Very Computer

This may be slightly off-topic, so apologies if you're just not
interested.

I have rendered a sequence of images using BMRT2.6 as part of an
animation running at 100 frames per second. That's a high frame-rate, I
know, but my animated objects are pretty nifty too, and in the absence
of motion blur in BMRT2.6 I thought I should run at that speed to try to
compensate. However, I'm finding it difficult creating a digital video
which can run at that speed and maintain image quality, so I started to
wonder if there are any practical schemes I could employ to simulate
motion blur by combining multiple frames, and thus reducing the the
number of frames. For example, combining every 4 frames in some way, and
running the animation at 25fps.

Many thanks for any help provided.

John.

Cheers,

Moritz

Ah yes, 3Delight. I have been tempted by this implementation, but the
fact that it doesn't support area lights put me off (I was well into
using BMRT by this time). I wanted the arealights feature so that I
could get soft shadows (real ones, not blurred shadow maps). So it was a
choice between motion blur and soft shadows; I choose the latter.

Anyway, I've now had a quick look at 3Delight's motion blur and I came
away a little disappointed. Maybe there's something I'm missing, but it
seems that both the Shutter (open & close), and the time interval over
which the Motion block is evaluated, are ignored. So motion blur occurs
over the whole frame interval (is that realistic?). The time sampling is
crude too, and I couldn't see any way of specifying the sampling rate
(in the same way I can set the rate for shadow sampling in BMRT). Anyone
out there managed to achieve smooth motion blur with 3Delight?

John.

Quote:> Anyway, I've now had a quick look at 3Delight's motion blur and I came
> away a little disappointed. Maybe there's something I'm missing, but it
> seems that both the Shutter (open & close), and the time interval over
> which the Motion block is evaluated, are ignored. So motion blur occurs
> over the whole frame interval (is that realistic?). The time sampling is

Quote:> crude too, and I couldn't see any way of specifying the sampling rate
> (in the same way I can set the rate for shadow sampling in BMRT). Anyone
> out there managed to achieve smooth motion blur with 3Delight?

Aghiles Kheffache

Quote:> > Anyway, I've now had a quick look at 3Delight's motion blur and I came
> > away a little disappointed. Maybe there's something I'm missing, but it
> > seems that both the Shutter (open & close), and the time interval over
> > which the Motion block is evaluated, are ignored. So motion blur occurs
> > over the whole frame interval (is that realistic?). The time sampling is

> Yes, sorry about this problem in the current implementaion. It will be
> fixed shortly.

Out of interest, what is the correct way to specify motion blur, keeping
in mind the operation of a real motion picture camera? Assuming a camera
running at 24 fps, then I would guess that each frame is exposed for a
period less than 1/24 seconds, depending on the aperature and the scene
being recorded, but never longer than this (though I don't know how that
could be physically done!). Is that right?

Quote:> Motion blur "smoothness" is directly linked to oversampling,
> specifying a higher 'PixelSamples' will produce better results.

John.

    http://groups.google.com/groups?ie=ISO-8859-1&as_usubject=sinc%20bmrt

Rick LaMont
Dot C Software, Inc.
http://www.dotcsw.com/

Quote:>Out of interest, what is the correct way to specify motion blur, keeping
>in mind the operation of a real motion picture camera? Assuming a camera
>running at 24 fps, then I would guess that each frame is exposed for a
>period less than 1/24 seconds, depending on the aperature and the scene
>being recorded, but never longer than this (though I don't know how that
>could be physically done!). Is that right?

So for starters, having the shutter open for half the frame time ought
to be about right.

I heard a story that "Tron" was shot with a shutter angle of 5
degrees, as the CGI elements couldn't be motion-blurred. This required
extremely high light levels; supposedly one night they left the stage
doors open and were visible from the Space Shuttle. I haven't been
able to confirm or disprove this, however.

<snip>

--
-Stephen H. Westin
Any information or opinions in this message are mine: they do not
represent the position of Cornell University or any of its sponsors.

Quote:> supposedly one night they left the stage
> doors open and were visible from the Space Shuttle. I haven't been
> able to confirm or disprove this, however.

There's another story that the City of Burbank threatened to shut down the studio
if they continued to draw 12,000 amps, but the matter was resolved in 45 minutes.

References:
    http://www.tron-movie.com/publications/american-cinematographer/
    http://www.tron-movie.com/p-kit/production_information.htm
    http://www.spaceflight.nasa.gov/shuttle/archives/year1981.html

Rick LaMont
Dot C Software, Inc.
http://www.dotcsw.com/