ILDA Format

[James Lehman]

Yes, absolutely, because Format 2 required more work on the part of programmers, and most programmers in the laser biz are pretty lazy.

Dude. If I was any lazier, I'd have no pulse.

That's the first bullet point on my resume'.

James.

[keeperx]

whats the deal with eggs? are they good for you or bad for you... and what the hell is cholesterol? how can there be GOOD cholesterol and bad cholesterol? why do they call them both the cholesterol?

(i started this as a joke but after typing it I realized that it was glaringly ironic)

[Pangolin]

Hi James,

Thanks for the thoughts, and thanks also for controlling your temper. This is nice!

I have some questions for you, but first some comments.

I really don't know anything about LaserBoy and it could be that many others do not know about it as well. I guess from this discussion, it becomes apparent that LaserBoy uses Palettes in what we can call an "old-fashioned" way. The term "old-fashioned" is not meant to be a slam -- just an observation.

Back before Windows 95, lots of programs uses palettes. I can say the same about real computer graphics platforms like Silicon Graphics, etc. But certainly over the last 10 years (if not the last 20), computer graphics has gotten away from palettes. Computer graphics used to consist of single frames with "palette rotation" as a means of creating animation. Today, computer graphics is like fully-rendered 3D stuff like what you see on Nintendo. Basically, the nature of the goal has changed.

Getting away from palettes makes certain things possible, and more easy. But also, I will agree that having palettes makes certain other things possible that wouldn't be possible otherwise. There are certain computer-graphics-related things that the old Chroma Zone screen saver did that might not even be possible today. This may also be true of LaserBoy.

I can say the same thing about laser projectors in general. The old Lasarium and AVI and LFI projectors made their "imagery" by means of prisms, gratings, crushed up aluminum foil, etc. Today, most people have projectors that are much more simple -- 3 lasers and one set of scanners. The newer "3 laser" method makes a kind of visual much easier (3D line art in virtually any color), but the soft, gause-like textures of the past aren't even possible with newer projectors. So "old fashioned" projectors make certain visuals possible, just as the use of palettes make certain things possible.

But here's the situation James. LaserBoy is the only software I am aware of, that even uses palettes any more. There could be others, but I am not aware of them. Pangolin got away from Palettes in 1992 and my impression is that all laser software these days uses direct RGB specification of colors for point data. So, if I understand some of your arguments, you are wanting people do things in a way where you see clear benefit, but nobody else does (at least right now).

---------------------------------------

So I have a question for you about Format 3. Lets say I have a frame that has only four points. They communicate four X-Y-Z locations, and also four colors. Lets say the frame is like this:

XYZ XYZ XYZ XYZ
BLUE YELLOW RED GREEN


And lets say we used a Format 0/2 way of transmitting this to (or from) LaserBoy. And moreover, lets say that the palette is organized like this:

Index 1 = RED
Index 2 = YELLOW
Index 3 = GREEN
Index 4 = BLUE

So by me sending (or receiving) the data with Format 0/2, each program communicates not only the X-Y data, but also the "palette" insomuch as the order of colors that the artist created (which of course assumes the artist could create this in their software in the first place).

OK, so my question for you is -- how does Format 3 help in this regard? Unless I am mistaken, this is what you would get:

Format 3 packet:

BLUE, YELLOW, RED, GREEN

Format 0 packet:

XYZ, XYZ, XYZ, XYZ

So unless I really misunderstand Format 3, it seems to me that Format 3 (just as you would argue for Format 4) results in a "loss of data" as well, doesn't it?

My guess is that, unless there was a kind of "Super Format 0" which could communicate more than 255 color indexes, that would be the only way to solve the problem that you are pointing out insomuch as "data loss" and such, if you have more than 255 colors, right?

--------------------------------------------------

I also have a thought about this. If ILDA files were communicated with Format 4 AND! Format 2, then (as long as there were no colors used in the Format 4 part that weren't in the Format 2 part), that would solve the problem you are talking about, right?

Anyway, I guess at least we know where your passion is coming from. You use palettes, and I think nobody else does. So you are arguing for their use, but really nobody else is right now. I can see how this fits into Format 2, but I can't see how this fits into Format 3 right now. Somehow I just can't see that.

Bill

[Zoof]

Just an observation:

If you need compatibility with palettes _and_ RGB color definition, the format 3 is the only ilda file format option to have it. Format 4 and 5 don't have the palette indexes, format 0 and 1 don't have RGB definitions. Format 3 had both.

[Pangolin]

Just an observation:

If you need compatibility with palettes _and_ RGB color definition, the format 3 is the only ilda file format option to have it. Format 4 and 5 don't have the palette indexes, format 0 and 1 don't have RGB definitions. Format 3 had both.

