FS7: The art of throwing away footage

P

Publimix

Veteran
Bad footage cost time and money, but what is the best practice to select the good stuff and get rid of the bad stuff?

I have terrabytes and hours of footage.
I try to keep my storage mean and lean.
1. That starts with shooting of course, finding a balance between enough and too much. Knowing what you need and stick to the plan.
2. selecting footage and decide what to keep and what to throw away. This is easier said than done. For this I use Catalyst Prepare.

But then I am always afraid to throw away the clips that I possibly need at the end of my project.

In a way I do this selecting in Davinci, I only fill my bins with footage that is good enough. But then the bad footage will be stored anyway.

Any thoughts?
 
Publimix said:
Bad footage cost time and money, but what is the best practice to select the good stuff and get rid of the bad stuff?


1. That starts with shooting of course, finding a balance between enough and too much. Knowing what you need and stick to the plan.

Any thoughts?
Click to expand...

When I first started shooting, the old sports guys used to say, "Only shoot the highlights". ; )
 
Storage is getting cheaper. B&H just ran a special on an 8TB Seagate USB3 drive for $139.00.
I am perhaps more sensitive to this issue than many since I shoot uncompressed CDNG raw 2K DCI, runs about 3 hours per TB at 24fps. Except for footage intended for specific fine art projects, I now compress the 12 bit linear camera offloads with Slimraw to lossless 10 bit Cineon log, roughly 2.5:1 compression. More than adequate for anything going to REC709 compressed delivery formats. Low value footage gets a lossy 7:1 compression for archival storage. I don't discard much, but shoot coverage like I used to shoot news and 16mm film.

Shooting news was a great discipline. One learns to think in terms of editing and story as events unfold. Picking shot combinations that move the story along and cover wides, mediums, closeups, cutaways, B-roll. Interviews get recorded completely, but the end product is likely to be a 90 second story at best with most of it being voiced over by the announcer with one or two sound bites a few seconds long in the mix.
With news we usually had at least a good notion of what the story was about going in, but we had neither time nor tape to just roll continuously. Didn't have the time to review or edit more than about 20 minutes worth of footage for a story working with the reporter. One 3/4" U-matic or Betacam tape per story. Before tape it was one 400 foot mag of 16mm, ten minutes. Fatal wreck on a mountain road in the dark, Shot or two of the rolled pickup off the side of a curve. Ambulance and sheriff's car blinking lights. First responders at work. A discrete shot of the covered bodies (never got used but we always shot one anyway). Couple of shots of the truck bed full of empty beer cans strewn across the road. Speech on the evils of DUI by the officer on the scene. Couple of good ole boys partied hardy and didn't make it home. Done.
 
Last edited:
It’s a bit like asking what’s the meaning of life...:happy:

Chances are, anything you delete, you might one day think “where was that clip?..”

Here’s my solution:

1) Get metadata straight from the start. Give clips a meaningful name in camera - establish a naming protocol and stick to it.

2) I cut in Premiere, so clip names are fully sorted in Finder before importing to PP. I use a free app called Name Changer to create sequenced names or prefixes/suffixes etc. This is my last opportunity to get the clip name meta data sorted.

3) In parallel with copying everything to my online edit scratch drive (G Tech 32TB Shuttle), I copy everything to a Sony ODA archive. These are 3TB once-write cartridges and I keep a set at the office and a set at home.

If you separate online, near-line and offline storage, you then don’t worry about keeping everything - and when it comes to restoring any edit project, it’s a simple case of copying the contents of the folder from ODA to scratch disk and clicking “relink footage”.

Ben.
 
I dont throw anything.. but for less important stuff I may not double/triple back it up.

You shlould also have a contract that you dont promise to deliver after initial delivery.. we are not an archive service for other people.

Of course when they do want it a year later charge to dig it out :)
 
I've had a few ideas about this very topic at work since we film a lot of marathons and races so there always ends up being a ton of footage that is pretty much the same thing over and over in the grand scheme of things. At the most recent race, one shooter was assigned to capture runners crossing the finish line so they just rolled the whole time and shot 13 cards. I think it was mostly 64 gig cards though but ~600+ GB of wasted space. They were not selective and didn't make an effort to get the best shots they could. The majority of it (like 95%) was bad footage and probably could have been the same result as an unmanned GoPro.

Our current workflow with these types of projects is to just pop out ProRes exports of our selects sequences. The original media gets archived to drives and LTO tapes. This works out pretty well for the most part but I've seen some issues with compression artifacts and banding pop up here and there, especially if the footage is underexposed. This might just be an issue if you're exporting S-log footage (we shoot S-log 3). I haven't seen it as much on Rec709 but I think the underlying issue is that the footage is being recompressed.

One idea I had to avoid those kinds of things is Premiere specific as it takes advantage of Premiere's Smart Rendering feature. Essentially that allows you to "passthrough" XAVC without recompression so long as you are not altering the files in anyway (lumetri color, scaling, etc...) in Premiere.

The idea is that you would cut your selects in Premiere in a sequence that matches your source footage and export that using the MXFOP1a container and codec that matches your source footage with the Smart Rendering checkbox ticked. This would essentially become your new "source" footage as it's ideally the good stuff without the bad stuff. The only annoying part is now it is all one file with no edit points. To sidestep this issue, you can flatten the sequence to one video track and export an EDL that matches your export. If you import this EDL it will recreate the flattened selects sequence with offline clips but they all have the edit points you originally made. The offline clips don't really matter in this case as we are only looking for the edit points. You can move these offline clips up a video track and place your selects export below to regain all of the edit points.

Then, if you really want to, you can potentially delete the original media to save on space. Only you can decide if the selects are good enough to make up for the fact that you would be deleting the original media.
 
My workflow varies from project to project, but generally I take a quick look at everything in Catalyst Browse as soon as possible and delete all the stuff I know I will never need while it is still fresh in my mind. For example, if I'm shooting wildlife or other uncontrolled b-roll I have a pretty good sense of what shots turned out to be nothing worth keeping. So I just delete all those garbage clips off the memory card before any backups are even made. If I have multiple takes of the same shot I keep the last one or two takes and delete all the rest -- knowing that if the first ones were any good I wouldn't have done more takes. Sometimes that isn't true, but a quick scan through any clip in question is fast and easy in Catalyst. LIke I said, I'm doing all this while it is still fresh in my mind. Once I've deleted the crap, then I make three backups of whatever is leftover and get to work.
 
You must log in or register to reply here.
Back
Top