Workflow for editing large projects?

[kye]

@joema thanks for sharing your methods and workflows

In a sense I'm thankful that I don't work with the footage quantities that doco folks work in - a shooting ratio of 600:1 is pretty killer!  Plus, the quality bar I have for my projects is exponentially less so there is that  

Part of the challenge I see is for NLE designers to provide methods for organisation and marking-up raw footage that is of sufficient flexibility and scalability to meet the challenge of the project.  The smaller and simpler the project the less comprehensive the tools need to be for doing it all within the NLE.  I also see there being a kind of scalability factor where below a certain size the built-in functions of the NLE will work, and above a certain size where doing things outside the NLE and even across multiple external tools is an acceptable overhead for the size of the project and the capability and capacity of the team of editors.  However I wonder if there is a middle ground where you have too much complexity for the NLEs integrated tools but the editing team can't afford the overhead of external tools and the additional admin overheads this creates.

This gap may not exist in FCPX or PP (I've heard that they're mature and well-featured editing packages) but I suspect that this might be where Resolve might be a bit lacking.  I'm not in the territory of this gap, I'm just kind of thinking out loud here.  My challenge (and my original call for comments) was that my challenge outstripped my previous technique of just doing a series of passes over the footage until I had an end product, so had to dip into some of the clip / timeline management tools available.

In future I'm contemplating doing highlight reels that cut across many years of footage and may end up including footage that didn't make it to a final edit of a previous project, so in a sense I'm kind of looking to a possible future project where my many hundreds of hours of footage are the source media.  Also, understanding a larger challenge normally helps when you're facing a smaller one, so there's both a longer and shorter-term benefit of this discussion for me.

[User]

Just found and heading into this:
 

 

[barefoot_dp]

I was going to go into a whole lot of detail about my post process for larger projects (I do a 10x23min travel TV show) but then realised I'd sound like a pompous wanker (which I am; I just would rather not sound like one), so instead I'll leave you with 2 tips:

PANCAKE EDITING. Once you've got your timeline with your rough selects on it, set up a workspace with two skinny timelines stacked on top of each other (like pancakes) - one with all your media, and one which is your working draft. Then you can quickly drop clips down from the top timeline to the one below, without ever having to go back to your project window or toggle between different timelines. Makes it really quick to scrub through and grab shots to drop into your edit.

SPREADSHEETS. If you haven't already, make a spreadsheet with notes on everything you did. Date, shoot day #, location, people, activities, cameras used, key events, hero shots, etc, etc. Normally I do this at the end of each shoot day while the footage is transferring. It will come in handy so often when editing bigger projects.

 

On 6/1/2019 at 12:28 AM, kye said:

I've heard of docs commonly being 100:1 and blockbuster movies sometimes go well beyond that IIRC, I remember a ratio above 400:1 which is just madness!

On big features, everything is planned in minute detail so shooting ratios are generally lower than on a doc. a 400:1 ratio would cop any director a serious grilling from the executives. Now obviously they allow time for rehearsals, dry runs, multiple takes etc, but as crew sizes go up, every extra run through a scene just adds to the bottom line. Of course certain scenes will take longer than others because they require more impactful & precise performances, but others are knocked over in one take to compensate.

[Thpriest]

Take care to consider the labels you give to any clips or files. It's amazingly easy to make life more difficult by labelling in a rush.

[kye]

4 hours ago, barefoot_dp said:

I was going to go into a whole lot of detail about my post process for larger projects (I do a 10x23min travel TV show) but then realised I'd sound like a pompous wanker (which I am; I just would rather not sound like one), so instead I'll leave you with 2 tips:

PANCAKE EDITING. Once you've got your timeline with your rough selects on it, set up a workspace with two skinny timelines stacked on top of each other (like pancakes) - one with all your media, and one which is your working draft. Then you can quickly drop clips down from the top timeline to the one below, without ever having to go back to your project window or toggle between different timelines. Makes it really quick to scrub through and grab shots to drop into your edit.

SPREADSHEETS. If you haven't already, make a spreadsheet with notes on everything you did. Date, shoot day #, location, people, activities, cameras used, key events, hero shots, etc, etc. Normally I do this at the end of each shoot day while the footage is transferring. It will come in handy so often when editing bigger projects.

 

On big features, everything is planned in minute detail so shooting ratios are generally lower than on a doc. a 400:1 ratio would cop any director a serious grilling from the executives. Now obviously they allow time for rehearsals, dry runs, multiple takes etc, but as crew sizes go up, every extra run through a scene just adds to the bottom line. Of course certain scenes will take longer than others because they require more impactful & precise performances, but others are knocked over in one take to compensate.

Cool tips. If you feel like you can share more (without crossing your personal PR threshold!) then please feel free  

In terms of ratios, I can't remember where I saw the 400:1 ratio but this gives a few with a couple of stand-out numbers...

I understand that knowing what the story is, having it planned, and knowing what you're going to shoot should give a lower shooting ratio than a doc (where none of that is planned) but obviously it's not the case for every film!  I don't know what would have made Fury Road and Gone Girl have ratios like they did...  it's nice to have options in post, but...  

[barefoot_dp]

12 hours ago, kye said:

 I don't know what would have made Fury Road and Gone Girl have ratios like they did...  it's nice to have options in post, but...  

Fury Road might be because of all of the practical effects & stunts, which usually get multiple angles of coverage as they're harder to repeat. 10 cameras shooting a single take = a 10:1 ratio, for example. So in that case, even though the shooting ratio is going up, the budget doesn't really change that much as they're not adding shoot days on to get the extra coverage.

