Page 1 of 4 1234 LastLast
Results 1 to 10 of 33

Thread: frame optimization through ballistic precomputation

  1. #1
    Join Date
    Oct 2006
    Location
    Cleveland, Ohio
    Posts
    2,334

    Default frame optimization through ballistic precomputation

    Given that properly tuned scanners at, say 30k using the ILDA test frame, have an approximately equal impulse response, it follows logically that point drawing (and reduction) optimizations could be applied, especially to fine detail and small movements, that make use of points that are never reached. This concept is nothing new to the laser show industry, frames have been produced and tweaked manually that rely on this.

    In the open-hardware software arena there are products such as anarchy that specialize in automatic optimization. Do these software optimization routines make use of the ballistic point-pulling technique in order to increase framerate?

    I am thinking of implementing this in some software in the future and would like some feedback.. would you consider it to be 'too hard' on the scanners?

  2. #2
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,489,205

    Cool

    To the best of my knowedge, Anarchy actually does the opposite of what you're describing. That is, the automatic point optomization routine will *add* points to smooth things out. Thus, if you optimize the ILDA test pattern, you'll end up with a dodecagon outside the square when you scan it.

    I'm assuming that you're thinking about implementing some kind of "pulled-points logic" in your software, where the ballistic arc that the scanner would take if fed a set of widely-spaced points could be pre-calculated, and then the points of the file adjusted to give the desired curve with fewer points... That would be quite interesting if you could pull it off!

    It could be harmful for the scanners, however. The ILDA test pattern is OK when you scan it at 8 degrees, but if you max out your scanners, it's no longer OK. The same thing would apply to a lesser extent with the images this new logic could produce. My suggestion would be to start out very conservative and then move up from there.

    Have you already thought about how to pre-calculate the scanner response? (Or did you plan to "calibrate" the the software to the scanner response based on a series of test images?)

    Adam

  3. #3
    Join Date
    Oct 2006
    Location
    Cleveland, Ohio
    Posts
    2,334

    Default

    Yes, that's what I was thinking, adding this for smaller jumps and details so that it would not be harmful. Ideally, response from all properly tuned scanners would be the same, but I would implement a few test images to calibrate the algorithm parameters on a per-scanner basis.
    Another interesting aspect is that the original vector oriented 'intent' of the optimized ilda file could be reversed into a proper-looking frame on a computer, given the right algorithm and parameters. This would help to make ballistic optimized frames portable across platforms and scanner types.

  4. #4
    Join Date
    Dec 2006
    Location
    Pflugerville, TX, USA
    Posts
    1,977

    Default

    I have been thinking of optimization lately, as well. I had a different idea but it could be applied to what you are proposing. I was thinking that if I created a utility that displays points, lines, and angles of different sizes I could display them on my scanners at different speeds and then figure out a formula that would give me the best blanking based on distance, speed, and angle. This would allow for adding the correct number of anchor points, specifying the correct color shift, etc for any speed or step distance. I suppose you could somehow use this information to determine how hard to pull points as well.

  5. #5
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,489,205

    Cool

    Sounds like a cool idea, Doc. Unfortunately, it puts all the weight on your shoulders to develop the logic!

    Still, I'd love to hear more about it if you decide to move forward with this...

    Adam

  6. #6
    Join Date
    Dec 2007
    Posts
    93

    Default

    Just thinking out loud but, for a perfect square, what if you actually let it overshoot, and use blanking to make it perfect.
    ex: you want a square, but you actually draw a '#' kinda pattern. You draw the 2 vertical lines first and the 2 horizontal next, and you use blanking to get rid of the 8 little overshoots ?

  7. #7
    Join Date
    Oct 2006
    Location
    Cleveland, Ohio
    Posts
    2,334

    Default

    Quote Originally Posted by carmangary View Post
    I have been thinking of optimization lately, as well. I had a different idea but it could be applied to what you are proposing. I was thinking that if I created a utility that displays points, lines, and angles of different sizes I could display them on my scanners at different speeds and then figure out a formula that would give me the best blanking based on distance, speed, and angle. This would allow for adding the correct number of anchor points, specifying the correct color shift, etc for any speed or step distance. I suppose you could somehow use this information to determine how hard to pull points as well.
    Yes, a combination of ballistic point pulling and velocity based laser modulation adjustment could be close to an ideal optimization set. If we could get this in code, along with the reverse calculations it could provide a nice boost to the frame complexity achievable by common scanners.

  8. #8
    Join Date
    Dec 2006
    Location
    Pflugerville, TX, USA
    Posts
    1,977

    Default

    Agreed. I would like to someday get away from having to specify images as a set of points and only as a series of segments, beziers, and shapes and then let the software figure out how to draw it based on the characteristics of your scanners. And, your scanner characteristics coud be easily discovered by a wizard that simply shows pictures and asks you to choose the best one as it learns... the same as how an eye doctor uses many different images during your eye test.

  9. #9
    Join Date
    Oct 2006
    Location
    Cleveland, Ohio
    Posts
    2,334

    Default

    This is pretty basic stuff, engineers in the electrical field calculate transfer functions and compensate for non-ideal impulse response every day. There are only a few variables to look at, a wizard would not have to be too involved. Frankly I'm suprised this hasn't already been implemented in everything, including a section for parameters in the ILDA format. Either there aren't enough engineers in the laser show industry or we are overlooking a fundamental flaw.. ?

  10. #10
    mixedgas's Avatar
    mixedgas is online now Creaky Old Award Winning Bastard Technologist
    Infinitus Excellentia Ion Laser Dominatus
    Join Date
    May 2007
    Location
    A lab with some dripping water on the floor.
    Posts
    9,543

    Default

    Quote Originally Posted by carmangary View Post
    Agreed. I would like to someday get away from having to specify images as a set of points and only as a series of segments, beziers, and shapes and then let the software figure out how to draw it based on the characteristics of your scanners. And, your scanner characteristics coud be easily discovered by a wizard that simply shows pictures and asks you to choose the best one as it learns... the same as how an eye doctor uses many different images during your eye test.
    DocL, please send me a email so I can send you a small attachment you might find interesting.

    I also have the equation of state for a galvo scanner, I will scan it in for you as a JPeg this evening. I would caution all of you that your neglecting hysteresis in the bearing friction, and that DSP scanner amps have only made the civilian market in the last 3 years.

    Another technique is to accelerate the scanner to a position P, then apply a precise timed pulse of 1/2 the magnitude and the opposite phase at a time Tau later, thus slamming on the brakes so to speak and eliminating the need for damping.

    I also caution you that good laser artists, in order to keep overall intensity even in large complex animations, often add parts of the image that will appear in the future, as blanked points, so the scanner scans them, thus there is no popping in the graphics or dimming as the animation changes.

    Steve
    Last edited by mixedgas; 10-24-2008 at 08:27.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •