sprockets Modeling demo by Michael Brennan Short film by Svetlik Making and Using Drop-On Poses TinkeringGnome's Atomic Rings PRJ 2001 Star Gate effect in A:M with PRJ Comparison of AO and Radiosity Renders Animated Commercial by Soulcage
sprockets
Recent Posts | Unread Content
Jump to content
Hash, Inc. - Animation:Master

Mandelbrot Circle Slit-Scans and More


Recommended Posts

After doing the 2001 slit-scan project (https://forums.hash.com/topic/53303-2001-slit-scan-effects-simulated-in-am-with-mufoof/) using resurrected 2001 artwork (https://youtu.be/dujQGB-2EXw), I wanted to try some other graphics.  The Mandelbrot Set is a wonderful source of interesting imagery.  My college roommate, Mike Segor, who also went to that 1968 showing of 2001, had sent me some Mandelbrot work.  Grist for the mill.  He also contributed music for the final movie.

The artwork:

2001-slit-scan-artwork-mandelbrot.jpg

John Knoll has a fascinating video about slit-scan techniques (and shows a graphic of the Trumbull 2001 set-up at 5:07):

https://vimeo.com/391366006

He discusses various approaches including circular slits and other, similar, symmetrical geometry like the Dr. Who Tardis/police call box/tunnel sequence.

For this experiment, I decided to start with the 2001 slit-scan technique and then try circular slits.  Did not use the soft-edged slit approach as much rendering on this project was complete before that idea was tested.

The movie breaks down into these parts:
Vertical slit scan created on left half of screen, flipped and inverted onto right half of screen; Composite=Add.  Has the expected acceleration effect created by moving the slit-screen model element across the artwork during MUFOOF frames and moving the artwork between frames:

M1-slit-scan-509.jpg

Circular slit scan (full screen effect), artwork movement between frames; has sweep, is pretty, but no dynamic acceleration:

M2-slit-scan-10.jpg

Another circular slit scan clip overlaid on first circular slit scan clip with Composite=Difference; more dynamic because of counter-movement, but no real acceleration:

M3-composite-difference.jpg

Circular slit scan with slit-screen movement Right to Left during each MUFOOF frame, artwork movement between frames; has dynamic acceleration but is asymmetrical:

M4-move-screen-R2L-90.jpg

Circular slit scan, slit-screen movement diagonally Lower Left to Upper Right during each MUFOOF frame, artwork movement between frames; has sweep and dynamic acceleration but is also asymmetrical:

M5-move-screen-LL2UR-0.jpg

I made other attempts at a symmetrical set-up, but they did not give a dynamic acceleration effect:
1. Moving artwork away from camera faster than slit-screen (z-axis), exposing more artwork, in each MUFOOF frame; moving artwork between frames (x-axis).
2. Changing size of circular slit, exposing more artwork, in each MUFOOF frame; moving artwork between frames.
3. Moving circular slit from below centerline to above centerline in each MUFOOF frame; moving artwork between frames.  Has the virtue of being symmetrical horizontally, though not vertically:

Move-screen-down2up-8.jpg

4. No slit, just an open circle in screen and no movement of screen in MUFOOF frames; moving artwork between frames.  Has sweep but no acceleration:

Open-circle-no-slit-11.jpg

I wanted to try one more effect before giving up on finding a symmetrical set-up that would give dynamic acceleration — rotation of artwork 360 deg around the z-axis during MUFOOF frames, moving artwork left to right between frames, and using the open circle shown above.  This idea went down a whole new rabbit hole where I spent several days, but it led to what appears to be a generalized solution for simulating slit-scans that create dynamic motion.

Initial results looked very strange.  Then I realized that as the animation progresses, the artwork moves left to right as intended but the rotation point travels with the artwork and rotates it out of the camera FOV.  After thinking about that problem, I decided the solution might be an additional bone to rotate the artwork (which I named Pivot) while the artwork itself is assigned to a child bone of the Pivot bone.  That worked and the result has dynamic acceleration.  A still from the final part of the movie:

Open-circle-rotateArt360-untitled-407.jpg

Here is the movie:

 

https://youtu.be/aD8xUaaGIF4

 

Here are files for the open circle scan with pivot rotating the artwork 360 deg during each MUFOOF frame, artwork movement Left to Right between frames:

circle-no-slit-scan-MUOOF-backward-mandelbrot1-rotateArt360-sys3.chocircle-no-slit-art-sys-MUOOF-backward-60sec-rotateArt360-sys3.actcircle-slit-screen-backward-artwork-moveL2R-sys3.actcircle-no-slit-screen-art-mandelbrot1-system03.mdlcircle-no-slit-mandelbrot1-rotateArt360-sys3.pre

Mandelbrot artwork appears above.

But wait, there’s more.  After successfully moving the artwork in relative space using the parent Pivot bone and then moving it in absolute space using the child bone in the second action, I realized this approach could be used for any slit scan animation, including the original 2001 project.  Proved to be a bit more complex than that simple statement.  I revised the 2001 slit-screen/artwork model with a new parent bone (Pivot) for the Artwork child bone.  Now the slit would be stationary, the artwork would move 6” during MUFOOF frames, and the artwork would be moved an absolute amount between frames.  Test renders did not look right.  

In the old method, the slit-screen element moved and the artwork was stationary.  In the new method, the slit-screen is stationary.  Where should that slit be located?  For simplicity I decided to try changing the camera location on the x-axis.  Eventually I found that moving the camera +3” on the x-axis gave test renders that were substantially similar to the previous 2001 slit-scan results.  However, the image did not reach all the way to the center of the screen.  Time to modify the model and MUFOOF action to allow for further experimentation.

I added a new bone (System2) to the slit-screen/artwork model to be the parent of all model elements so I could move the whole model during the MUFOOF action. At action start, the model was located at -3” on the x-axis (and the camera was returned to zero in the cho).  I experimented with end-of-action values for bone System2; eventually +3” gave the desired result — these are the same values as in the original move of the slit-screen element, except now we are moving the whole model during MUFOOF frames.  This model move is analogous to panning the camera slightly during its travel down the track which Trumbull mentioned.

Here are revised 2001 slit-scan files using the new method:

soft-slit-scan-backward-2001artwork-60sec-PivotSystem2.chosoft-slit-screen-2001artwork-system2.mdlsoft-slit-screen-backward-2001-artwork.actsoft-slit-screen-MUFOOF-backward-60sec-movePivotR2Lsys2L2R.actsoft-slit-artwork01.tgasoft-slit-scan-backward-2001artwork-60sec-PivotSystem2.pre

2001-slit-scan-deconstructed-artwork.jpg

Here is the movie created using the new method which is substantially similar to the demo movie of the soft-edged slit-screen in the 2001 post:

 

As the I Ching says, persistence in a righteous course brings reward.

 

Link to comment
Share on other sites

  • Replies 0
  • Created
  • Last Reply

Top Posters In This Topic

Popular Days

Top Posters In This Topic

Popular Days

Posted Images

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...