Illustrations in Corian are carving with a strange pattern and little detail

[amanda0101]

I have tried everything! I got a request from an artist to transfer some of her work to corian, for some reason it just isn't working. There is a pattern in all the light areas like the background and the skin. The rest of it just looks very undetailed. I tried turning it into a picture first so I am not scanning on paper, getting rid of any pattern that would be there because it was painted, blurring the pixels, I am just out of ideas. Has anyone ever carved anything that wasn't a photograph? This is a new machine, had to send the other one back but I don't think that would have anything to do with it. Attached is an example, this isn't her work but after screwing hers up so much I thought maybe using a piece that was done by a different artist might work better, no such luck. Any ideas would be great as I keep asking her to wait for a finished product since I can't figure out what I am doing wrong or if it is even possible at this point.

Edit: I am saving the originals as PNGs not JPGs

[dougmsbbs]

Any chance we could see one of the PNG files? Then maybe we can find out whats going on...

[amanda0101]

Here you go. The only thing I can think of is that it was done on grainy paper making it show up like that. But it is all of my illustrations even my sketches so I am not 100% sure.

[ChrisAlb]

I have tried everything! I got a request from an artist to transfer some of her work to corian, for some reason it just isn't working. There is a pattern in all the light areas like the background and the skin. The rest of it just looks very undetailed. I tried turning it into a picture first so I am not scanning on paper, getting rid of any pattern that would be there because it was painted, blurring the pixels, I am just out of ideas. Has anyone ever carved anything that wasn't a photograph? This is a new machine, had to send the other one back but I don't think that would have anything to do with it. Attached is an example, this isn't her work but after screwing hers up so much I thought maybe using a piece that was done by a different artist might work better, no such luck. Any ideas would be great as I keep asking her to wait for a finished product since I can't figure out what I am doing wrong or if it is even possible at this point.

Edit: I am saving the originals as PNGs not JPGs

Hi Amanda,

If I'm understanding this correctly, You have both "scanned" the art work which is on paper, And taken a picture of the artwork and tried it both ways?

When you scanned it, what did you set the scan DPI at?

Chris

[amanda0101]

Actually I went to get it developed from the scan onto photo paper. I took a picture about 10 minutes ago and put it into the machine. The dpi on the one I attached is 72 but the one that I used was 300 another try at 400. Do you think taking a picture of the art would work? I have seen this quality before when I used JPGs instead of PNGs but I am not sure if it is for the same reason.

[dougmsbbs]

Did you resize the image you uploaded to a smaller size? I just took a look at it and it's only 256 colors. If thats what your sending to designer, it's not enough, it needs more to work with.

[ChrisAlb]

If you haven't yet, and you have a reasonably good camera say at least a 3.1 mega pixel. Try setting it up so there's no glare or uneven light on it and take a close up (Macro) shot of it. Maybe a couple. Bring the photo (jpg) in to your graphics software and convert to PNG just as is (no resizing or gray scale converting). Once it's a PNG then do what you would normally do to prep it. (I've seen your work, no way I could instruct you there...lol)

See if that yields any better results (at least on screen) in Designer.

Chris

[amanda0101]

The image is an 8 x 10 and I made it a 5 x 7 in photoshop. I didn't even think about the colors since I convert it to greyscale, so to be honest I was so busy figuring everything else out I didn't look.

[HighTechOkie]

Try zipping the photoshop file and attach it or the png that you are importing into Designer. Just not enough detail in the pic posted to determine what is going on with Designer.

Rob

[mtylerfl]

The image is an 8 x 10 and I made it a 5 x 7 in photoshop. I didn't even think about the colors since I convert it to greyscale, so to be honest I was so busy figuring everything else out I didn't look.

Hello Amanda,

The Designer software only "sees" 256 shades of grey. Starting with 256 shades should be fine for any photo - I don't see the need to have more then 256 "colors" at all - Designer can't recognize them anyway.

1) Scanning or photographing a textured paper is probably the primary reason for the "patterned" look that you're getting.

