How does one do that? Is a master clock just a device you buy, plug, and play? Are there any that are both good and cheap?
The camera(s) and the audio recorder(s) need to be designed from the beginning to accept and/or generate time code (many mid-priced recorders, Sound Devices "T" recorders [because the "t" stands for Time Code - duh!] and almost all Zaxcom devices). Then they all run off of a common sync. I don't even know what they're using these days. SMPTE (Society of Motion Picture Technicians and Engineers)? Black burst? GenLock? But, even at the mega-budget level there really is no sync for film; they may use digital slates and a dozen other fancy devices, but film and production sound are still synced manually to this day.
Mmm, partly correct. When you start getting into the guts of it, synchronisation is quite a complex subject area. Part of the complexity is due to the history and different methods used to achieve sync. In the old analogue days pilot tones and neo-pilot tones were used to help sync cameras and sound recorders. Then came LTC (Longitudinal Time Code) to sync film or tape based cameras to DAT recorders (Digital Audio Tape recorders). Today with file based audio recording we can achieve extremely high sync accuracy but it's complicated by terms such as, bi and tri-level sync, word clock, video ref, black burst, genlock and others. It's not quite as complicated as it seems and I'll try and keep this a simple as possible:
There are two sides to synchronisation: Positional Reference (where are we) and Timing Reference (how fast we are going). Knowing how fast we are going (Timing Ref) is essential to maintaining sync but we need to have a common starting point, this is why we need a Positional Reference (time code). Think of synchronising running athletes, not only do they need to run at exactly the same speed (Timing Ref) but they also need to have exactly the same starting point (Positional Ref).
Positional Ref: File based audio recording systems record a single time code stamp in the audio file's metadata, which provides an initial Positional Reference. As there is no continuous time code recorded in digital audio files, Time-Code (Positional Reference) cannot be used to synchronise file based digital audio equipment, just provide a common starting point for cameras and audio recorders.
Timing Ref: Timing ref is a little more complicated, the timing reference signal used by all digital audio equipment is called "Word Clock". Whenever synchronising any type of digital audio equipment (mixers, DAWs, converters, recorders, etc.) Word Clock must be used, either with one piece of equipment (usually the A/D converter) acting as the word clock master and all other equipment set to be word clock slaves or by having a dedicated master word clock unit with all the digital audio equipment slaved to it. What makes timing ref a little complicated is that word clock is not used for sync'ing picture, bi-level sync (or in the case of high def tri-level sync) is used. The signal which carries this picture sync info is often called Black and Burst or more accurately and commonly today, Video Ref. How we generate, distribute and combine these two different timing protocols (video ref and word clock) depends on the equipment being used. If the audio equipment only has word clock input we will need a master clock unit capable of generating a video ref signal and a word clock signal from a single internal clock with outputs connecting (via bnc cables) to the video ref input on the camera/s and simultaneously to the word clock input of the audio recorder.
Higher end audio recorders designed specifically for A/V applications have a video ref input (which is then converted internally into word clock) and therefore only a video ref (sync) generator is required. Sync generators are generally much cheaper than master clocks (sync gens) with multiple output protocols but then audio equipment with video ref inputs is usually more expensive, so you pay for it one way or another! BlackMagic Design do a good sync generator for under $300.
There are in fact quite a few different ways of generating and distributing timing ref though. About the best method I'm aware of and becoming increasingly common is GPS based sync. This is a special GPS unit attached to each camera and audio recorder. The timing reference is extracted from a GPS signal which is generated by an atomic clock. This system is extremely accurate, very flexible (because it's wireless) and I should imagine pretty expensive.
Once all the picture and audio equipment is locked to the timing ref signal the system is said to be Genlocked, Fully Resolved or Frame Edge Aligned.
What I hope is obvious from all this is that it's obviously impossible to achieve this level of sync unless all the cameras and audio equipment actually have the facility to generate or receive a timing ref signal (video ref and/or word clock) and unfortunately, most pro-sumer level equipment does not.
BTW, even with generated sync, there will eventually be sync issues on extremely long takes.
This statement is not correct. Setting up a frame edge aligned system should allow for close to sample accurate sync over extended time periods. IE, sync accuracy of better than a quarter of a frame in a 24 hour period. It should be noted though that frame edge alignment must also be carried over in post production. IE, A sync generator locking both the video card and DAW to a single master timing ref signal. In the case of ProTools, this means a ProTools HD system with a Sync HD, a sync generator and a video card such as the BM Studio II or HD Extreme which have a video ref input.
G