General question re: horizontal motion and stuttering/skipping/jittering/etc

I'm working on a program right now that includes some graphic elements that were put together in Motion. The graphics include a lot of fine lines (graph paper), and horizontal motion. Here's a sample:

[dl.dropbox.com] (81mb quicktime)

In this particular example, there are two keyframes that control the horizontal movement--beginning and ending. Both keyframes are using Bezier interpolation. 29.97fps.

It might be difficult to see in the sample I posted, but in playback, we're seeing some stuttering/skipping especially on the horizontal movement. I've tried to slow down the movement somewhat to reduce the effect with some success, but am still seeing it. I see it when playing back the file on our apple tv and plasma display, I'm seeing it in desktop cinema preview in FCP 6, and I see it when playing back a quicktime from my mac pro.

Is it an optical illusion that I'm seeing because of the graph paper/fine lines and a progressive scan video? Is it actual frames being skipped in Motion because of the fast horizontal movement? Or is it something as simple as a bad render or errant keyframe? (I don't see any errant keyframes, and I've rendered several times, so I think I can safely rule that out) I'm rendering out of Motion using best quality prores 422, and dropping in a FCP timeline that doesn't require render in order to playback these files.

Would love to hear your ideas on this--even if those ideas include never using a graph paper treatment in conjunction with horizontal movement.

Clay,

It does no good to furnish us with a progressive and possibly resized version, and in a highly compressed delivery codec (H.264). That conversion itself might mask some of your observed behavior. Could you post a clip in the native size and format?
(You could try the following suggestion first and see if that helps.)

I looked at the clip that you posted. Even with the progressive H.264 format, I still see a bit of "shimmering." This may be an artistic choice, but why not blur the graph paper background? It probably doesn't need to be shown "literally." With a little Gaussian blur, you'll still know what it is but it may look a lot better.

And, yes, thin lines are like fonts with thin strokes or serifs. They all suffer from interlacing and may exhibit "strobing" behavior when moved at certain speeds.


-Dave

Sure, it's 314mb, but here you go:

[dl.dropbox.com]

clay Wrote:
-------------------------------------------------------
> Sure, it's 314mb, but here you go:
>
> [dl.dropbox.com]
> v

Thanks. DL-ing now.

You could also try a couple other things (still downloading)...

Enlarging the graph paper background should help with the thin lines. You could also consider rotating the graph paper a bit in the x-y (viewing) plane, or even tilt it in 3-D (around either the x- or y-axis, plus maybe a bit around the z-axis). Reducing the contrast between the white lines and the dark background may help, as well. Adding motion blur to the pan movement may also help (but would make things less sharp). Some of these may help to de-emphasize the background and lure viewers eyes to the inset movies, etc.

Joey and others here, who do a lot of motion graphics, may provide even better solutions...



-Dave

Flicker filter might help in FCP

mark@avolution Wrote:
-------------------------------------------------------
> Flicker filter might help in FCP

I tried this filter and it seemed to work the best out of a few "deflicker" filters I tried (I had to use the maximum setting to get rid of all of the flickering).

I did notice that your clip had no interlacing field order set (it seemed progressive?). Did you set your sequence to field order "None" and render that way (which would explain the exported test clip)?

The flicker severity changes with the speed of the horizontal motion.

(FYI, I monitored your clip via an external broadcast CRT HD monitor.)


-Dave

Field Order set to "none" in both Motion and FCP sequence. The clip I posted was direct from Motion, so I don't think that FCP would be causing any issues in this particular case.

I will probably go back and try to make some of the minor adjustments that were suggested to try and minimize the flicker effect in the original files. i suppose I could also try to slow down the move, but at this point, messing with the timing too much will just create more headaches, i think.

Thanks, and if anyone else has suggestions or ideas, feel free to post.

i haven't seen the clip,

but "strobing" on pans or moving objects is a real headache with not much of a solution if your movment is at the "wrong" speed.

in a film i cut last year we had a slow pan across a building with tall windows and columns, then back again.
not slow enough!
very distracting flicker.

the lab tried all sorts of tricks to add bur, and even slow the shot own slightly.
none of those tricks worked.
what alleviated the problem was reducing the contrast (dark window shutters against cream walls made the stobing really obvious)
still, a test film-out showed that the problem was still very severe.

finally the director had the idea to make the scene a bit more "day for night" reducing the contrast even more and hiding the problem in the shadows, so to speak.


watching "True Grit" the other day i saw strobing on the wide shots of a silhouetted horse running against an orange sky,
then the worst strobing i have ever seen in a released film on their end roller.
maybe my cinema was playing the print too fast, or maybe the cohens are as sick of 6 minute title crawls as i am, and just threw in the towel.


nick