Double system sound and sync drift?

[David Morgan]

Looking for confirmation or correction of the following:

Time code: To use TC, your camcorder and sound recorder must accept an external
TC signal from a common source.

Sync: In order to stay in sync with each other, the units must be driven by a common clock or genlocking source. The units must have physical TC in and Genlock connectors on them.

Conclusion: TC is good for lining up video with audio but only if it's laid down in sync and forced to stay in sync.

Pretty much excludes most pro-sumer camcorders like my XH-A1 (not the G1) and most Sony camcorders.

The same would be true if trying to lock two cameras.

any comments?
thx

[Roshdi Alkadri]

Ok, your camera and recorder dont necessarily need TC to sync up. Its sweet if you have TC IN/OUT on the camera and/or recorder, and you can send TC out from the camera into the recorder or the other way around.

I have a TC capable recorder but the camera doesnt have TC IN/OUT. I can set the recorder to 23.976 frame rate so it matches with the camera frame rate.

The best way is to use the good old clapperboard and to record that sound onto your recorder and to film it on your camera. then you have to manually line up the sound of the clack along with the clapperboard closing. this technique has been used since the inception of double system and it works. Yes, you have to line up every take and thats why its good to assign scene/take numbers to everything you shoot. it makes it easier in editing and syncing.
It works nicely for me.

[Steve House]

You are correct for the most part. There are two elements to fully syncing audio and video, a positional reference and a speed reference. Timecode only provides the position reference, really it's just a fancy slate. To prevent drift, you must also have a speed reference, that is, the camera and the audio recorder must share a common timebase. There are several ways that can be accomplished. If one has a camera that accepts genlock, one way is to use a master clock to simultaneously send genlock to the camera and wordclock to the audio recorder. Sound Devices 7xxT recorders offer a variation of that where their internal clock is a module made by Ambient that is similar to a Lockit box built right in to the recorder, and another Lockit can be tuned to it to provide genlock to a camera that accepts it. Another is to use an audio recorder such as the Tascam HD-P2 or Sound Devices 788T that accepts video coming from the camera and slaves its sample clock to the embedded blackburst (SD) or Tri-level (HD) sync signal. Sure wish someone made a battery-powered and affordable converter box that would accept a composite video signal from a camera and generate wordclock from it but AFAIK none are on the market at the moment.

[Carlos E. Martinez]

Some time ago I got a mail from Sound Devices confirming the following, which I really didn't know:

"The 702T, 744T, and 788T all handle time code in the same manner. When
recording with time code the recorders internal timecode generator is always
running and is available as an analog signal via the 5 pin LEMO connection.
When the file is written, the starting time code value is stamped into the
header information of the file. When the file is brought into a program that
can read the time code stamp, it usually can insert it at the appropriate
point on the time line. When the file is played back on the recorder the
recorders internal timecode generator is again triggered and SMPTE time code
is again present at the LEMO out. If one needed time code striped to an
audio track for say, transcription purposes, you could run a LEMO 5 to XLR
cable back around to an available input and route it to a track on the
recorder while recording audio on the others."

Apparently all TC recorders work in the same way, though not all allow going in with a video signal, as Steve points out, to lock to and really be in sync.

[Carlos E. Martinez]

Elaborating further about this same question, if the TC just sets a positional reference and it's stamped once at the beginning of the take, there was something Ambient did offer some time ago.

I think it was called TC burst, and it allowed you to use crystal-sync controlled non-TC machines, like Nagras, DAT and MD recorders. During transfer to a computer or perforated tape you would feed that TC burst into a TC generator that would re-construct the complete TC signal and stripe the audio file with it.

Taking that to present times, a similar routine could be implemented on non-TC CF recorders, even if the speed reference set by the genlocking video would not be accomplished. Then only shots up to perhaps 30 minutes should be in sync. On thing I don't like about the genlocking system is that you would need a cable to link camera and audio recorder. I'm not sure you should genlock through a wireless link.

[Steve House]

Originally Posted by Carlos E. Martinez

Elaborating further about this same question, if the TC just sets a positional reference and it's stamped once at the beginning of the take, there was something Ambient did offer some time ago.

I think it was called TC burst, and it allowed you to use crystal-sync controlled non-TC machines, like Nagras, DAT and MD recorders. During transfer to a computer or perforated tape you would feed that TC burst into a TC generator that would re-construct the complete TC signal and stripe the audio file with it.

Taking that to present times, a similar routine could be implemented on non-TC CF recorders, even if the speed reference set by the genlocking video would not be accomplished. Then only shots up to perhaps 30 minutes should be in sync. On thing I don't like about the genlocking system is that you would need a cable to link camera and audio recorder. I'm not sure you should genlock through a wireless link.

