Keeping Track of Footage pt.2

Ok, I’ve finally come up with a working framework for recording all my documentary footage information. It’s still evolving, but here is what I’ve got now.

My existing method requires the creation of the following files:

  1. Most basically, the tape log for each tape, a single summary-level document recording:
    1. title – actually more of an identifier, like “13J.2” for “2nd tape of footage shot on Oct 13th with the J camera”, the ID designed to sort properly on the file system. The name of the file will be based on this ID.
    2. summary – like “13 min – interview with John Doe; 25 min – interview with John’s mom”
    3. spreadsheet which is an excerpt from the footage summary file, below

    Since some tapes were used on several different days, or different locations, they have multiple names- one tape represents several “reels.” This is also handy if someone reset the timecode on accident.

    What this means is one file is one tape, but it will have a longer filename, like “13J.2-14J.1” – this will be aliased to look like two other files, “13J.2” and “14J.1”

  2. footage summary spreadsheet file. This is the giant overview of every tape- every soundbite, every notable shot, every time someone defines something is recorded in this spreadsheet. Each row represents a single shot, and records:
    1. reel ID as described above
    2. timecode
    3. take description
    4. good? meaning, “this is a notable shot which should definitely be sampled”
    5. soundbite? meaning, “this is a self-contained phrase or shot”
    6. definition? meaning, “this is a straight-up definition of something that appears elsewhere in the footage
    7. anecdote? meaning, “this is a longer take whose value depends on its entire content”

  3. tape-summary is another spreadsheet which condenses the footage summaries in each individual tape file into a spreadsheet. This is used for a reference when you can’t remember what tape a given interview was on- it’s much shorter than the footage summary, and you don’t have to open a million individual tape files

  4. list of tapes is for data on the physical tapes. Actually after I reviewed everything this list wasn’t so useful

  5. diagrams list is a list of all the diagrams and other graphics I realize I need for the final product, while I am editing.

  6. transistions list is for jotting down obvious transitions while reviewing. Like I’ll be watching something and think “wow that sounds really familiar- didn’t someone else say that?” and then look back in the log, and sure enough! In the final product those two clips can be edited to be close to each other.

  7. outline is the outline… duh. Actually I am not quite to this stage yet; I may end up using another system to write the outline.

That’s it!

You know what would be really nice? Management software tailored for documentary editing. Because unlike a narrative film editing process, where you have the script as a guideline, the documentary’s structure is potentially in flux the entire time you are editing. Since you don’t know what subjects will be discussed until after you finish filming, you can’t very well write a definitive outline before the shoot.

And lots of new material emerges during the shoot itself which you couldn’t possibly have predicted. So, simple editing software, and even screenwriting software is inadequate, and doesn’t really let you see the emergent themes that come out of the footage.

I was thinking something that records entries on reel, timecode, location, speaker, clip length, and a list of subjects that have come up in the footage. Each clip could record any number of subjects or themes, and then you would be able to sort on those themes so all the similar clips would be grouped together. Probably you’d also want to define shot attributes as well so you could edit clips together with a similar look.

ANTHONY: Use FileMaker!

In many ways, FileMaker would be ideal- it’s a database with multiple display options where you can lay out multiple forms and templates populated by the same database of data. There is probably some way to serve all this data too… the only problems would be

  1. setting up and maintaining a FileMaker server
  2. every client would presumably need FileMaker- not so bad if there is only one (local) client

Anyway, in the spirit of “not invented hereism”, how about this:


  • Clip data recording:

    1. reel ID as described above… note that I should have started a new reel every time the timecode got broken. Even if it was a gap of 10 seconds because a battery died!
    2. timecode as above
    3. date and time shot with multiple cameras, you really need this to keep the different shots on different tapes in order. Unfortunately we didn’t keep reliable notes during the shoot- maybe if we had the tapes with the chips in them, or if we just hit the “display” button at the beginning of every take.
    4. take description
    5. take location – I didn’t separate this from “themes” but really, it is a completely different attribute.
    6. themes, the “value add” in this application, a list of existing themes defined in previous entries, as well as the option to define new themes for this clip. Unlike in spreadsheets, this is potentially a list of multiple values.
    7. good? meaning, “this is a notable shot which should definitely be sampled”
    8. soundbite? meaning, “this is a self-contained phrase or shot”
    9. definition? meaning, “this is a straight-up definition of something that appears elsewhere in the footage
    10. anecdote? meaning, “this is a longer take whose value depends on its entire content”
  • Maybe a separate Theme editor, which would let you append notes and analysis to a given Theme (attached to and defined in the Clip, above)
  • Reel entity, representing the “reel” notion described above, basically being a collection of Clips
  • Tape entity, representing the physical tape (as above), containing Reels and displaying Clips, and also recording information pertaining to the physical tape.
  • Footage List – a browser of all these clips, sortable by various fields, including “show all clips with theme X”
  • A Sequence would be a list of clips – this is where the magic comes in. This Sequence is the core of the tape review process; you could drag the clips into a new sequence or an existing sequence, either as a placeholder to stage like-themed clips or to arrange them into their final order
  • Outline would have a bunch of line-entries which are either plain text (notes), or a Sequence or Clip. The Notes could be replaced by Clips or Sequences of Clips as footage review continued. Ideally, by the time you’d reviewed all your footage, you will have filled in all the entries, and you’ll have a rough cut outlined before sampling any footage!

Implementation Technology

  • Since it’s fast, easy to install, and free, the database would be MySQL
  • PHP forms the basis of the server end, as well as the presentation layer for web clients. PHPMyAdmin makes creating and maintaining the database easy, and dbForm/ADOdb is a good framework for fast deployment of forms which populate a database
  • PHP has a few SOAP / WSDL frameworks, to make this a bona fide web service.
  • Safari or Mozilla web browser for the front end to view, edit, and create data

  • To be extra slick, and since I still have yet to implement this two-headed system for any of my software, a Sherlock or Cocoa-friendly client. Given the requirement of low latency and the “drag” device for the better user experience, a dedicated client is preferable to a web browser.

Well great. I’ve succeeded in generating yet another project for myself which will probably never get done.

Okay, back to editing!