn't even think of. I'd forgotten that most desks don't have XLRs, they're all 1/4" RTS and well, 'we had some adaptors here yesterday' ended up being solved with a quick soldering iron job.
Bob.
ottowr wrote on 11/28/2005, 5:21 PM
Sony, HHB, Tascam and others make CD players with word clock inputs.
farss wrote on 11/28/2005, 8:52 PM
Many thanks,
I'll look further into them.
Bob.
ottowr wrote on 11/28/2005, 10:01 PM
One useful side-effect of locking the cd player to word clock is that the player will Sample-Rate-Convert the output to match the word clock rate, e.g. 44k1 disc played with 48k clock will output as AES-EBU 24bit @ 48kHz.

The HHB unit looks good.

http://www.tag.com.au/jsp/displayNewsItem.jsp?id=109&newsType=News
farss wrote on 11/29/2005, 2:20 AM
I'm not certain if even this is going to achieve what i want to do.
When I capture a CD into Vegas or whatever the digital values are read and each sample then placed such that there's 44,100 of them per second of the T/L. This doesn't depend on how fast the CD player spins the disk.
But when I play the same CD in even one of these players OK the data is clocked such that the samples are locked to the word clock but is the speed that the CD rotates at also locked, this player provides +/- 10% speed control, I gather that even if the CD is spinning 10% faster the output clocks are locked to word clock but clearly the speed of the CD isn't.
What I want to ensure is that the recorded output of the CD player precisely matches what I get when I digitally capture the same CD, even on the rather expensive players in the venue there's an error of around 0.18%, not a big amount but over the length of a number enough to get things out by a beat or two. Easy enough to fix in Vegas but if I'm going to be doing a lot of these gigs which involve around 40 tracks it's a chore I'd like to avoid, or at least know there is a way to avoid, first we've got to get the XL-H1s though.
Maybe I'm worrying about nothing, maybe this player does do exactly what I want, guess I need to talk to TAG to find out for certain.
Bob.