Timeline/Animator Revisited, Again ;)

I sitll want this so bad, I can taste it. When I have some spare time, I play with the idea in order to discover and define more what these items(s) are and how they should work.

It is possible, but complex, (using the Dialog Box, for instance) to simulate something that has an external transport, pseudo-timeline, and event recording.

The limitations and requirements for a timeline include:

* Realtime control and recording, hopefully with visual representation, (ala Flash). I am not sure if the recording has to be complex. In other words, it could just be a "Tap" of a button to note that something should happen at that point and then, the "playback" would be up to internal events You might be creating tirggers for a short cut-scene, or you could be working with an hour-long video, (now that a 16,0,16 background allows for objects to appear over video, this is even more applicable).

* Compactness of the "event file" that would most likely have to record even the passage of non-events somehow over the duration of the episode at a given time resolution. For instance, a period of no events could be a marker and time reference in the file, if you know hat I mean.

* Precision: There should be an internal precision that allows for triggers to be evented and executed within the resolution of MMF's own loop. As an example, it is very hard to trigger events using the resolution of the Direct Show object. You often have to allow for a "greater than" and then close the window on the trigger. There may also be a need to consider more than one form of synchronization, i.e., actual time, video position, etc.

* Visual Representation: The timeline/recorder needs to run as a window that opens during the application, obviously. It would have options that allow for selecting a file to play, (that could then be used by the appropriate object in action) ad saving/loading event files. The object would also be part of “playback” for the final aspect of the application the object was used with. The author would then use a list, (both numbered and with a time reference – possibly with more data like “event type”, etc.) and set their own triggers as the object plays back the file, (event: 1 Triggered, or Reference: 100200 triggered). Visually, it would be good to have a scrolling row of checkboxes, (simple or complex) that represent, at the minimal, where events are recorded. They should be easy to add manually and delete.

* Keyframes (optional in V1): Sometimes, you want to record a duration for an event. I would refer to those as keyframes here. You would record either single triggers or the start and end of a keyframe that could then be used as an on/off for the true of an event line in MMF, etc. While keyframes, in this sense, make it more complex, they add a greater degree of flexibility and control.

That is my update on the timeline idea and represents my wants/thoughts on this tool to date. Obviously, this is different from the old MMF timeline and represents more flexibility and realtime control. In a way, it is like merging the old timeline idea with the step recorder in a progressive way.

I did not go into animator here, but I wanted to mention it because it does relate, in form and function, to the timeline. I would often be using both together for specific purposes. We do need to find out if an “animator” is possible to do. I have spelled it out before, but animator would, in contrast to the timeline, be be able to control a defined set of objects in various ways. It too, should be able to allow realtime control from a panel, (speed, rotate, move, etc.) somewhat like a puppeteer's control. While it could be a PART of timeline, if not, it could respond to timing triggers:

Timeline Object Trigger 2 reached --> Play animation set “bird flight”.

I hope it is very clear where I am going and why this is important. Even if we don't get the object(s) a clear definition of the form and functions might be an inspiration for native tools in MMF3. As for the old, now removed, timeline, I was all for its being canned. These ideas are a level above it, with greater scope, function, and intent.

"external transport" - What do you mean by external transport?
I totally don't understand here.

"pseudo-timeline" - What do you mean?
Do you mean like to control behaviour?
@16.35sec start animation 1
@16.55sec start animation 5
etc

"event recording" - What do you mean?
Do you mean like a log file, eg
Collided with active(B) at 1,2 at 234.54sec
Collided with active(C-123) at 100,100 at 250.34sec
Moved from 10,10 to 100,100 @ speed 30.

(now that a 16,0,16 background allows for objects to appear over video, this is even more applicable)
This doesn't work on any of my computers, I don't know why. I compilied the standard examples and they don't play the video, so I wouldn't rely on it.


Hmm, we went away from timeline stuff, but I suppose an extension could be made to do it.

[:"blue"]
Here is my idea for a timeline extension
ie. you could have a setup dialog box made up of a grid
columb setup:
timeline set (string), time(from trigger time)(float), event number(interger), psuedo event name (short string), description (String), event on/off

Actions
- Start timeline x (where x is the set number)
- Stop timeline x
- Restart timeline x
- Next event timeline x
- Previous event timeline x
- Jump to line event timeline x
- Change time for line event on timeline x
- Repeat last timeline x event
- Add timeline x, at time, event number, psuedo name
- Delete(or turn off/on) timeline x, at time, event number, psuedo name
- Modify timeline x, at time, event number, psuedo name
(possibly swap events as well)
- Clear all error
- load next error into buffer
- Generate error user code
(save Timeline x to file)
(load Timeline x from file)
(oh and you could also copy timelines and offset timeline too)

Conditions
- Timeline x event n true/false
- Timeline x running true/false
- Timeline x started true/false
- A Timeline has started (any timeline) (negative as well)
- Timeline x line n occured tre/false
- Timeline x restarted true/false
- Timeline x completed
- Errors have been added to buffer

Expression
- Last Timeline x event number
- Number of events for timeline x
- Time
- Number of timeline sets,
- Psuedo name of timeline set n,
- (able to retrieve any one item from a time line)
- number of errors
- error number in buffer
- error description in buffer
- error time of error

[/]
Hopefully this is something that you could base your timeline idea on.

What do you think novabrain?

Thanks for the on-spot response <img src="/center/images/graemlins/wink.gif" alt="" />

I apologize for not defining my terminology.

A transport is a term used for a set of VCR controls, if I am not mistaken. It is generic for the play/pause/ff, etc. buttons.

The psuedo-timeline was simply a slider. A real timeline would be, to be, more lke a scrolling line of boxes than tic marks. How the actual timeline works would be similar, at least for the data structure, to what you illustrated. We need to consider that more.

I realize that the term "event" is often associated with MMF's lines in the editor. In this case, I mean it explicitly as in something that happens at a point. Perhaps the term should be changed, if needs be. I always suggest some consideration of wording used, but we have to start somewhere. I am open.

Thanks for outlining the c/e/a ideas. That is helpful to this goal.

I guess the extension running like a dialog box would work, as well.

As far as Animator goes, there are still open-ended questions about it and its relation to timeline. For the sake of ease of use, perhaps they could be individual extensions that have some interactive capacities? I am just guessing here and relating that to precision, as one can easily imagine using c/a/e's for this.

Also, for those who are new here, there are many potential uses for these extension(s): Cut-scenes, movies, illustrations for education, etc. The idea is to simplify timing and animation on a more complex level. Eventing by hand can become very time consuming and inflexible in this case, as most of us know.

Oh, and I am dissapointed to hear that the 16,0.16 is not universal. That spoils my excitment about it, since it is then impractical to use <img src="/center/images/graemlins/frown.gif" alt="" /> It makes me wonder, since it is there and does work on some machines, if there is a way for it to be exploited and made to work across-the-board? I will have to ask Yves and Francois.