Advice needed: Editing as we are shooting

[Michael Jarvis]

Hey there,

So I have a music video shoot coming up and could use some advice. I got this idea from Chris Probst and his workflow on the Taylor Swift videos. The studio shoot material that we will be shooting will be one piece of the puzzle. We will be overlaying this footage over really great live performance material that we will already have prepared. We will already have soundlock and the live footage in place.

For the studio shoot, I want to set up a video feed directly from Cam A (most likely with a Teradek to the video village) and capture this content redundantly as it is being shot. I will have a team of producers at the editing table and will be cutting the video together as it is being shot. Later we will go back in and replace the proxy footage with the R3D footage.

My question is....what gear/methods would you suggest to make this the smoothest workflow possible? Primary editing suite is Adobe Premiere. The main cam will be a Scarlet Dragon.

What complications do you see and can you suggest anything that might make this process as efficient as possible? Have you ever used this workflow on set for a music video?

We will have a second computer on site dedicated to backing up media. One computer will be dedicated just to creating a rough cut of the video as we shoot it.

Any advice would be helpful.

Thanks

 

[Brendan H. Banks]

This is interesting. In terms of relinking, the biggest things coming to mind are your proxy naming convention, in/out points, and framerates.

1) You'll need to match your proxy file names to your master R3Ds so that they relink properly later. I think the live nature of the project makes this tough, but doable.
2) So if you just want to relink and have it all fall into place, you're going to have to have the in/out points be PERFECT. I am not sure if there are any wireless systems that pass-through a record trigger...seems possible but you might need a more robust VTR system.
3) This is all assuming you're shooting the exact same framerates that are being pushed out wirelessly. The second you go into higher than supported framerates (dependent on your wireless system) you will have real trouble getting the timing or relinking to work in any way that makes sense.

Those would be my flags and things to look into. Never worked this way but it's interesting.

 

[Michael Jarvis]

Hey Brendan....here is a quote from Chris. Sounds like he matches take and timecode manually. Also, to clean up my verbiage, we're not capturing live concert footage. We have live concert footage that will already be edited to the sound locked audio in advance of this shoot. We are just taking the studio shoot footage and dropping it into the structurally edited video where needed.

On the first day, there is a blank timeline. We know which sequences go where in the song. So we shoot a scene and drop it in that part of the timeline. It's all in Joseph's mind and masterplan, we just need to fill in each missing hole. At the end of the shoot, we hit play and watch the completed video to usually about 95% accuracy. On both Taylor videos, she was able to watch the rough cut, maybe request an alternate take for a different facial expression, and aside from the post work and vfx shots needed, we locked picture there and then.


We capture the videotap feed live and use that to edit immediately as we shoot, and then the editor on set gets transcodes as they come on from the on set DIT and replaces the videotap with the transcoded footage, matching take and timecode manually. This has proved the fastest workflow for us and enables Joseph to edit the split second after the camera cuts to confirm the precise timing/matching he's after.

 

[Brendan H. Banks]

Ahhh, gotcha. Sounds me to like they had Free-Run TC that embedded back into the wireless feed, which makes sense. If you can embed the timecode, that will help the syncing issue -- plus if you dont mind relinking by hand, that makes sense.

I would guess they needed to do this because of the way Swift runs her business -- she gets final say on all things and is at least lightly-versed in a lots of different mediums. I would say that unless you have these same requirements, not sure what is to gain from doing it this way. I might prefer that budget go into more time on set for options and a previz to make sure angles worked. That being said, I don't know the content and it's not my production, so it's not my call

If you have that embed and some post machines ready for the DIT, sounds like you probably have a solid situation on your hands. Though a proper test-day will be the ultimate decider in how it'll go on the day.

 

[Michael Jarvis]

Thanks for the advice Brendan. Really appreciate it.