Ximagic Quantizer

[eL_PuSHeR]

Good day.

I have found a bug in "GrayDither" plugin version 2.6.0.

If you press the OK button without UPDATING PREVIEW first, the process enters an infinite loop.

[brodriguez]

Thank you very much for the report.

I published a new version (2.6.1) with the bug solved.

[eL_PuSHeR]

Thank you very much!

[eL_PuSHeR]

Did you know this page?

Also, Wikipedia article about Dither

I am googling to see if someone has already made the 10x10 and 12x12 cell 45º Ulichney filters. 16x16 is somewhat coarse. I would also suggest you renaming difussion filters to Bayer, because the "diffusion" name is misleading. I am talking about your GrayDither plugin, which rocks, btw.

Best regards.

[eL_PuSHeR]

Good day.

I have found a nice way of crashing the GrayDither plugin: if you create several custom dithers (from image) and then try to delete them fast, GrayDither will surely crash.

[brodriguez]

Did you know this page?

Yes, I already know it. It is the only example I know of dithering with edge enhancement.

Also, Wikipedia article about Dither

This is one of the references in the GrayDither page.

I am googling to see if someone has already made the 10x10 and 12x12 cell 45º Ulichney filters. 16x16 is somewhat coarse.

Have you found it?

I would also suggest you renaming difussion filters to Bayer, because the "diffusion" name is misleading.

I agree diffussion is a bad name. I'll change the classification and names in the next version to reflect the type (dispersed/clustered) source and size/angle of each filter.

I have found a nice way of crashing the GrayDither plugin: if you create several custom dithers (from image) and then try to delete them fast, GrayDither will surely crash.

Thanks for the notice, I'll solve that in the next version. Most probably this bug is also in the Quantizer.

I expect to publish the next version this week.

Thanks for your help.

[eL_PuSHeR]

Thank you for your replies. I spent a few hours yesterday trying to draw a 10x10 dither cell and although I started with a 4x4 one, which seemed right, the 10x10 didn't. Is there any trick for drawing those cells? It can't be that complicated. the extremes for the cell gets overlapped and give a nasty banding effect.

[eL_PuSHeR]

Good evening.

Check this Google Books page I found while Googling.

[brodriguez]

I spent a few hours yesterday trying to draw a 10x10 dither cell and although I started with a 4x4 one, which seemed right, the 10x10 didn't. Is there any trick for drawing those cells? It can't be that complicated. the extremes for the cell gets overlapped and give a nasty banding effect.

Time ago I've tried to design a screen with poor results. From what I read there is a difficult task. I remember that some screens are slightly excentric to avoid that the points become lines. I suppose this is what you see as a banding effect. If the threshold values in the tangents between points are too high, the points will be connected.

Check this Google Books page I found while Googling.

Thanks for the information. I don't find the time to work on Stevenson-Arce until now, but I'll use this information when I find it.

[eL_PuSHeR]

Thanks. In the meantime, I will continue with smaller cell sizes, like 6x6. Come to think of it, how do you choose palette values?

If you design, let's say, a 6x6 cell (37 shades of gray), do you always start at the center with palette index 0 (black)? or is it better to start at palette index 256-36?, thus always ending at 255 (white). I will try that too. Back to trial and error method.

[eL_PuSHeR]

Hello.

Alms for the poor!

I have found an interesting page by a man called Gernot Hoffman -> Link

I had a document lying around that describes the creation of halftone screens and it even gives code for it. It resembles Turbo Pascal (which I know a little), but it says you cannot implement it using TP. Could you check it out?

Original link here
Alternative Document Download (< 700KB)

Regards.

[Drahken]

Personally I don't have much use for dithering functions, but I find your denoiser plugin extremely useful.
There's one annoyance about it though, scrolling the preview window. It'd really be nice if you could drag the image around instead of using the scrollbars, but I would at least like to have the image scroll in real time when you use the scrollbars. The way it is right now you have to take a guess at the right position, then let go & see if it's right, then scroll again & stop, repeat ad nauseum.

Other than that, the denoiser is a great plugin, even capable of making a decent looking image out of an over-compressed jpeg.


edit: One other thing, it'd really be good if you made it so that we can type values into the various fields instead of having to use the up/down buttons.

[eL_PuSHeR]

I have a question about the denoiser. Can it be used to demoire a scanned image? If so, what combination of filters should I use?.

Thank you very much for all your hard work.

[brodriguez]

If you design, let's say, a 6x6 cell (37 shades of gray), do you always start at the center with palette index 0 (black)? or is it better to start at palette index 256-36?, thus always ending at 255 (white). I will try that too. Back to trial and error method.

You should use 0-35. This range is scaled up to the 0-255 to cover the full gray range using the max value of the threshold matrix (supposing 8 bit gray).

I have found an interesting page by a man called Gernot Hoffman -> Link

I had a document lying around that describes the creation of halftone screens and it even gives code for it. It resembles Turbo Pascal (which I know a little), but it says you cannot implement it using TP. Could you check it out?
Regards.

For what I could read for the code, it calculates a clustered dot dither with 4 shapes (Dot, Square, line-x, line-y and circle) and provide rotated patterns rotating a image sized pattern.
The code could not be used, because there are several procedures not included XPowerA, Framelin, FMemGrYIQ, ImiToScr, RolTranP, ColToImi.
But if you only want to create a threshold matrix, you only need the procedure Dots and the code between { Build threshold array } and { Find actual image size in frame buffer } which apparently don't need anything more.

Personally I don't have much use for dithering functions, but I find your denoiser plugin extremely useful.

Since it is about the denoiser I reply you in another thread (I hope I don't break any rule )
http://newsgroup.xnview.com/viewtopic.php?t=17111

I have a question about the denoiser

Same as above, I'll Reply you in
http://newsgroup.xnview.com/viewtopic.php?t=17111

[eL_PuSHeR]

Using a separate thread for the Denoiser, seems like a good idea to me.