Page 2 of 4 FirstFirst 1234 LastLast
Results 11 to 20 of 33

Thread: LaserBoy on Saturday the 14th!

  1. #11
    Join Date
    Apr 2007
    Location
    Toronto Canada
    Posts
    1,120

    Default

    This last image just made my 3ds max crawling like a dead turtle!
    Can you loose 90% of all details. There are just too many of them and there is no galvos in the world which can do that. BTW I finished recording TraceIT tutorial and 3ds max to Illustrate to Laserboy tutorial just have to resize them and post them on the web.
    I hired an Italian guy to do my wires. Now they look like spaghetti!

  2. #12
    Join Date
    Jan 2006
    Location
    Akron, Ohio USA
    Posts
    1,754

    Default

    I see what you're talking about...

    You're looking for a smart data reduction algorythm. I guess if you can describe exactly what conditions to look for in the data you can do whatever you want to with it.

    It's definately something to look at.

    One thing that helps is to select the whole thing and make it all one color. Then when you do the optimizations it puts together as many vectors as it can.

    James.
    Last edited by James Lehman; 06-15-2008 at 21:55.

  3. #13
    Join Date
    Jan 2006
    Location
    Akron, Ohio USA
    Posts
    1,754

    Talking

    I haven't had my head really into the guts of LaserBoy for a while as I got most of that to work well years ago.

    The biggest realization was the need for something between a vertex or a vector and a whole frame. I came up with the "segment" A segment can be anything from a single vertex to several vertices in a row. A frame is a segment, but it is also possibly made up of many segments. Each vertex is a possible segment. Obviously a good place to split up segments in a frame is to look for the strings of lit points between the blank spans. A segment as a C++ class can be smart. It can tell you things about itself, like it's first lit anchor location, its last lit point, its height, width, depth, center, top, bottom, front, back, left, right, and... it knows how to reverse its order so that it can be stroked in the other direction.

    I believe that my algorithm goes something like this... find the start or end of a segment that is nearest the center of the universe, look at it's other end and find the nearest point that is part of any other segment. If it is the first point of the next segment, take that segments as it is. If it's the last point of a segment, reverse the direction of that segment and take it. Do the same thing with the other end of that segment and so on...

    I do remember I was struggling with a super optimization idea. That's why you can fracture and bond all segments. My thought was that if you put two closed polygons near each other, they should each be drawn with a segment that begins and ends on those points in each polygon that are nearest, thus minimizing the blank span between the two objects.

    I think I ended up finding that you needed to know the best first vertex to start and that was not apparent until you were all done... or something like that. I'd really like to get back into this stuff. It's fun!

    James.

  4. #14
    Join Date
    Jan 2006
    Location
    Akron, Ohio USA
    Posts
    1,754

    Default

    Ooops! found another bug. You can not traverse backward through the first lit segment and you can't color the first segment in the palette transform menu. Crap.

    Oh well. It's 3:03AM. Good Night!

    James.

  5. #15
    Join Date
    Apr 2007
    Location
    Toronto Canada
    Posts
    1,120

    Default

    Quote Originally Posted by James Lehman View Post
    I see what you're talking about...

    You're looking for a smart data reduction algorythm. I guess if you can describe exactly what conditions to look for in the data you can do whatever you want to with it.

    It's definately something to look at.

    One thing that helps is to select the whole thing and make it all one color. Then when you do the optimizations it puts together as many vectors as it can.

    James.
    If there were a smart color reduction algorithm I would be more then happy.
    Here is a picture what I have in mind. First one how image can be converted into smart colors and the second is how Traceit traces both original and smart colors image.

    I know there are few cartoonizers like IntoCartoon but they don't do batch conversions.
    Attached Thumbnails Attached Thumbnails jdex.jpg  

    convt.jpg  

    Last edited by Dr Laser; 06-16-2008 at 01:30.
    I hired an Italian guy to do my wires. Now they look like spaghetti!

  6. #16
    Join Date
    Jan 2006
    Location
    Akron, Ohio USA
    Posts
    1,754

    Default

    That is pretty neat!

    I know you discussed this in your tutorial, but imagine if you made some video with the express intent of turning it into a laser trace. You could use a black background and florescent colors, lit with black light to make them really pop out and be fully (over) saturated. You could even put base makeup on the people to light up their faces and stuff!

    James.

  7. #17
    Join Date
    Dec 2006
    Location
    Pflugerville, TX, USA
    Posts
    1,977

    Default

    ^ Huh? That's how I look on a daily basis. Are you making fun of me.

  8. #18
    Join Date
    Apr 2007
    Location
    Toronto Canada
    Posts
    1,120

    Default

    Hmmm... the problem is a color shades. Traceit recognizes them as different colors. I wish there was a program that can see different shades of one color and convert it into that 1 color. For example you have a yellow skirt. But because of lighting it is not 100% yellow. It has variations of yellow and red. So the program recognizes that yellow/red area and changes variations of yellow and red to solid yellow. I know it can be done.
    I hired an Italian guy to do my wires. Now they look like spaghetti!

  9. #19
    Join Date
    Jan 2006
    Location
    Akron, Ohio USA
    Posts
    1,754

    Default

    Quote Originally Posted by Dr Laser View Post
    Hmmm... the problem is a color shades. Traceit recognizes them as different colors. I wish there was a program that can see different shades of one color and convert it into that 1 color. For example you have a yellow skirt. But because of lighting it is not 100% yellow. It has variations of yellow and red. So the program recognizes that yellow/red area and changes variations of yellow and red to solid yellow. I know it can be done.
    That is exactly what I am talking about! Use primary colors of flourescent pigments. Well lit with a lot of UV, they will glow; actually give off light of their own. This could easily over expose a camera so that it sees almost a solid color with no shade. Once you have it that close, cleaning it up digitally should be easy.

    BTW I have a really old (Windows 3.1 era) picture editor that lets you do a "median cut" color reduction to any number of colors from 1 to 256. It's really neat looking. It globs colors together based on how close they are to each other and keeps going until there are only the number of colors you want. The end result looks a lot like a paint-by-numbers picture.

    James.
    Last edited by James Lehman; 06-16-2008 at 12:09.

  10. #20
    Join Date
    Jan 2006
    Location
    Akron, Ohio USA
    Posts
    1,754

    Default

    Quote Originally Posted by carmangary View Post
    ^ Huh? That's how I look on a daily basis. Are you making fun of me.
    Really! I image you as a skinny guy with a huge head who just can't stop dancing.

    James.

Posting Permissions

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