David Skidmore
Well-known member
Answer this question first as it renders this post otherwise pointless:
Does the Epic 4.0.18 lock onto TC at a subframe level when fed externally?
If the answer is no, or don't know, then I'm hoping someone can lend insight to this issue:
We are shooting two EPICs running 4.0.18, each being fed TC from their own Lockit (model ACL202), running 23.98 across the board. We are not utilizing HDRx. Otherwise it's pretty standard, 5K FF, 8:1, 23.98, 256GB RedMags.
The sync is off by a single frame between 60-80% of the time. It's always off from the slates by at least by 1 frame, (which I'm assuming is the denicke electronics delay - as previously I've shot for 6 months with the slates being off by 1 frame from the recorded TC and no post complaints about sync).
It's challenging to diagnose with absolute precision in RCX as the frame doesn't magically line up with the clap of the slate as there is some crossover when you get in extremely close (perhaps being so close isn't helping), but I've adopted the general idea that if the slate is closed, that's the frame that the start of the audio clap should be in. I'm only being this precise to try and identify with impunity which clips are 'in-sync' and which clips are 'not in sync'. 1 Frame is acceptable in some circles, not so much in this one.
In some clips it's easy to tell when it's not in sync because the clap in the sound comes ahead of the frame of the closed clapper in the video, which in theory is impossible. It's practically indiscernible at 1x playback speed, but post keeps complaining that a huge chunk of the shots are off by a 'frame or two'. We don't have Lockits that produce genlock, so we aren't able to test the 'Genlock fixes sound sync' theory. In theory genlock shouldn't make a difference but I've learned to never discount something before trying it.
We ran through all the standard troubleshooting, swapped the lockits, checked the settings 3-5 times, jammed more frequently, but the problem persists and is infuriating in it's randomness. Borrowing from the genlock theory (and possibly grasping for straws), it's almost like the sensor and the T/C are not perfectly aligned at a subframe level and just slightly out of sync when REC is pushed leaving it in-sync sometimes and off sync other times. IE it's half way through a frame when REC is pushed, and it starts at the bottom of whatever timecode it sees 12:23:45 when the timecode could be at the top of 12:23:45 and 1/10 of a frame away from :46. My theory being that if genlock fixes it, it's timing the sensor exactly to the beginning of each frame, and when it's not genlocked, it could be starting anywhere during the frame.
This theory also lends itself to the sound clap occurring before we see it, because the camera could be anywhere from 1/10 - 9/10 of a frame behind. There are also plenty of anecdotal reports of genlock sync fixing these sync issues. Other anecdotal reports are happy being within 1-2 frames of drift which obviously is better than not having the lockit attached. haha
Maybe I'm out to lunch, but maybe, just maybe, we can finally say with impunity that a genlocked EPIC is the way forward for 'perfect' sound sync. Or it could be something completely different, because it sure doesn't explain being two frames out.
Does the Epic 4.0.18 lock onto TC at a subframe level when fed externally?
If the answer is no, or don't know, then I'm hoping someone can lend insight to this issue:
We are shooting two EPICs running 4.0.18, each being fed TC from their own Lockit (model ACL202), running 23.98 across the board. We are not utilizing HDRx. Otherwise it's pretty standard, 5K FF, 8:1, 23.98, 256GB RedMags.
The sync is off by a single frame between 60-80% of the time. It's always off from the slates by at least by 1 frame, (which I'm assuming is the denicke electronics delay - as previously I've shot for 6 months with the slates being off by 1 frame from the recorded TC and no post complaints about sync).
It's challenging to diagnose with absolute precision in RCX as the frame doesn't magically line up with the clap of the slate as there is some crossover when you get in extremely close (perhaps being so close isn't helping), but I've adopted the general idea that if the slate is closed, that's the frame that the start of the audio clap should be in. I'm only being this precise to try and identify with impunity which clips are 'in-sync' and which clips are 'not in sync'. 1 Frame is acceptable in some circles, not so much in this one.
In some clips it's easy to tell when it's not in sync because the clap in the sound comes ahead of the frame of the closed clapper in the video, which in theory is impossible. It's practically indiscernible at 1x playback speed, but post keeps complaining that a huge chunk of the shots are off by a 'frame or two'. We don't have Lockits that produce genlock, so we aren't able to test the 'Genlock fixes sound sync' theory. In theory genlock shouldn't make a difference but I've learned to never discount something before trying it.
We ran through all the standard troubleshooting, swapped the lockits, checked the settings 3-5 times, jammed more frequently, but the problem persists and is infuriating in it's randomness. Borrowing from the genlock theory (and possibly grasping for straws), it's almost like the sensor and the T/C are not perfectly aligned at a subframe level and just slightly out of sync when REC is pushed leaving it in-sync sometimes and off sync other times. IE it's half way through a frame when REC is pushed, and it starts at the bottom of whatever timecode it sees 12:23:45 when the timecode could be at the top of 12:23:45 and 1/10 of a frame away from :46. My theory being that if genlock fixes it, it's timing the sensor exactly to the beginning of each frame, and when it's not genlocked, it could be starting anywhere during the frame.
This theory also lends itself to the sound clap occurring before we see it, because the camera could be anywhere from 1/10 - 9/10 of a frame behind. There are also plenty of anecdotal reports of genlock sync fixing these sync issues. Other anecdotal reports are happy being within 1-2 frames of drift which obviously is better than not having the lockit attached. haha
Maybe I'm out to lunch, but maybe, just maybe, we can finally say with impunity that a genlocked EPIC is the way forward for 'perfect' sound sync. Or it could be something completely different, because it sure doesn't explain being two frames out.