-
The Magnetic Timeline – What’s The Paradigm?
-
Craig, I see no reason why AV foundation versus quicktime should make any difference to how a GUI displays something as basic as a timeline. It may allow interesting new approaches to editing but as I just posted there are numerous ways to visually represent and manage assets in a timeline.
-
[Michael Gissing] “I see no reason why AV foundation versus quicktime should make any difference to how a GUI displays something as basic as a timeline.”
Nor should you. AVFoundation is Apple’s new framework for dealing with time-based media. It replaces their deprecated QuickTime frameworks and is Apple’s preferred framework for handling time-based media moving forward. It does four basic things: read media, play media, assemble media, and export media. It allows a developer to do these things without writing their own code to do so. There is nothing about calling upon AVFoundation that would necessitate the Magnetic Timeline.
Best,
Andy -
David,
i was compelled by the recent controversy and your general approach to join the forums and contribute some thoughts. i think this is a fruitful level of discussion, and important to editing as craft and art.
first i think it is important to to point out that the idea of time as linear is, in itself, a certain mental model which recording media in general (and film in particular) have contributed to. in fact, it is such a pervasive way of conceiving of time that most will accept it as a given without a second thought.
but i think for the purposes of the discussion it’s not a bad starting point.
i think the real question that your post ignores, however, relates to the marketing of the new software, and specifically the idea of the “trackless timeline”.
and so we should ask first – is this timeline really working without tracks?
i haven’t worked with the software. i’ve seen some demos and screenshots and read a lot of discussion. so far as i can tell, the software does, in fact, organize media in tracks. there are tracks of video. there are tracks of audio. sometimes they are married. they have borrowed the idea of multi-channel tracks and applied it to clips (some clips are stereo; some are mono; i assume that an 8 track HDCamSR would yield an 8-track clip – if you could figure out a way to get it into the software), but when there are two or more clips playing at once, they look very familiar – two parallel tracks. … correct me if i am wrong with any of that.
what it does seem that they have tried to do – and this is the new part – is make the management of tracks automatic. so marketing has called it “trackless” but in fact it seems to the “auto-track manager”.
this seems to me to be the fundamental question that editors have with the software – how well does it manage this task?
one thing that struck me about your first post was that you claim that, in the project discussed, there was a realization that “a real-world metaphor … was unnecessary” in designing the editing software. while i admit that i didn’t thoroughly read the accompanying document (very interesting though), it does seem to fundamentally rely on a real-world metaphor – that of tracks. tracks are physical models – after film reels and mag sound, or channel strips in a mixer (both references apply to NLEs, i think).
tracks as a paradigm of editing have been around for a long time – so far as i can tell, NLE’s inherited the idea from the audio world (or perhaps dual system projection), but the idea was expanded upon. tracks offer a very simple and yet flexible and powerful way to organize and arrange media. in fact, you will find many editors use tracks in different ways to solve different problems and organize their ideas and approaches and solutions in different ways.
it seems this software doesn’t really throw out a model – it just proposes that auto-management is a better way to do it (and the only way). and that is the question that it will be tested against – does it answer the needs that an editor has? does it offer the same flexibility and power as the old, self-managed model? does it offer more?
it seems that it was developed to solve the problem of clip collision. i think this is generally a problem for people new to editing.
personally i can think of many tasks that i wish were more (or better) auto-managed in FCP; track management wasn’t one of them.
Franz.
-
Thanks Craig. I’ll do my best although I hope this doesn’t mean I have to read the framework guide now 😉
To be honest, my understanding of AV foundation not deeply technical so I’ll be looking to you and other experts to fill in the gaps. I’m primarily interested in what AV foundation, i.e. an object data model means for FCP UI in broader terms. Where I go with it may not be where you expect but I hope it will be of interest.
_______________________
David Lawrence
art~media~design~research
propaganda.com
publicmattersgroup.com
facebook.com/dlawrence
twitter.com/dhl -
Great comment Franz!
You’ve touched on so many huge issues. I don’t have time to get everything now but I hope to address the question of what a trackless timeline means in a future post. I think that’s a key question and you’ve made some interesting observations.
Re: Physical world UI models. You’re correct, the document I posted very much depicts a physical world model. It went away in the next version.
Re: mental models of time — that’s a conversation I really need a beer for! 😉
_______________________
David Lawrence
art~media~design~research
propaganda.com
publicmattersgroup.com
facebook.com/dlawrence
twitter.com/dhl -
[Craig Seeman] “I’ve been trying to explain to people how AV Foundation impacted the fundamental changes in FCPX GUI/paradigm but couldn’t have said it as elegantly as you. Trying to model FCPX with tracks would have been an awkward and technology defeating shoehorn.”
Craig, I’m still not sure I follow here — what is it in AV Foundation that suggests a required magnetic timeline to you?
