Topic: Total rewrite of wave optimization stuff!

[James]

For this version:

http://akrobiz.com/laserboy/code/LaserBoy_2009_11_06.zip

I've spent the last couple of days completely reconstructing the whole system of adding dwell to the data for wave optimization.

I have not put this much work into the core of LaserBoy for quite some time.

EDIT: This is not a good version! See below.

James. 

[drlava]

laserboy sounds like it's making some good improvements.

[James]

Ugh!

Never mind....  

I just tested it some more and it's all screwed up.  

Oh well.

I'm on the right track. I just need to keep working on it.

Unfortunately, I have to sleep. 

James.  

[James]

I'm trying to make it give you a good, working wave output with as little fuss as possible.

If you want to mess with the settings, you can. But it should do a respectable job right out of the box.

James.  

[meandean]

  Gee, just when you had optimization down good, you had to go and frak it up! The XY now sounds like vinyl, with clicks and pops!  

[James]

Yes. I know.

I am writing a wave analyzer to look for the max step from one sample to the next in X and Y.

I'm also adding stuff to the frame set stats, so you know what section of samples you are looking at, from the beginning of time, in every frame. So I can find that crap!

If I can see it, I can probably figure out where it is coming from.

When it's all figured out, it will make a huge difference in the quality of the image.

James. 

[meandean]

I am writing a wave analyzer to look for the max step from one sample to the next in X and Y.

  Sacrilege! I thought that was one of the main reasons for using wave to begin with: you don't have to re-invent the wheel.

[James]

Huh?

I just wrote a function that rips through a wave looking at the difference between every two consecutive samples. So I get two numbers that are the largest step for X and Y and the sample numbers where they occur. If either of these numbers is a couple thousand or more, then the wave is probably not safe for scanners and it will have those ticks in it that you hear.

If the wave was made in LaserBoy and either of the numbers is greater than max lit or max blank vector length, then something went wrong in the process of making the wave and now I know where to look for it.

James.

[James]

Ha Ha!

I think I fixed it! 

(again) 

Anyway, I can make waves all day long that no longer exceed my max vector settings!

Now I'm going to actually test it on the scanners before I release it! 

James. 

[James]

Well I'll definitely sleep better tonight than I did last night.

It's really starting to look good.

I found the problem that was putting long jumps in the waves and got rid of that. Now I just have to tweak in the settings for distance span and work out the best formula for dwell.

It's all going in exactly the right places now, I'm sure of that!

James. 

[James]

Still working on it. It's unfolding into more details that I had anticipated.

It always does!

Oh yeah...... and my head exploded today.

Now, LaserBoy is busted in a totally new way !!! 

James. 

[James]

I did a considerable amount of nit-picking and testing and I'm pretty sure LaserBoy is good for another release.

Coming soon!

James. 

[James]

OK. I think this has been tested pretty well and found to produce favorable results and no crashes.

http://akrobiz.com/laserboy/code/LaserBoy_2009_11_12.zip

New things to look for:

In the wave utilities menu, you can choose 'h' or 'i' to go through a wave looking at the difference between every two consecutive samples' X and Y values for the maximum and the samples at which each occurs.

This is a very accurate way to tell if a wave has scanner dangerous transients in it. I'll probably add other wave data tests here in the future.

There are also some rearrangements in the frame stats that you see with '?' .

But the main thing is that dwell is now really accurate.

Things have been set up so that I can do further research and adjustments to it and make it even better.

Oh yeah! The included in.ild has a new frame, #17, that is quite usefull for adjusting and seeing dwell.

The obvious angles in the middle of the lines are 0, 10, 20, 30, etc... degrees. YOU can do the trig to figure out all the angles at the edges! Hint: all the long lines are the same length. You should flatten it before you make a wave. I made it 3D cause I thought it looked cool that way!

James. 

[meandean]

  X-Y sounds as clean as ever, but just a little different.

[James]

The difference is that now EVERY place where dwell should be added is getting its fair share.

Before, that was not true!

I haven't worked on that part of the code for literally years!

Since I have added a lot to my skills and to LaserBoy's ability to do forensic examinations of wave and other kinds of vector data, I decided to take a much closer look at what I really had.

I found there were a few situations where dwell was being missed completely; like at the ends of the blank line that must be added to join the last point of one frame back to the first point of itself or to the first point of the next frame, and in cases where the last lit vector in a frame is actually a dot (a vector with no magnitude or direction). That's all fixed now.

James.