stevel@tybalt.caltech.edu (Steve Ludtke) (09/21/89)
Hi. Does anyone out there know how time base correctors (like the ones used to sync 2 video signals) work ? If so, could you post a general description of how they work ? Just curious ... --------------------------------------------------------------------------- stevel@tybalt.caltech.edu CS-> | <-Ph stevel@citiago (Bitnet) \|/ I DO ... 72335,1537 (Compuserve) ?
bkc@image.soe.clarkson.edu (Brad Clements) (09/21/89)
From article <11991@cit-vax.Caltech.Edu>, by stevel@tybalt.caltech.edu (Steve Ludtke): > Hi. Does anyone out there know how time base correctors (like the ones used > to sync 2 video signals) work ? If so, could you post a general description > of how they work ? Just curious ... > > --------------------------------------------------------------------------- > stevel@tybalt.caltech.edu CS-> | <-Ph > stevel@citiago (Bitnet) \|/ I DO ... > 72335,1537 (Compuserve) ? There are two basic types of TBC's on the market, analog and digital. They vary in the number of lines that can be stored (corrected) (known as 'window') some units are full field and frame window units. Analog ones usually max out at 8 or 16 lines of correction. Basically VCRs that can be connected to TBC's (not all can be connected to analog models, but full frame units known as frame synchronizers will handle any video input) have a video input that accepts advanced compensation sync signals (ADV COMP). The idea is that a tape player will speed up and slow down minutely while playing a tape, and these 'broadcast' VCR's can be told to speed up and slow down by sending an advancing sync signal into the VCR. Analog and small window units (16 lines is considered small these days) advance the signal coming from the VCR until the vertical retrace signal is aligned with 'house sync' (the signal that you're trying to lock on to). Then, regular wow and flutter in the VCR causes the signal to advance or delay by a certain amount, at which point the analog and digital line delay units jump in to compensate for the speed change. Thats the basic synchronization part. (of course the ADV COMP signal is going all the time and the line delay is always working, its quite fast). Some TBC's also offer drop out compensation (usually handled by the VCR) in which a lost video line is substituted with the previous line by the TBC. Most VCR's do this (up to 4 or 5 times in a row) before actually showing a snowy line. Drop outs happen all the time and the DOC in the VCR's work quite well. Most TBC's also get into the actual time-base- correction process, in which horz and vert sync, and color burst are stripped off the incoming signal and digitally regenerated, and added on the final output to give a 'clean' output signal. Some TBC's have a proc amp (process amplifier) section which allows video level, color level, hue and pedestal adjustments. You can also buy stand alone proc amps that don't have the TBC functions. Full frame TBC's, also known as frame synchronizers, have enough digital memory to store an entire frame of video, therefore the advanced comp sync signal is not required, but may be used anyway. Digital units (the common ones) have 8 bits for luminance, 4 bits for saturation and 4 bits for hue (8-4-4) (don't quote me on this one). Some TBC's are dual units, containing two frame buffers that handle two input signals, extra cpu power can be added to mix and match the input buffers to produce effects on the output side, such as compression, push off/on, tumble (needs lots of cpu). I speak from experience w/ the following TBC's FOR-A FA-200 Microtime S320D, T120D E120 (yuck). Quantel Adda hope that answers your question. -- | Brad Clements bkc@omnigate.clarkson.edu bkc@clutx.bitnet | Network Engineer Clarkson University (315)268-2292
brian@ucsd.Edu (Brian Kantor) (09/21/89)
Simple time-base correctors work by using a double-buffered two-line digital memory. The take the incoming video signal and digitize it, then store one horizontal line at a time into one of the line buffers. That buffer is then swapped with the other one, and the data just received is then sent back out using the station horizontal sync and chroma reference while the other buffer is being filled. Clearly this introduces a one-line delay in the video signal path. Often all the other inputs to the switcher console will be run through a 1H delay line to correct this. This simple scheme does not correct for vertical sync errors. When using a pair of TBCs to edit tapes into an effects console, for example, the playback decks have their vertical timing synchronized mechanically to station sync. A full-frame TBC, often refered to as a "framestore", has enough memory to perform the above on a full frame (sometimes just a field), and can thus correct for both horizontal and vertical sync errors. They can also can be made to do scan rate conversion, stillframing, sampled-slo-mo, etc. You'll often see these in use for outside-broadcast work, such as "minicam remotes". Since they clock the input signal in from its sync, and output at station sync, they have the interesting effect of providing stillframes when a seagull lands on the microwave and the input signal disappears. You've probably seen remotes where the picture kept freezing; this is often the reason why. The resolution of the stuff we have here (for student-production educational use) is 512 pixels luminance, 128 pixels I-carrier, and 64-pixels Q-carrier per line. (I and Q are the Inphase and Quadrature subcarriers used to transmit chroma [color] information in NTSC.) Each is stored in 8 bits. It actually looks pretty good. - Brian
riner@avsd.UUCP (John Riner) (09/22/89)
Video time bace correctors (TBC's) are simply an elastic menory which reads in a signal with its source sync and reads out from another sync (usually a common to the rest of the system sync). The common implementation today is to A/D the input video and use digital FIFO techniques to read it out then D/A convert. This is very simplistic, the actual implementation can be varied depending on correction range (amount of memory and complexity of control logic) quality of A/D/A conversion and accuracy of analog circuitry.
ISW@cup.portal.com (Isaac S Wingfield) (09/24/89)
The first time-base corrector I ran across was in one of the early model quadruplex video tape units (probably Ampex VR-1000 or VR-1200) The TBC was implemented as an L-C delay line, with the C's being voltage variable capacitors. No, it didn't store an entire frame; maybe not even a whole line, I don't remember. Isn't history wonderful? Isaac isw@cup.portal.com