You have been so adamant about this that I’ve dug through the AV Foundation documentation a bit to try to understand this better. I suspect you’re looking at AVMutableComposition [link], which has the very magnetic-timeline-like features of being able to ripple-add and ripple-delete time ranges — but there’s also the lower-level AVAsset [link] and AVComposition [link] classes, in which tracks exist in specific time ranges almost exactly as they do in QuickTime.
In any case, aren’t these are just the frameworks that the applications use to read and write the media? Why would they necessarily dictate an application’s own internal data model for media?
What am I missing?
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 -
[Franz Bieberkopf] “it seems this software doesn’t really throw out a model – it just proposes that auto-management is a better way to do it (and the only way). and that is the question that it will be tested against – does it answer the needs that an editor has? does it offer the same flexibility and power as the old, self-managed model? does it offer more?”
The new model is clip relationships.
Let’s say you have some interview footage. The speaker mentions some subject, and you cut away to a shot of whatever he’s talking about. That cutaway is timed to… what? To an absolute time index within the sequence? No. It’s timed to the underlying shot of the interviewee.
In a conventional track-based timeline, the software understands only the former. Any relationships between shots are only implied (to the human editor) by the fact that the shots are positioned at particular absolute time indexes. The software knows nothing about such relationships.
In FCP X’s timeline, such shot relationships are explicit. The editor communicates them to the software in a way the software understands.
[Franz Bieberkopf] ”
it seems that it was developed to solve the problem of clip collision. i think this is generally a problem for people new to editing.”Avoiding clip collisions is not, by itself, the primary design consideration. It’s merely one instance where the software can act more intelligently because it now understands shot relationships. There are many others. Moving a clip, for instance. Connected clips come along. Or performing a slip edit; connected clips follow right along there as well.
But even beyond these behaviors, the clip relationship model simply makes sense conceptually. It brings the software’s understanding of the edit closer to the editor’s own mental model.
In many respects, FCP X is simply following what has been a consistent trend in software design over the last decade or more: the trend of adding more structure to data. Whether you’re talking about the ‘nav’ tag in HTML5, or Spotlight’s metadata importers, or IBM’s Watson QA system, creating structured data where only unstructured data previously existed is a major focus for computer science these days. Sometimes it’s done though automated data mining techniques. Sometime it’s done by having humans help computers out by adding a little structure themselves. Both of these approaches are used in various places in FCP X.
The more structure there is, the more the software understands, and the more the software understands, the more it can display and manipulate data in ways that make contextual sense.
—
Digital Workflow/Colorist, Nice Dissolve.You should follow me on Twitter here. Or read our blog.
-
I agree with Chris that the primary goal of the magnetic timeline is to make the implicit relationships between clips on a timeline explicit.
Since the universe could explode if I admitted to unqualified agreement with him, though, I’d add that I don’t think clip connections would be incompatible with a more traditional multi-track timeline. In fact, I think that arbitrary freeform clip connection on a multi-track timeline could be incredibly cool — though you would need some mechanism for resolving potential clip-track collisions on moves.
Looking at my own work in motion graphics and animation (also track- and time-based), I’d love to try applying the connection notion to keyframes, so moving one element in time could also move properties in other elements which should remain synced. As it stands now, I have to remember which keyframes in a timeline should be connected and which should remain separate — or I have to continually re-discover that through trial and error on more complicated shots.
David, thank you for starting a conversation on how our NLEs think, instead of just another thread on what we think of our NLEs. I can’t wait to see where this goes!
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 -
[Walter Soyka] ” I’d add that I don’t think clip connections would be incompatible with a more traditional multi-track timeline. In fact, I think that arbitrary freeform clip connection on a multi-track timeline could be incredibly cool — though you would need some mechanism for resolving potential clip-track collisions on moves.”
Exactly what I was saying about how the audio world has already solved this problem over 15 years ago by allowing clip layering within a track.
No need to have a magnetic timeline as it seems to me a clumsy way of solving such a non issue if you have non destructive clip overlaying on a track. But I see that no-one in the video world seems to want to take advantage of the problems solved so long ago by DAWs. Curious myopia methinks.
-
[Michael Gissing] “Exactly what I was saying about how the audio world has already solved this problem over 15 years ago by allowing clip layering within a track.”
That’s certainly one approach, and the more I think about it, the more I like it — but it would require the same sort of non-tracked rethinking that the magnetic timeline does. For example, not all supers could lie on the same video track if they were layered over other clips.
[Michael Gissing] “No need to have a magnetic timeline as it seems to me a clumsy way of solving such a non issue if you have non destructive clip overlaying on a track. But I see that no-one in the video world seems to want to take advantage of the problems solved so long ago by DAWs. Curious myopia methinks.”
Those who cannot remember (or never knew) the past are condemned to repeat it, right?
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
Reply to this Discussion! Login or Sign Up