2) 72-dpi resolution is way too low (72 to 96 dpi is for computer display resolution only). I usually work at at least at 300 dpi resolution or higher in my photo editor.

3) I think the MAIN problem is the source material itself. Low contrast, overall darkness and possibly not enough variation in shades of grey in the drawing to yield an accurate representation as a carving

4) Playing with the histogram settings could possibly help - and/or adjust shadows, midtone and highlight settings in your photo editing software. You might need to do a lot of trial and error (edit, save as png, convert to pattern, put on the board, see how it looks...then try again, and again if it doesn't look right), so try not to get discouraged too early.

[dougmsbbs]

I would have to disagree. While it's true that Designer will only 'see' 256 shades of gray, the change comes in on the step before that, during the conversion to gray scale. There are several algorythmns that are used, all based on the way the human eye perceives the color densities of the red, green, and blue colors. The three are not treated equal. The more information it has to work with, the better will be the conversion.
While DPI has a larger part to play, this conversion can really make a difference. Going from 256 colors to grayscale produces a more uneven result than going from 16 millions colors will.
The first picture below was my starting point. The second was taken from 256 colors, and the third from 16 million. If you zoom in a bit on the last two you can see how the last is 'softer', without looking near as blocky. Sometimes you really can't see the difference, sometimes it's dramatic. I feel it's important to give too much information than not enough...

[mtylerfl]

I would have to disagree. While it's true that Designer will only 'see' 256 shades of gray, the change comes in on the step before that, during the conversion to gray scale. There are several algorythmns that are used, all based on the way the human eye perceives the color densities of the red, green, and blue colors. The three are not treated equal. The more information it has to work with, the better will be the conversion.
While DPI has a larger part to play, this conversion can really make a difference. Going from 256 colors to grayscale produces a more uneven result than going from 16 millions colors will.
The first picture below was my starting point. The second was taken from 256 colors, and the third from 16 million. If you zoom in a bit on the last two you can see how the last is 'softer', without looking near as blocky. Sometimes you really can't see the difference, sometimes it's dramatic. I feel it's important to give too much information than not enough...

Thanks for the tip and illustrations, dougmsbbs. Food for thought alright.

[Reset]

Here is another anomaly to consider when scanning, converting, and importing photographs.

http://en.wikipedia.org/wiki/Moir%C3%A9_pattern

Tommy

[dougmsbbs]

Reset,
Your right, that CAN play a part. But I don't find its really a problem just going to gray scale. Other types of conversions bring that up a lot more. Moiré patterns also show up when converting to dither patterns, and a lot of work has gone into preventing them. A dither pattern is used to convert a color image to a single bit of color depth, black or white, nothing in between. I'm working on producing dither patterns for use on the CW, to give us another way to produce images.
Here's one I did Friday, to give you an idea of what they look like. This one was done on Corian.

[deemon328]

I just wanted to add another thing to try. I'd try scanning the image at 600 DPI, then run a couple dust and scratches filters on it to see if it reduces the lines. I think it was paint shop pro that I liked for this filter choice.

[liquidguitars]

The Designer software only "sees" 256 shades of grey. Starting with 256 shades should be fine for any photo - I don't see the need to have more then 256 "colors" at all - Designer can't recognize them anyway.

mtylerfl, can you show us a test to substantiate this? :confused::confused:

LG

[mtylerfl]

mtylerfl, can you show us a test to substantiate this? :confused::confused:

LG

Hi LG,

See post # 11 above - dougmsbbs took the time to do a test with a color photo at both 256-colors and 16-million colors before converting to greyscale. From that particular photo, it does appear the 16-million color original made a smoother conversion to the 256-greyscale than did the reduced-color version before converting.

[ChrisAlb]

I would have to disagree. While it's true that Designer will only 'see' 256 shades of gray, the change comes in on the step before that, during the conversion to gray scale. There are several algorythmns that are used, all based on the way the human eye perceives the color densities of the red, green, and blue colors............I feel it's important to give too much information than not enough...

My take on DPI and color count.