Hi Zoof,

As far as I know (and gosh, I hope I would know...) that's not how Format 3 really works. With Format 3, the very first RGB you receive is the RGB for the very first point in the frame. RGBs from that point forward follow the color of each point in the frame. Format 3 doesn't have indexes, and wasn't really meant to work with indexes.

In fact, if a frame was sent with both Format 3 and Format 0 portions, the indexes of the Format 0 portions weren't even guaranteed to be valid or interesting (it could have been simple "index 0" and "index 1" to indicate "blanked" and "non-blanked" respectively). If we had ever supported Format 3 from LD2000, that's how we would have done it.

Remember, if "indexes" were anywhere, they would be within the Format 0 or Format 1 part of the file. But these only give you up to 255 different colors (not including black).

So it could be that there is a misunderstanding somewhere -- either with me, or with you and/or James' interpretation of what Format 3 is supposed to do. Format 3 really isn't supposed to have anything to do with indexes at all per se. Format 3 only gives you the complete RGB of each point in the Format 0 or Format 1 part.

Also, there are a few other problems. None of this has ever been documented very well. It could be that people have "read stuff into" the document that was never there to begin with.

Anyway, I would like to hear from James, about what benefits he feels that Format 3 really had, and how he believes it was supposed to work -- how, specifically, indexes fit into it, and how it gave you palettes with more than 256 colors. This is not a trick question -- I just want to understand where James is coming from. Until recently, I didn't know that LaserBoy relied on palettes so heavily.

Bill

[Zoof]

Hi Bill,

I understand that it was not the intention of format 3 definition at the time but it is how I used it briefly with LDS 2007 and it worked. It allowed to define ilda files with RGB information in my own software and use the same ilda file with the palette oriented LDS.
I made the index values in the format 0/1 part point to the proper color in the externally defined palette while the format 3 part had the full rgb definition. It worked but the palette handling was always as hassle.

Matthijs

[James Lehman]

WOW! My head is spinning.

Now I see that there is a huge gap between what we all "understand" about formats 2 & 3.

Format 3 should never come into play until the number of colors in one frame of format 0 or 1 exceeds 255.

If that happens, it is perfectly legal to first save a format 2 palette (<256 colors) that is a best color match for all of the >256 format 0 and 1 frames that follow. All of these frames would have a color index that points into this format 2 palette. Each of these format 0 or 1 frames would be immediately preceded by a format 3 color table.

Please read this (written almost 4 years ago):
http://www.akrobiz.com/laserboy/ilda_file_format.html

BTW, format 0 & 1 have a byte for blanking and byte for the color index. They are not the same thing. This is what makes it work with no defined palette at all. With only format 0 or 1 information, at least you know when to turn the laser on and off.

Here is the big point I'm trying to make...

Individual laser frames (at least in tradition) are not likely to have more than a couple thousand points at most. If each one of them has a unique RGB value, that is still a small set of RGB values and it IS a palette of sorts. It just has more than 256 values. It doesn't really matter if you save the RGB values with the XYZ values. It's still a small set of RGB values.

The fact that there are 16.8 million combinations of RGB bytes was true even back in the day of 256KB palette video cards!

A palette is not a restriction. It's order and control.

Ask yourself this. If all you had was the ILDA file format, would you be satisfied?

Don't you think the rest of the world deserves a decent format?

James.

[James Lehman]

......

James.

[Pangolin]

Now I see that there is a huge gap between what we all "understand" about formats 2 & 3.

Yes, it certainly seems that this is the case...

Format 3 should never come into play until the number of colors in one frame of format 0 or 1 exceeds 255.

Well, maybe, or maybe not. If someone sends a single Format 4 file, then the recipient has everything they need to reconstruct THE FRAME. This would result in a smaller file being sent, than would be the case if Format 0/2 or Format 0/3 had been used.

And the purpose for the ILDA file format is frame interchange -- nothing more. If something more gets accomplished, then it is a benefit, not an intended consequence.

If that happens, it is perfectly legal to first save a format 2 palette (<256 colors) that is a best color match for all of the >256 format 0 and 1 frames that follow. All of these frames would have a color index that points into this format 2 palette. Each of these format 0 or 1 frames would be immediately preceded by a format 3 color table.

Could be but... now you are sending a lot of data, most of which would not even be needed for, what I would argue is most software systems out there.

Also, there is one other thing that is "legal" and I would argue useful in some situations. It is perfectly legal to have an ILDA file with only a Format 2 section in it. That way, a software system could communicate the palette to another software system. The "interleaved" concept really wasn't an intention actually. It was done in 2002 for the Laser Jam. That's the first time I have ever seen such a thing done... The ILDA format docs mention nothing about this possibility...

