Warning: A non-numeric value encountered in /home/customer/www/dannylaceyfilm.com/public_html/wp-content/plugins/social-warfare/functions/social-networks/googlePlus.php on line 73
It really was my worst nightmare come true.
We spent over 17 hours filming Host, in one day. Lots of footage shot on the beautiful RED EPIC at full 5K resolution, working with DP Philip Bloom, a dream. That was until I came to sync up the sound and discovered there was serious issues with EVERY shot we took.
In this latest video blog I’ll explain in detail what the frame dropping issues were and how I’m going to try and overcome this problem to lift this project from near disaster.
One or two frames we dropped from EVERY second of footage we shot on the EPIC. Something we couldn’t have spotted during filming. There was a serious fault with the camera, which Philip Bloom has since sent back to RED to be fixed.
Watch this space to see how we get on with editing Host. All is not lost just yet!
You can also keep up to date with the films progress on the Facebook page http://www.facebook.com/hostfilm
Thank for stopping by and please don’t forget to support my Movember campaign, growing a mustache in November to help raise what I hope will be £1,000 for The Prostate Cancer Charity. You can donate on my page here mobro.co/dannyfilmtash
Surely the dropped frames will show as a stutter in the footage
(If you remember the 30p 5d I did a lot of experiments on dropping and blending frames to get 25p .. most were disasters)
Say you have frame 23 and 25 it might be worth creating a 24 being a 50% blend of both – I guess that may look better than dropping
How you find those frames or automate the process in beyond me
Then you get your synch back
What I would do is a pile of tests before moving down any specific road
That's a very good point and you raise some interesting suggestions. Looking at all options right now to see what can do the best job at attempting to repair the film.
So sorry to hear this Danny. I think you should get RED to pay for a reshoot, after all it is their fault the camera didn't work properly?! You will notice the dropped frames sadly. Sorry mate, I've had dropped frames before and it's a nightmare.
Good luck and look forward to seeing the film.
Nice thought, not sure they'd be so quick to part with any money though.
I am a little worried that the dropped frames are going to stand out way too much. We'll soon see.
Thanks and best wishes, Danny.
Not sure what NLE you will be using but it might be worth trying to conform to a higher frame rate "stretching" the video just to get sync, then reverse telecine to try and recover/recreate frames. I hope you stomach has stopped bottoming out. That totally sucks.
If the frame dropping is happening in a strictly predictable manner, say every 15 frames or so, I maybe have a solution to your problem. (I'm a VFX guy)
It involves an automated script that recreates a new frame by looking at the frames before and after the drop, using Twixtor. I just tried it on some footage which I deliberately cut up to make it drop frames.
The result is near perfect, if you just watch it play there is nothing out of the ordinary to see. If you pause the footage on the recreated frame and A/B it against the original there is only difference in small details. Much, much better than dropping or frame blending.
If you're interested in me helping you fix this problem please let me know…
I think the positive publicity could be good for them (RED), the last thing they'll want is the reputation of their best camera scrutinised. PB has such a reputation it's worth a try.
Nothing ventured etc…
The real question is whether there is a stutter indicating that frames are actually dropped.
The appearance of a missing timecode on the clapperboard may simply be due to the frame rate on clapperboard not being precisely the same as frame rate on camera – in which case no frames are dropped at all and the camera is working fine.
Unfortunately, after looking at the shots closely, we can clearly see the dropped frames as the movement on screen jumps further than a normal frame to frame would.
The camera is definitely faulty and is now in for repair with RED.
It can't be worse than piecing together Metropolis. =)
RED doesn't care about their reputation. They never have…
Check your time base on the EPIC. It defaults to 23.98, so if you're shooting 25fps, it will conform it to 23.98. The time base it what you want the footage to conform to, giving you more options for overcranking.
For example, if you were shooting something at a 24p time base, and shot 48fps, it will obviously slow the footage down in half. If you're shooting 25fps on 23.98 time base, you'll get the same effect by fractions of a second (the problem you're having).
Definitely look into an optical flow solution to re-creating the dropped frames. It'll be a pain in the butt to set up but will look much better than frame-blending. Find an old copy of Shake, or use Nuke or Twixtor or… well there's tons of options really.
[…] https://dannylaceyfilm.com/2011/11/major-technical-issues-with-footage-from-host-vid… […]
Looks like the project timebase was set at 24, and the framerate was set at 25. The RED may be fine.
Thank you so much for sharing! Was this Phillips first venture (with dialogue) with the camera? I am sure that he has not had the camera long and with his tendency to shoot more non scripted footage (i.e. time lapse work) I am assuming he had not run into this problem so much shooting that type of footage? Also, I am wondering if this would have been the same occurrence if the film had been shot at a lower frame rate (i.e. – 23.98)? Have you heard anything from him concerning this at all? Anyway, I have been considering shooting the EPIC and this gives me some insight on how to test the camera before shooting with it. I shoot frequently on the MX and I have become so comfortable with the camera that I am sure that comfort would have caused me to be somewhat lackadaisical in testing the EPIC. So thank you again for posting this video! Hoping for the best with your footage and film.