Not sure about Gone Girl though - I guess just lots of takes trying to get those nuanced emotions?

My next biggest editing tip would be to work backwards from the end product. Do a paper edit, or edit it in your head first, before you start anything else (ideally you'd have the end product in your head before you even start shooting!). Decide on roughly how long it is going to be, what your climax is going to be, your opening and closing shots, what sound bites you want to use, what the pacing is, etc. Then once you actually start cutting, put the best shots/sequences in first in roughly the right places, then get your soundtrack & audio bed sorted out, then fill in the holes. If you work out the edit on paper (using the aforementioned spreadsheet) before you even start cutting, you'll be able to do the actual task of editing with a lot more purpose and the decisions will almost make themselves.

 

[KnightsFan]

In one of the BTS features for The Hobbit, Peter Jackson said that one side effect of digital was they would roll through multiple takes. It makes me wonder how much of that is a "usable ratio," or actual footage where the actors are acting, and how much is between takes, clapperboards, etc, how much is from multiple cameras as mentioned, and how much is from content that was cut or done in multiple takes.

[kye]

1 hour ago, barefoot_dp said:

Fury Road might be because of all of the practical effects & stunts, which usually get multiple angles of coverage as they're harder to repeat. 10 cameras shooting a single take = a 10:1 ratio, for example. So in that case, even though the shooting ratio is going up, the budget doesn't really change that much as they're not adding shoot days on to get the extra coverage.

Not sure about Gone Girl though - I guess just lots of takes trying to get those nuanced emotions?

My next biggest editing tip would be to work backwards from the end product. Do a paper edit, or edit it in your head first, before you start anything else (ideally you'd have the end product in your head before you even start shooting!). Decide on roughly how long it is going to be, what your climax is going to be, your opening and closing shots, what sound bites you want to use, what the pacing is, etc. Then once you actually start cutting, put the best shots/sequences in first in roughly the right places, then get your soundtrack & audio bed sorted out, then fill in the holes. If you work out the edit on paper (using the aforementioned spreadsheet) before you even start cutting, you'll be able to do the actual task of editing with a lot more purpose and the decisions will almost make themselves.

 

Yes, Multicam would sure bump that ratio up..  and combined with @KnightsFans comments below about rolling through multiple takes you sure could keep a DIT (or three) busy on set!

In terms of working forwards or working backwards, I've seen some conflicting approaches:

• One approach is to make a pass through all your source footage making selects and adding them to the timeline, then you duplicate the timeline and cull the duplicates and lesser shots, then rinse and repeat until you have a final edit.  The advantage of that is that once you've eliminated a piece of footage you never have to watch it again, so it's efficient in that sense.
• The other approach is to review all your source footage and flag the truly great clips, then start with only the great clips on a timeline and then only pull in whatever else you need to make a final edit.  This approach has the advantage that you're not watching and re-watching clips that haven't yet been culled.

In narrative it's a little different as you've got an idea about what you're aiming for (a script) but in unplanned / uncontrolled documentary-style shots (like weddings, docos, events, etc) it's hard to know where you're aiming in terms of which shots will make the final edit.  I've mostly tried the first approach, culling and gradually fine-tuning, and it works but it's pretty laborious.  I suspect that the second approach works much better if you're able to efficiently remember what clips you have and are able to find them quickly.  Some people are able to watch lots of footage and then remember it pretty well, but I guess that would depend on how much footage there was and the level of control - ie, it's probably easier to remember that there were three takes of that scene that worked rather than there were three reaction shots that might have made a great insert into a dramatic scene that you haven't even edited together yet.

1 hour ago, KnightsFan said:

In one of the BTS features for The Hobbit, Peter Jackson said that one side effect of digital was they would roll through multiple takes. It makes me wonder how much of that is a "usable ratio," or actual footage where the actors are acting, and how much is between takes, clapperboards, etc, how much is from multiple cameras as mentioned, and how much is from content that was cut or done in multiple takes.

I once edited a video of a race where one camera was an in-camera shot looking forwards and seeing the dash and out the windscreen.  It was a huge long file of multiple sections and the stuff in-between, and I was having trouble working out how to tell which bits were stages and which weren't.  That is, until I noticed that during the stage the RPM gauge was fully maxing out and between times the car was basically idling..  problem solved!  I would imagine that once you've worked out what to look for (like lots of crew on set) you could easily make selects.  It would also make it easier as you only have to sync the footage once then you can cut it up easily.

[barefoot_dp]

4 hours ago, kye said:

In terms of working forwards or working backwards, I've seen some conflicting approaches:

 

This is where the pancake timelines come in handy. On the top level you have the timeline that you're gradually culling so it's just a whole bunch of footage on a single track (maybe with some markers or colour labels to break up different sections or key clips). Then on the bottom timeline you have your clean slate that you can drag and drop things onto and build it from the ground up.

So you can essentially be doing both techniques at the same time.

[tweak]

I probably have a similar amount of footage to you on most occasions. The feature I'm working on now I shot about 3TB of footage over 3 months.

Steps I take -
I usually try and cull the bad shit from the start, once I transfer to hard drive I kill anything I know I wont use for sure and also try to give every clip a name.
Once I'm in the editor I make different timelines for different things, pull the shots I know I might use in there to just check them out and be easily able to see them.
If I've done the first step well the footage is already in folders on the hard drive that are pretty self explanatory. 
The last part is making seperate timelines (if the project is too big) of the edit and then putting them all together in one timeline at the end.

I dunno how people usually approach this stuff, but for the things and style I shoot this is the best workflow I can come up with because usually my videos are not narratives.