ILDA Format

[Pangolin]

Hi James,

I still don't have an answer to a few of my questions. Not that I am pushing for one... just wondering if we have really gotten to the end of this discussion.

My questions are:
* How does Format 3 really help the "palette" issue at all?
* A related question is: Does sending a file with Format 2 and Format 4 accomplish what it is that you think is accomplished by sending a file with Format 3 and Format 0?
* I also asked a direct question of LaserBoy -- does it support palettes with more than 256 colors?

I can see that, if your system is really wrapped around palettes, how Format 2 MIGHT help. But I really don't see how Format 3 ever helps in that regard (at least not completely).

Also, one thing that has been made clear -- at least in my mind -- is that if you have a software system that uses palettes, but most other software systems do not, then communicating frame data even when using Format 2 doesn't really help the end-goal much, if your end-goal is to maintain "the artists palette", if there was no artists palette to begin with. For example, lets say that my fully-shaded, fully-rendered picture mentioned earlier only had 256 colors, and it was sent to you using Format 2/0. Yes, there is a collection of colors there but not really a "palette" in the classic sense of the word. So yes, you would get the picture and colors, but wouldn't really be able to do anything valuable in terms of "palette rotations" on such a picture, would you?

Have we really come down to the point where we have identified a unique part of the LaserBoy software -- basically something unique whereby this software can do something that other software can not, because of its use of palettes? If so, then I guess that's the end of the story. You, and others can revel in the fact that your software has a unique ability, but also hopefully understand and appreciate how others can't (and in some cases don't want to) use such an ability, or have their own flavors of this ability (i.e. alternate ways of accoimplishing what would appear to be "palette rotations").

Is that really where we are at now? If so, then I will be happy, because I think it means we have full understanding by all parties, and no more arguing .

Bill

[keeperx]

Bill, attached is an image that should help answer the question about palettes over 256 colors.
you can see the palette on the right side of the window.

the small palette is the simple 62 color palette and the large one is a true color palette

[Pangolin]

Bill, attached is an image that should help answer the question about palettes over 256 colors.

Thanks for the contribution. I don't seem to be a jerk, but what you are showing here doesn't really answer any of my questions. My questions have mainly to do with programmers' perceptions of Format 3, and whether or not Format 3 really aids in the use of "palettes".

the small palette is the simple 62 color palette and the large one is a true color palette

Well, again, not wanting to seem like a jerk but... we have a bit of a terminology problem here. The term "true color" is mutually exclusive with the term "palette". Please see the Wikipedia page about it here:
http://en.wikipedia.org/wiki/Truecolor

Pangolin has been "True color" since 1992. We conciously made the decision to get away from palettes at that point in time.

We can make an abstract just like the one you showed here. Just as one quickie example I can lay my hands on, you can take a look at some abstracts from LAStudio here:
http://www.pangolin.com/LA_Studio/Abstraction.htm

None of these abstracts use a palette in the way you are showing, or in the way that LaserBoy uses them. In fact, some of these abstracts use a concept called "RGB Oscillators", which is actually something you can apply to images as well in some of our softare.

Color effects come in many shapes and sizes. They don't need to be manifested through the use of a "palette" like this.

Anyway, I hope we can get back on topic and hope that I can have an answer to my three questions posed above. Otherwise, I suppose we can consider this matter to be closed...

Bill

[James Lehman]

Oh boy!

I've got some writing to do!

I need to come up with a brief, tight TEXT explaination of palettes, color tables and LaserBoy!

I need coffee for this!

James.

[James Lehman]

OK. That took a LOT of coffee!

For clarification,

RGB is a 24 bit number made of (3) 8 bit numbers; a byte each for red, green and blue.

There are 16,777,216 unique combinations of RGB.

The maximum number of elements that can be stored in any of the ILDA format sections is 65,536.

I will use the word "palette" to describe a numerically ordered list of RGB values that number <=256. A palette can be (random access) indexed with a single byte.

I will refer to a "color table" as a similar, numerically ordered list of RGB values that number >256. A color table can be linearly indexed by its natural order (very similar to a palette, but you can not index it with a number that can be stored in one byte). The order of the color table is exactly the same as the order of the vertices that follow it in a format 0 or 1.

.....................

When LaserBoy first starts up and loads into memory, it loads a set of palettes. Each palette has a name. The first one is "Default_". This is the well known palette of 62 colors.

There are a total of 16 different palettes that might be of some use to the laser artist.

None of this has anything to do with the contents of an ILDA file..... yet.

When you open a .ild file, LaserBoy takes it in section by section. If it encounters a format 0 or 1 before anything else, it must assume that the "Default_" palette is to be used. This complies with the thousands of .ild files that have been out there all these years.

If LaserBoy encounters anything other than a format 0 or 1, then it takes a different action. If the next section is a format 2, then it first looks at its name (in its section header) and checks if it is already one of the pre-loaded palettes. If it is, it skips it. If it is not, it adds it to the end of the list of pre-loaded palettes. In either case, it flips the mode of the way LaserBoy will display any format 0 or 1 sections that follow it to now be shown with THIS palette.

frame, frame, frame, frame, palette, frame, frame, frame, frame...

In this stream of data, the first four frames (format 0 or 1) would be in the default palette of 62 colors and then, after the palette, the four frames that follow would be in the new palette.

One palette (format 2) has a one-to-many relationship with frames that follow it.

If you think about it, any single color laser system would require nothing more than a palette of intensity values from 0 to 255 to cover its entire range of possible levels. That is most definitely a palette!

