Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Creative Community Conversations Kinda hijacking the REELS thread from below…

  • Bill Davis

    October 30, 2012 at 1:06 am

    [Chris Kenny] “and Apple seems to be backing off of supporting them without having presented a comprehensive alternative.”

    Well, I actually see X’s robust range and/or file renaming functions to be precisely that “comprehensive alternative” but I suspect I define that term differently than you do Tom. Not saying mine is better or worse than yours – just necessarily different since we have different needs.

    I think that’s different than asking the software to accommodate any one arbitrary manufacturer standard (like a RED or Alexa format ReelID) that the majority of the software’s users might not have any need to ever use.

    I know that someone working for a network affiliate expects every operator to be labeling things to a “house standard” and to have a similar ID structure.

    But in my use of X it’s extremely common for me to use a dozen or more more timeline content sources. It’s not just digital tape roll-offs and DSLR and GoPro generated files, it’s those plus stock video clips, , buyout annimations, H4n audio files, stock sourced still photos and Zeus knows what else.

    What I’m saying is that in a well budgeted “shop” world, I totally know you want to track numbers which are consistently attached to every tape that enters your particular workflow.

    Out here in the “other” world of production, it’s nowhere so consistent and organized.

    The big question is whether the shops need to bend in order to accommodate the needs of the wider world – or the wider world has to bend to accommodate the needs of the shops.

    The lesson of FCP-Legacy – to my mind at least – is that you let the program do the best it can for the largest pool of users at first – then let it evolve to accommodate the needs of smaller constituencies as it can.

    This is tough for the professional industry, I think, because in the history of video – the pros got the tools first and only afterwards did those tools “trickle down” to the masses.

    FCP has always been an example of development in the other direction. It started with what was actually a very small subset of tools for DV Video editors (more “the masses” than the pros after all!) – and only evolved as a great tool for professional use over time.

    And here we are again. I think that X version 1 will be a lot like Legacy version 1 in that regard.

    But we’ll see.

    Know someone who teaches video editing in elementary school, high school or college? Tell them to check out http://www.StartEditingNow.com – video editing curriculum complete with licensed practice content.

  • Michael Phillips

    October 30, 2012 at 12:44 pm

    A few companies have tried this – a concept of “virtual KeyKode” Where KeyKode was a unique identifier for every frame in a production. In all my years of dealing with Kodak based production did I come across two frames in the same production with the same number. Fuji film was more apt for that to happen.

    The issue is settling on a format that encapsulates the identification as well as ensuring it is unique. RED cameras had this issue early on where two cameras could create the same file name till they started using the last 4 characters as a random hash. Things get complicated with formats that don’t use metadata in the header (TIFF) where both historical and versioning nomenclature is contained within the name of the file. Then there are formats that can contain metadata in the header and is not used by all processes. Not to mention the fragility of just the filename on its own.

    A good example is BMCC that only uses the folder containing the DNG files to be unique, but every file within those folders are identical to any other file – and there is nothing in the header to identify them after the fact. On the files I have been getting, I use a file renamer to at least take the folder metadata and apply it to the files themselves. Also, all user metadata is contained within the XML portion of the WAV file – be sure to record audio!

    I digress – some of these companies got pretty far in their proposal, but the maintenance and logistics of assigning and such was too much for one company. These are one of these things that should be part of a SMPTE type effort, and probably is – haven’t checked. But as with all standards, it takes years to get everyone to agree and get on board. And speaking of timecode, SMPTE should also start allowing for timecode beyond 23 hours – for production it makes sense, but for archival and post processes it would be nice to have an additional 76 hours of unique identifier in timecode as well as define a 50 and 60fps timecode – and maybe a 24fps DF, not that we need confusion, but cutting 23.976 for broadcast it would be helpful to not have to add an additional 30fps DF ruler to the UI.

    Michael

  • Chris Kenny

    October 31, 2012 at 4:57 am

    [Michael Phillips] “A few companies have tried this – a concept of “virtual KeyKode” Where KeyKode was a unique identifier for every frame in a production. In all my years of dealing with Kodak based production did I come across two frames in the same production with the same number. Fuji film was more apt for that to happen.

    The issue is settling on a format that encapsulates the identification as well as ensuring it is unique.”

    Obviously getting everyone on board with the same solution is quite non-trivial, but the technical problem is pretty simple. Just generate a full-on UUID for the first frame of every captured clip, and increment it by one for each subsequent frame of that clip. That gives you unique frame IDs — just just on a given production, but globally (from Wikipedia: The probability of one duplicate would be about 50% if every person on earth owns 600 million UUIDs), but they’re still sequential, meaning that an app trying to figure out which of a bunch of media files contained a given frame would only have to read the first and last frames’ UUIDs out of each file it was examining.

    Of course UUIDs are not easily human-readable, being 128 bits long and usually represented in hexadecimal, so you’d probably still want to keep more human-friendly identifiers like reels and timecode around. But when push comes to shove, software should have a reliable way to uniquely identify frames.


    Digital Workflow/Colorist, Nice Dissolve.

    You should follow me on Twitter here. Or read our blog.

  • Michael Phillips

    October 31, 2012 at 1:05 pm

    Exactly – the format takes on the concept of what TinyURL does for long URL names. It is then up to the system (NLE or other) to interpret as needed on a per frame basis. We are seeing more and more frame based metadata where many vectors change on a per frame basis coming from the lens and motion control world with similar challenges and solutions.

    Michael

  • Walter Soyka

    October 31, 2012 at 7:18 pm

    [Michael Phillips] “I digress – some of these companies got pretty far in their proposal, but the maintenance and logistics of assigning and such was too much for one company. These are one of these things that should be part of a SMPTE type effort, and probably is – haven’t checked.”

    SMTPE is in fact doing quite a bit of work around metadata, especially in MXF. I don’t know how long this will take to actually filter into production, though.

    [Michael Phillips] “It is then up to the system (NLE or other) to interpret as needed on a per frame basis. We are seeing more and more frame based metadata where many vectors change on a per frame basis coming from the lens and motion control world with similar challenges and solutions. “

    And as with medicine, the first rule should be “do no harm.” Don’t use it if you don’t need it, but don’t muck it up for other systems later in the tool chain.

    As I think more about how I might use reel/tc metadata for synthetic images in my own workflows, I’m realizing that it could be very useful for effects and compositing systems to allow source metadata from camera images to flow through.

    In other words, uniquely identify the synthetic frames, but also preserve the source footage’s UUIDs so that the provenance of the image is knowable to any software that cares. This could get pretty complicated, but it could also be very powerful.

    Walter Soyka
    Principal & Designer at Keen Live
    Motion Graphics, Widescreen Events, Presentation Design, and Consulting
    RenderBreak Blog – What I’m thinking when my workstation’s thinking
    Creative Cow Forum Host: Live & Stage Events

Page 2 of 2

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy