#3DS book name contest!
9 messages in this thread
That's what I thought, but how would 3DS knwo how to generate the tweens needed
for the odd fields?
Your asking the wrong guy *that* question. I assumed that 3DS created both
fields all the time and just sent them in interlaced mode as "fields" when you
told it to. Not so?
DT
The best way to generate field-rendered video from pre-existing images is to
use the ASDG Exabyte->Abekas converter. If you render the images at 720×243,
and render 2x the number of frames, it'll interleave them together for you.
Works GREAT!
– G
Thanks! I'll check into it. For now, I think what I have will do. But if
someone actually pays for this transition, I may have to re-think things…
You're right, you can't do that. When 3ds renders in field mode, it renders
half of the keyframe time on one set of scan lines, and then moves a half frame
in time and renders the other set of scan lines to create a full video frame.
If you put your flic in the background and render fields, the _time_ data is
just not there. You'll get identical images on both sets of scan lines.
– Jack
Jack – What if you map your flic onto a big flat box? Would that work?
David Taffet
If the data ain't there, ain't nothin' gonna get it back. Here's the situation
in a microcosm: You make a flic of a ball moving. In the first frame in time,
the ball moves one inch. If you render to frames, that first frame contains a
single ball. If you render it to fields, half the scan lines show the ball in
one location, and the other half show the ball in another location. Neither of
those positions is the same as where the ball will be at the beginning of frame
2 (regardless of whether you're rendering to frames or fields), and neither is
the same as what you got when you rendered it to frame 1 in frame mode. Thus,
once you've rendered to frames, you don't have the field data that describes
the ball's movement, and you can't get it back, short of rerendering the
keyframe data in field mode.
– J
Thanks, Jack. I think I'm *beginning* to "get it" with the rendering to fields.
I "got" the part about the 2 interlaced images in 1, but am only now realizing
that 3DS interpolates motion to generate the fields (as it has to) rather than
just taking pieces of frames one and two and interlacing them (which is
initially what I thought was happening).
DT
Sounds like you've got it now.
– J