Elktrekker 0 Posted March 11, 2015 Okay, I've got an audio sync question for you all. I can't go into detail about the situation, but it's an object coming into contact with a vehicle. I'm transitioning into forensic A/V work, recreation, that sort of thing. If you could help educate me, I would appreciate it. Scenario: 8 cameras multiplexing (assumed), recording 7-8fps of each camera at 2CIF. Audio is only recorded on the first camera. Using an IDIS clip player to play back the cameras. The video information says it is Timelapsed. Three questions humbly asked... 1) Is my math correct in that the frames per second is 7-8 per camera, multiplied by the 8 cameras, that it is recording approximately 60 ips? 2) If the video is multiplexed, how is the audio recorded. Separately? 3) I have determined when the audio hit the vehicle based on marks, etc., but the point of impact does not line up with the audio (examining in Premiere, looking at the audio waveform)... it seems to be a couple of frames ahead of the video). I'm assuming this is due to the missing frames of video in relation to the audio file. My question is, is there a way to know how many frames it would be off and whether it would be in front or behind the impact frame? Thanks for your help. Share this post Link to post Share on other sites
SunnyKim 2 Posted March 12, 2015 Lip Sync is a tough job. I do not know how the system is composed of what kinds of chip set. The first thing top off my head is to ask IDIS to make a programmable audio buffer delay option. In general, audio and video keeps flowing into the system. If the sampling digitizer (ADC) for audio and video signal is driven by the different clock source (crystal), timing mismatch can be observed in the long run. If the camera chip handles video and audio together, it would be a lot better off. The problem still appears when decoding (playback) audio and video. The decoding video uses different amount of buffer from that of audio. The following is my explanations to your questions. 1. 7 or 8 fps seems to drop frames because the back end processor is not powerful enough of handling 25/30 frames of incoming videos. For 8 cameras, yes, around 60FPS can be recorded. I agree to your math 2. audio can be recorded as pcm, just digitized samples -no compression, as the data amount is relatively smaller than the video data. These can be recorded separately or together, depending on the system makers. Also time stamp for synchronization could be present at the recorded bit stream. Share this post Link to post Share on other sites
Elktrekker 0 Posted March 12, 2015 Lip Sync is a tough job. I do not know how the system is composed of what kinds of chip set. The first thing top off my head is to ask IDIS to make a programmable audio buffer delay option. In general, audio and video keeps flowing into the system. If the sampling digitizer (ADC) for audio and video signal is driven by the different clock source (crystal), timing mismatch can be observed in the long run. If the camera chip handles video and audio together, it would be a lot better off. The problem still appears when decoding (playback) audio and video. The decoding video uses different amount of buffer from that of audio. The following is my explanations to your questions. 1. 7 or 8 fps seems to drop frames because the back end processor is not powerful enough of handling 25/30 frames of incoming videos. For 8 cameras, yes, around 60FPS can be recorded. I agree to your math 2. audio can be recorded as pcm, just digitized samples -no compression, as the data amount is relatively smaller than the video data. These can be recorded separately or together, depending on the system makers. Also time stamp for synchronization could be present at the recorded bit stream. Thanks, that does clarify things. I'll see if I can get more information about the processing chip in the camera, but your explanation seems to match what I'm experiencing. My only problem is trying to figure out if there is a definitive way of telling how many frames off the audio will be. I can only approximate. Thanks for the info. Share this post Link to post Share on other sites
SunnyKim 2 Posted March 12, 2015 No way to predict the timing mismatch. There could be chances: Video processor at back end may drop a video frame when it feels a short of processing power to restore the processing power back. So the interval of skipping frame can not be fixed all time. Most of all, the crystals driving audio chip and video chip can be different, even though they are both claiming, say, the same 1 Mhz. But any commercial crystals can have their own PPM numbers, physically drifting all time. Share this post Link to post Share on other sites
Elktrekker 0 Posted March 12, 2015 No way to predict the timing mismatch.There could be chances: Video processor at back end may drop a video frame when it feels a short of processing power to restore the processing power back. So the interval of skipping frame can not be fixed all time. Most of all, the crystals driving audio chip and video chip can be different, even though they are both claiming, say, the same 1 Mhz. But any commercial crystals can have their own PPM numbers, physically drifting all time. This information has been very helpful and time-saving. Thanks, so much. Share this post Link to post Share on other sites
