Topic: Delayed Blanking

[James]

I don't think it has anything to do with the way you read them from the file. I mean, they are in natural order and they are independent data elements.

The issue is in how you interpret that data when you try to draw it.

Count from element 1 (not zero) and draw the line from element to element - 1.

I'm not sure I see any problem with the test pattern.

James. 

[James]

Being able to load a frame set into memory and rip through it looking for heuristics like "lit segments", corners, straight lines, etc... and control them as individual concepts; being able to rearrange the order for the shortest path, including the ability to reveres the stroke of vectors within a lit segment is all in the magic of LaserBoy! 

That's why you need to hire me! 

James. 

[dtewksbury]

The problem with the test pattern is that the blanking timing test shows the top line completely blanked, and the line just below it all on.

I am not sure I follow what you are saying about counting from element 1, do you mean I should ignore the first vertex all together?

Here are the test pattern and LaserBoy images without shifting the blanking and color.

[dtewksbury]

It does sound like I need to hire you.

[James]

Bear in mind the ILDA test pattern has all kinds of stuff in it that would not be found in any optimized frame.

There are lots of vertices that stroke image that are supposed to be blank.

You don't completely ignore the first vertex. You just go one past it and look back at it!

Draw your line from (index) to (index - 1). Your color and blanking are in (index).

James.  

[James]

It does sound like I need to hire you.

Well......... I'll make you sick to death of laser math and code! 

I do go on....   

It's more than an obsession at this point. 

James. 

PS. Neato looking graphics, by the way... 

[dtewksbury]

Thanks for the compliment.

mmmmm I just loaded the ILDA test pattern into LaserBoy and it does look the same as my previous image.

That being the case I don't quite understand how the blanking timing test can work. Unless you render to the laser differently that on the screen.

What I do, ignoring adding points due to traversal distance and point to point angle (which I calculate when outputing to the laser, BTW I disable this for test patterns). Is to read a point, and the color, and send that to the DACs, that's all. So looking at that ILDA test pattern, it would seem that it is incorrect (I am obviously wrong here). It would mean that the scanner would not be moving at the top line, and therefore the blanking position could not be set.

My brain is confused.
 

[drlava]

See my first post in this thread, regarding details on the ilda test frame.  Remember, it's designed for use with glavo blanking which is much slower than dpss blanking, so blanking shift compensation must be done during display (either in the playback software or preferably in the DAC controller) to 'equalize' the timing between the galvos and the lasers.

[James]

The ILDA test pattern is specifically designed to be a color laser projector obstacle course.

It is not meant to be wholly understood by any single human! 

This is why we must form communities! 

And have LEMs !!!   

James. 

[dtewksbury]

OK now we are getting somewhere.

I have attached an image of the LaserBoy ILDA as displayed by my small laser test rig (532nm DPSS).

This is with the auto traversal enabled because the image is point optimised.

It is a little hot on the joins, but getting there. I am using an opto coupled intensity control, so I have made any point that is brighter than 0.0 full intensity, I haven't finished the analog modulation yet.

[James]

I'd have to say that looks freakin' awesome!

It's interesting though that you can not tell that his pupils are pentagons.

Other than that, the way the shapes begin and end together is almost seamless and invisible.

NICE! 

James. 

[dtewksbury]

The pupils would be pentagons if I turn the minimum point distance down even lower, but the scan would get slower of course. There are probably no extra points being generated for the pupils here, I could turn the minimum distance down more and they would become pentagons. I also add extra dwell on the transitional points between blanked and non-blanked, that is why those joins are a bit hot. Plus I can't seem to tune the scanner amp identically in both axis, this is because the vertical mirror is longer that the horizontal one. There is a filter to compensate for this on one amp, but I notice that one of the damping controls is at full, so the symetry is a little off.

Thanks for the compliment.

It is interesting, now I can put points a long way away from each other and it draws perfectly straight lines between them, I have another one for you. I will post it in a minute.
 

[dtewksbury]

Hi James, I thought you might like to see your Avatar in Laser. I traced it in my program.

Here you go.

[James]

Dude!

That's awesome!

The thing on his shirt is supposed to be a letter Q.

That is Quisp, the cereal alien. He was designed by the same group that did the original Bullwinkle cartoons.

The cereal came from Quaker Oats, which is a company that started in Akron, Ohio.

James. 

[dtewksbury]

Ah, I have never seen him before.