BTW, format 0 & 1 have a byte for blanking and byte for the color index. They are not the same thing.

Yes, I know that. You have to remember the origin of all of this. All of these ideas came from LGRASS (Laser Graphics Symbiosis System) created by Steve Heminover. It used a Brooktree Video Dac with 64-byte palette inside of it. It also used a separate bit (blanking bit) because many companies actually used a separate blanking scanner back then, in addition to the color DACs.

BUT!! As far as I know, the ILDA format requires that "color 0" be black as well. At least this is the case in the latest spec.

This is what makes it work with no defined palette at all. With only format 0 or 1 information, at least you know when to turn the laser on and off.

On some level I agree, but I have seen so many mis-implementations of the ILDA file format, that I would bet if you saved a file with all "color index 1" and merely manipulated the "blanking bit", it would break many ILDA readers out there...

Individual laser frames (at least in tradition) are not likely to have more than a couple thousand points at most. If each one of them has a unique RGB value, that is still a small set of RGB values and it IS a palette of sorts. It just has more than 256 values. I doesn't really matter if you save the RGB values with the XYZ values. It's still a small set of RGB values.

On one level I can agree. BUT, if we think about a palette in the classic sense of the word -- some fixed set of colors that an artist uses over and over again to create artwork -- like the old timey paint palettes that an artist holds in one hand, while painting with the other, then your statement above breaks down, because, with modern imagery it is probable that the mixture of colors would change in every single laser frame.

Take another look at the picture that you see in this post:
http://www.photonlexicon.com/forums/...gner#post64756


That picture probably has a few thousand separate RGB values. In the next frame, because the lighting, shadows, and even lets say there is a color-cycling texture -- all of those colors could change on just the next frame, and the next frame, and so forth. These are "changing colors" -- its really not a palette in the classic sense of the word.

A palette is not a restriction. It's order and control.

I would say it can be both. On your system, it appears it is indeed a way of bringing order to things, and controlling things. It may also bring other benefits (i.e. animation by way of palette rotation). But for many people, it is indeed a restriction. Again, referring to the picture above, there is no way to do that with a "palette" in the classic sense of the word. There is also no way to do raster imagery with a palette.

I think this is why classic computer graphics has really gotten away from palettes. At one point in time, computer graphics WAS a single frame with palette rotations. Now computer graphics is fully ray-traced, fully rendered, real as life imagery. Basically, the nature of computer graphics has changed.

This doesn't mean that a single frame with palette rotations can't be beautiful of their own rite. Just that you don't find this any more as a feature built into 3D Studio MAX, Silicon Graphics, or even Adobe Premiere...

Ask yourself this. If all you had was the ILDA file format, would you be satisfied?

Hehe. No way! We were NEVER satisfied with the ILDA format... There are no inherent safety features in the ILDA format. There are no cataloging features in the ILDA format. There are no timing-related features in the ILDA format. There are no security features in the ILDA format. There are, essentially no animation-related features in the ILDA format. There is no way that the ILDA format conveys how the data is to be interpretted (as splines, lines, points, etc.). There is no face data or normal-data in the ILDA format, etc. Basically, there is more that the ILDA format lacks, than what it has to offer!

Bill

[keeperx]

i think the reason that palettes makes sens is that laser art is just that ART.

I agree that palettes may be antiquated when it comes to modern display technologies, they are certainly still the standard when it comes to ART.

i use a few vector/raster drawing programs (i like the Corel products over Adobe), and they rely heavily on palettes. And you can store the palette info with the art so its there when you come back to it..

If im creating a drawing, and i want it to have a certain "mood" or "look", one major way i control that is with the palette i use.

I can also send a file to someone to be used, reviewed, edited.. and they can see the palette thats being used even if i haven't drawn a single dot in the file.

I may recieve a palette from someone because they want me to use a very specific set or colors.

I may need a palette because i only have certain colors im aloud to work with, so i have to use a palette that is a gradients of those colors.

I always start a piece of computer art palette up, and many times that palette is one that is going to be shared with other proplr working on the same project.

so if i want to send joe the plumber the proper palette for our project, i would have to create a frame that has a POINT for EVERY color in my palette (since the RGB is stored with the point data). Its an ART document not a video file.

Storing a palette to be recalled by another person in the same software is a prety standard practice in the art world.

so technicaly speakin, the RGB method is great for displaying the data, its not all that great from the artists prospective.

Laser art is art.. the standard filetype that we trade art in should be one that keeps the artists needs in mind.

if there are 2 formats that that can both work equally well at storing data, and one of those formats offers the artist more freedom flexibility control options, then i would think the choice would be clear.

Ask yourself this question.. is this file format designed for storing data that goes to the scanners, or is it a file format for CREATING STORING and RECALLING frames of laser ART???...