Daisy-Chaining 3 Sony PMW-350K Cams

[John Blankenship]

For those who have done a lot of multiple camera shoots, what is the general consensus of the best way to link 3 Sony PMW-350K cameras if I don't have tri-level sync?

 

1. Feed the first camera with my timecode (SB2a -- just timecode, no tri-level sync) and from there each camera feeds the succeeding one with both timecode and sync signal.

 

2. Let the first camera be the sync master, feed the others as above and feed the timecode out from the last camera to my Deva.

 

Is there any danger in #1 above, as all three cameras will be locked with each other, or is there is there a chance that feeding the first camera just timecode could cause an issue?  It seems to me that either should work fine. 

 

The downside of doing #2 is that I'd need to go wireless to the slate(s), fed from the Deva, as I don't trust Sony timecode to be stable over time.  Not a big issue, just one more thing to be concerned about.

 

Thanks.

[Marc Wielage]

I don't think it can be done very well without tri-level sync and timecode boxes. Without that, the cameras are going to drift, in my opinion. 

 

It may be possible that the drift is manageable by the editor, and if you let them know in advance and all the cameras have a scratch track, they can just compensate in the edit. They have to be prepared to jig and jog a little bit as required, especially for long takes (over 5-6 minutes). I bet it'll hold OK for a normal 3-minute scene... but a disaster for a 90-minute concert or something like that.

 

Renting 3 SB-T's or Lockits shouldn't cost more than $150 or so. The camera department can't afford this?  

[John Blankenship]

 Thanks, Marc,

 

But, here's where I'm puzzled...  Shouldn't the cameras stay in sync with each other since they are getting both timecode and a sync signal from each one to the next? 

 

It's the way Sony recommends in the manual to do it.  I'm just adding the part about audio timecode (before? or after?).

 

Is Sony all wet on this (it wouldn't be the first time)?

[Marc Wielage]

I think the sync will get glitchy (that is a technical term I just made up) if one camera feeds another, particularly on a battery change or a reboot. Genlock references have to always be there and always be accurate, or else all bets are off. 

 

I see that diagram on page 64 of the manual about tying all the cameras together, but that sounds like a lot of hooey to me. In the real world with studio cameras, they all get external sync from a master sync generator in another part of the broadcast plant. In most location situations, I think the camera operators would be driven mad if every camera had two umbilical cords attached to the other cameras, limiting their movement and position. If this were a lockdown and the cameras never had to move, like during a live concert situation, then maybe it could work. I'd do a test beforehand. 

 

I still think renting 3 SB-T's is a better solution, plus it'll allow the cameras to run around on stage, in the audience, and all that stuff without potentially knocking people down with wires. 

[John Blankenship]

Fortunately, on this shoot, camera mobility is not an issue. The tether scenario Sony recommends is what the camera department is planning. That's kind of a given at this point in time.

Any thoughts on which way you'd approach timecode under this scenario? I'm kinda leaning toward being at the end of the chain. That way, I'm not imposing any variables, camera-sync-wise, into what they had already planned.

Thanks, Marc, for the feedback.

[Sound Art Film]

My opinion on this would be to do option 2, as it would be the least fault prone of the two options. 

 

With option 1, there would only be TC feeding camera 1 without any sync.  Over a given time (usually longer periods), there could be an issue between the sync box feeding TC and the camera's clock causing some drift. 

 

With option 2, you can follow camera's TC and Sync.  Your audio files would match exactly and everything should be preserved.  Marc does offer some good advice though about the potential pitfalls of option 2, but given an either or scenario, I would go with 2.  Now, if the sync box had TC and Sync, you could probably do option 1 just fine. 

 

My humble 2 cents here...

 

Paul

[John Blankenship]

Thanks, Paul.  I appreciate the input.

[Marc Wielage]

That would be the other approach: get one SB-T box and then use a DA to feed the TC out and Sync out to all three cameras. That way, if one camera is powered off or has a technical problem, it won't affect the other two. The SB-T could also feed TC and genlock to the sound recorder, and everything would be referenced identically.

 

I still think it's dicey to daisy chain any video devices like this, because there's too many opportunities for failure. Any cheap video DA could be used to distribute the reference, and we're talking maybe $250 max for that. Timecode is just an audio signal, so splitting that four ways is pretty easy. But I definitely think in a multicamera situation, everything should have both identical timecode and identical reference in order for it to work reliably. 

 

Is it possible to do a camera/sound test prior to the shoot? That would be the other thing I'd advise, particularly with (say) a 10-minute timecode slate test with a clap at the head and a clap at the tail. Pull all three camera files into a timeline, along with the WAV files, and see if they line up and work correctly.

[john Quinn]

Daisy chaining is just a big pain in the neck for everyone, it is cheaper time wise to use lockits and more reliable because you have less chance of getting a broken BNC .As one broken BNC at the beginning of the chain will affect 

all timecodes further down the chain ie all cameras will revert to their internal generators and quickly drift apart .

Best John

[John Blankenship]

Thanks, John, Paul, and Marc for the comments.

[edward chick]

I have linked three cameras together via just timecode and never had a problem. I've done this a lot for Dateline type shoots.

I think in a perfect world sync boxes would be the way to go, but when you are working with owner-operators who occasionally have to tie multiple cameras together, and given network budgets it's not always an option. I have not had to add an outboard recorder to the mix however. Usually a 442 feeding everything. 

[Philip Perkins]

Having the A camera be the TC and genlock master is a very normal way of syncing pro video cameras, and has been since we've had camcorders, certainly for longer than we've had syncboxes.  Sony cameras will usually automatically "take" external sync and TC if it appears at the correct input jacks, so we just do what you said--wire the cameras A>B>C.  Anymore most DITs I work with don't like this arrangement very much, and would prefer to have a master sync generator and TC source, but the old way will work.  The problem is how your recorder will interface, since it doesn't take trilevel sync.  That's a reason to go the Lockit route.  It seems like in a hardwire daisy chain you could get away with just having the Lockit on the A camera and hardwiring the rest to it for sync and TC?

 

philp

[John Blankenship]

Thanks Edward and Philip.  I appreciate the feedback.  I should know more tomorrow and hopefully we can run a couple of tests.

 

I've been involved numerous times when the cameras were linked thusly, but haven't synced an audio recorder into such a scenario.  When shooting with a single Red or an Alexa I just feed them my timecode and all has been fine.  Fortunately, for this one everything will be edited, not switched, so that should make it less problematic.  They're likely to use the camera audio anyway, which will be the mix.  I decided that since it's a less than ideal acoustical environment, I'd give them a mix plus isos so they have post options, if needed.

 

Thanks, again, everyone.

[john Quinn]

Daisy chaining does work no doubt I used to do it myself many years ago,But all those extra long bnc cables lying around the floor is an  accident waiting to happen .Loosing a picture on a monitor is an easy thing to spot but once you start filming TC drift is not so easy.And as has been said many times before the sound department get the blame when there are TC problems not the camera department.

Best John

[cmassey]

Hey John...did just as you are thinking a couple of weeks ago.  However it was my 552 at the end of the chain, as the recording was for transcription purposes only.  I never heard from an editor or producer who complained about the TC.  And believe me, if transcriptions are screwed up I hear about it.  Of course it could have been off a frame or two and for those purposes never be noticed. 

 

Good luck...let us know how your tests go today...

 

cleve