Addressing Audio-Video Lip Sync at the Baseband Level
Part 5: Checking Lip Sync at Each Point Along the Signal Path
With your lip sync test setup ready as well as a survey of signal delay producers in your signal path, it's now time to go through the baseband signal chain step-by-step and measure audio-video lip sync errors, and insert corrections where needed.
To get an idea of the amount of overall delay (most likely audio) that will be needed, do a quick check of the final facility audio and video output that feeds the transmitter(s). If everything is okay here, you may decide not to continue. But it would still be beneficial to ensure that each stage of the signal path is in sync.
Starting at the beginning, look at the frame synchronizers. Is there an internal audio delay? Does it have a setting to automatically insert the same delay as video? If so, does that properly correct for lip sync errors? If not, ultimately, find out why. In the meantime, make manual adjustments to correct.
If there is no internal audio delay in the frame sync, then an external one must be added. Use your test audio delay to verify the amount needed.
Don't forget the studio, especially if you're using CCD cameras. Use a shot of a talker that has been close mic'd (like with a lavalier) and compare the audio and video. Are they in sync?
Take time with the video production switcher. Each mix effects bus setting and digital video effects can add video delay. Are there standard M/E settings for your productions? Try using GPI outputs of the video switcher to trigger specific audio delays for each setting. (Of course you will need an audio delay unit that can accept GPI triggers, and can change the delay settings in an imperceptible manner.) The audio to be delayed would be the final production mix from the audio room.
Check edited pieces. Did audio-video sync slip somewhere in the edit process? With nonlinear editors where audio and video can easily be adjusted on the timeline, it's not hard for one to get out of sync with the other.
Check recording devices like VTRs and especially video servers. Continue to follow a logical route until you've completed the entire signal path.
While this can be a time-consuming process, you will be assured that you have corrected for any lip sync errors under your immediate control and know that the signal is correct entering any MPEG encoder. What can happen in the compressed domain is a whole other story, but you can at least say, "It's fine leaving here."