The format 2 palette also makes it possible to import DXF data with the correct colors. DXF has a fixed palette of 256 colors that is not stored in the DXF file. One of the pre-loaded palettes in LaserBoy is "DXF_____". When LaserBoy reads a DXF file it knows to render it with the DXF_____ palette. If you wanted to keep the exact same colors you designed your DXF art with, and save that data as a .ild file, it would look something like this.

palette, frame, frame, frame, frame, frame, frame...

The general rule is, if LaserBoy encounters anything other than a format 0 or 1, it flips the mode of the way it will display any format 0 or 1 sections that follow it...

The possibilities for things that LaserBoy might encounter other than 0, 1 or 2 would include format 3. This is the color table. It is exactly the same structure as format 2, except that it's called a '3' in its section header and it has more than 256 RGB values.

If a format 3 is encountered, it is read and held in memory. Then the next section is read. This must be a format 0 or 1 and this frame must have the same number of points as there are (RGB) colors in the preceding format 3.These two sections are put together and the frame is rendered with the correct set of colors. There is no palette to add to the palette list or show in the current frame palette area and the name of the "current palette" is displayed as "-24-bit-".

Format 3 is known to have a one-to-one relationship with the frame.

color_table, frame, color_table, frame, color_table, frame, color_table, frame, color_table, frame.....

There is an interesting "fit" that happens when you put these two different ways of doing color together!

frame, frame, frame, frame, palette, frame, frame, frame, frame, palette, color_table, frame, color_table, frame, color_table, frame, color_table, frame, color_table, frame, palette, frame, frame, frame, frame, palette, frame, frame, frame, frame.....

In this example, the first four frames would be in the "Default_" palette, then a different palette takes over color and four more frames use that. Then something neat happens. A palette that was made to be a best match for all of the variations of colors that will be found in the next bunch of "-24-bit-" frames is loaded. Since these frames are format 0 or 1, they have a one byte index into the best match palette, but since each one of them is preceded by a format 3 color table, they are rendered by LaserBoy as "-24-bit-" frames.

If there was such a thing as an ILDA file reader that knew how to open format 2 and reliably skip anything other than 0, 1 or 2, then this would look like this:

frame, frame, frame, frame, palette, frame, frame, frame, frame, palette, skipped, frame, skipped, frame, skipped, frame, skipped, frame, skipped, frame, palette, frame, frame, frame, frame, palette, frame, frame, frame, frame.....

And the last occurrence of the palette, just before the section of "-24-bit-" frames, would serve as a best match for all of the "-24-bit-" frames, which actually have 8 bit palette indices!

Hopefully, this will clear some things up and raise some new questions.

As I mentioned in an earlier post, this leads to the possibility of yet another logical variation on this theme. It would also be of value to have a color table (which may or may not have more than 256 RGB values) that has the same color-to-point relationship as format 3, but also the one-to-many relationship of format 2 with any continuous series of format 0 or 1 frames that follow it.

But this would be completely outside of anything ILDA ever suggested.

James.

[James Lehman]

Bill, attached is an image that should help answer the question about palettes over 256 colors.
you can see the palette on the right side of the window.

the small palette is the simple 62 color palette and the large one is a true color palette

Nope. Those are both palettes of 256 RGB values or less.

There is no way (yet) in LaserBoy to see a color table laid out in order like a palette.

It is possible that it might contain as many as 65,536 RGB combinations.

A color table is an integral part of a frame, due to the one-to-one relationship between color tables and frames.

A color table only becomes a separate entity when saving the information to an (ahem.... ) ILDA file.

James.

[James Lehman]

Answer me these questions three, dare the other side yee see!

* How does Format 3 really help the "palette" issue at all?

Format 3 functions as a palette of sorts and it relies on, and enhances, the use of formats 0 & 1.

* A related question is: Does sending a file with Format 2 and Format 4 accomplish what it is that you think is accomplished by sending a file with Format 3 and Format 0?

Format 2 and format 4 are totally incompatible and should never be found together in any logical formation of an ILDA file. Format 2 is JUST a palette and format 4 is a set of points with colors.

* I also asked a direct question of LaserBoy -- does it support palettes with more than 256 colors?

In and out of ILDA (ish) files, yes!

James.

[James Lehman]

Another thing about palettes, in general...

There are basically two kinds of raster image files.

Photorealistic: millions of different colored pixels.

Logo: only a small set of colors.

Both of these image types have their place and not understanding when to use which will almost always make a mess of things.

Photos look the best as jpg files. Logos look best as gif or png.

As you might already know... jpg files are considered to be "true color" and gif and png are always "palette".

There is absolutely NOTHING outdated about the use of color palettes!

As long as we use video cards with RGB color, there NEVER will be!

James.

[James Lehman]

..... Otherwise, I suppose we can consider this matter to be closed...

Bill

You've got to be kidding me. Right?

James.

[Zoof]

Another thing about palettes, in general...

There are basically two kinds of raster image files.

Photorealistic: millions of different colored pixels.

Logo: only a small set of colors.

Both of these image types have their place and not understanding when to use which will almost always make a mess of things.

Photos look the best as jpg files. Logos look best as gif or png.

As you might already know... jpg files are considered to be "true color" and gif and png are always "palette".

There is absolutely NOTHING outdated about the use of color palettes!

As long as we use video cards with RGB color, there NEVER will be!

James.

OK, here are a few magic words of _raster_ computer graphics:

filesize

compression

lossless

I know how they apply to gif, jpg, tiff, png, etc

How do they apply to laser graphics?