The genlock signal is at a much higher frequency than an audio signal or a SMPTE timecode burst and regular wireless transmitters and receivers would be not be appropriate for it. Eliminating the cable is where Lockits come into play. The Lockits have highly stable oscillators that can be precisely tuned to each other. One is connected to the camera's genlock in and provides the video reference clock. Another is connected to the audio recorder's wordclock in and provides the audio reference clock. (And they also provide TOD timecode to both camera and recorder.) As I mentioned before, the Sound Devices 7xxT series of recorders have the equivalent of a Lockit already built-in so if you're using one of them you don't need the external box for the audio recorder - all you need is a Lockit for each camera.

I think one of the reasons this whole timecode sync thing is confusing to a lot of people is back when TC was first introduced, the post workflow involved transferring the sound from analog tape or DAT to magnetic perf or, after Avid replaced the Movieolas and Steenbecks, essentially capturing it in real time from the DAT tape into the file format the audio workstation needed, much like the process today of capturing a DV tape from your camera's playback into your computer for editing. The analog and DAT tapes had LTC recorded alongside the audio when the original recording was made. That TC signal playing back off of the tape was continuously compared to a standard reference clock during the dubbing or transfer process and controlled the playback speed, speeding it up or slowing it down so it matched the reference. But modern file-based workflows where the production audio is recorded directly into a wav, bwf, or mp3 file on a hard drive or CF card don't work like that.

[David Morgan]

thx to all for the replys. I guess I have to keep syncing the hard way as it seems u have to pony up a boatload of money to do multi-camera and double sound. Most, if not all, of these prosumer cameras aren't built for it.

[Carlos E. Martinez]

Originally Posted by David Morgan

thx to all for the replys. I guess I have to keep syncing the hard way as it seems u have to pony up a boatload of money to do multi-camera and double sound. Most, if not all, of these prosumer cameras aren't built for it.

Well, I wouldn't say so. The problem with most prosumer cameras is that their internal mic preamps suck, and even more most of the mics they come with.

If to that you add that using an on-camera position for the mic is most of the times far from good and you need a boom man and/or lapel mics, wireless or hardwired, then doing double system sound may be the way to go.

This may not work for long takes, longer than 30 minutes or so, but it does work superb in all other situations.

You may sync using a clapper, a bloop light or simply sync the audio wave-forms in the editing program. But after that everything goes fine.

In any case, if you are doing multiple camera situations you should feed the same sound to all cameras, so you would still have to sync things.

[David Morgan]

So question is, would it work to take the video out of my Canon XH-A1 and connect it to the video clock input on the Tascam P2 recorder? I realize that this is not TC but would it lock the speed of the tascam to the Canon to prevent drift?
thx

[Carlos E. Martinez]

Originally Posted by David Morgan

So question is, would it work to take the video out of my Canon XH-A1 and connect it to the video clock input on the Tascam P2 recorder? I realize that this is not TC but would it lock the speed of the tascam to the Canon to prevent drift?
thx

You need a genlock signal, which I don't know if the Canon provides.

You can also feed an external TC signal into the Tascam P2 with a Denecke generator, like the SB-3.

[Steve House]

Originally Posted by David Morgan

So question is, would it work to take the video out of my Canon XH-A1 and connect it to the video clock input on the Tascam P2 recorder? I realize that this is not TC but would it lock the speed of the tascam to the Canon to prevent drift?
thx

Yes, the Tascam is expressly built to sync its internal sample clock to the blackburst imbedded in the video signal if you give it one. Ot if you want fancier, the new Sound Devices 788T also accepts video and will sync to it. Alas the older 702T and 744T do not.

[David Morgan]

um... I thought gen lock was the vertical sync pulse. I've heard u can simply use a black generator or the P2 would sync to the vertical sync pulse of the cameras video out.

[David Morgan]

thx Steve. The tascam is now tempting. I've used it a couple of times. However, I don't like the business with the limiter switches. It's some murky setup where ur using the mic inputs (XLR) and the limit switches are supposed to work with them. The manual is pretty useless. However, I'm sending line level from a mixer into the them. Already a shakey deal. I swear I can hear the clipping even though the "limiters" are activated.
Next thing is, I sure wish there was a 4 channel version.

:-)

[Steve House]

Originally Posted by David Morgan

thx Steve. The tascam is now tempting. I've used it a couple of times. However, I don't like the business with the limiter switches. It's some murky setup where ur using the mic inputs (XLR) and the limit switches are supposed to work with them. The manual is pretty useless. However, I'm sending line level from a mixer into the them. Already a shakey deal. I swear I can hear the clipping even though the "limiters" are activated.
Next thing is, I sure wish there was a 4 channel version.

:-)

There's always the Sound Devices 788T 8-channel model, a mere 6 grand <g>. I thought Horita made a battery powered synchronizer that accepted video and sent wordclock but I was checking for it last night and have'nt been able to locate it.

[Carlos E. Martinez]

Today I was discussing this drift question with a friend of mine that works with video, and he tells me that when shooting with several cameras they go to "free run" and then jam sync them all to the same clock. They then go their separate ways and shoot without any links.

So he asked me why there should be any drift in these audio machines, which are basically digital based and shouldn't "slip-away" as DAT or MD discs might do.

Isn't any clock reference recorded along with the audio and why isn't it as accurate as that on the video cameras?