Computers render images based on algorithms designed for our “eyes” to see. Up until now, that’s all they had to do. It only had to “look” good to the “eye” on the only output sources available, i.e. the screen or printer. Now the game has changed as we have an “additional” output source….the CW which, unlike the screen or printer, has a physical limitation to how “small” it can go. The limitation is the 1/16” size of the carving bit. At 1/16” (HUGE by comparison) to a pixel, it can only place 128 dots per inch of travel. Screens and printers are capable of producing thousands of dots per inch of travel thus “fooling” the eye into seeing a smooth, apparently transition-less change from one shade of (gray or color) to another.

Think of it this way. If you were to draw 128 dots in a one inch line. The CW (just touching the surface enough in depth to make a mark) can carve each dot separately and clearly. If you go above the 128, the dots will start to overlap and thus, the clarity begins to diminish. Adding more depth of cut increases the loss of clarity because the bit, obviously, is tapered.

Now while it’s true that (on screen or printer) a higher DPI and or Color count will produce a seemingly smoother “looking” result, you’re forgetting about the algorithm the CW then uses after that. It takes (and I don’t care how many dpi or colors you start with) all the pixels and translates them to the Nearest Matching Shade of Gray of which, as we all know, there are only 255.

Soooo…. You may have 1000 shades of red, in 1000 dpi that make what appears (to the eye) to be a pink line. But the CW has ONLY 255 shades of gray to work with so it will take all those (very close) shades of red and convert them to a shade, or shades of gray. If you have more than 255 colors, it simply takes ALL the varying shades around one of them and converts it to the closest shade of gray as that’s all it can do. So while you may “Think” you’re producing a “smoother” image (and to your EYE, on screen, you are), the CW simply doesn’t care because it CAN’T.

And I STILL wish pixels were ROUND...lol

Ok, enough babble, time for a second cup of coffee…LOL

Chris

[rjustice]

My take on DPI and color count.

Computers render images based on algorithms designed for our “eyes” to see. Up until now, that’s all they had to do. It only had to “look” good to the “eye” on the only output sources available, i.e. the screen or printer. Now the game has changed as we have an “additional” output source….the CW which, unlike the screen or printer, has a physical limitation to how “small” it can go. The limitation is the 1/16” size of the carving bit. At 1/16” (HUGE by comparison) to a pixel, it can only place 128 dots per inch of travel. Screens and printers are capable of producing thousands of dots per inch of travel thus “fooling” the eye into seeing a smooth, apparently transition-less change from one shade of (gray or color) to another.

Think of it this way. If you were to draw 128 dots in a one inch line. The CW (just touching the surface enough in depth to make a mark) can carve each dot separately and clearly. If you go above the 128, the dots will start to overlap and thus, the clarity begins to diminish. Adding more depth of cut increases the loss of clarity because the bit, obviously, is tapered.

Now while it’s true that (on screen or printer) a higher DPI and or Color count will produce a seemingly smoother “looking” result, you’re forgetting about the algorithm the CW then uses after that. It takes (and I don’t care how many dpi or colors you start with) all the pixels and translates them to the Nearest Matching Shade of Gray of which, as we all know, there are only 255.

Soooo…. You may have 1000 shades of red, in 1000 dpi that make what appears (to the eye) to be a pink line. But the CW has ONLY 255 shades of gray to work with so it will take all those (very close) shades of red and convert them to a shade, or shades of gray. If you have more than 255 colors, it simply takes ALL the varying shades around one of them and converts it to the closest shade of gray as that’s all it can do. So while you may “Think” you’re producing a “smoother” image (and to your EYE, on screen, you are), the CW simply doesn’t care because it CAN’T.

And I STILL wish pixels were ROUND...lol

Ok, enough babble, time for a second cup of coffee…LOL

Chris

Chris,
I have probably over simplified the "resolution" issue before by saying that the "Z" depth is limited to 1" of travel divided by the 256 shades of grey gives us a resolution step of .0039. I had the response below in another thread, and wondered if you could apply your knowledge/theory to the response. I would really like to understand this better... and it sounds like you have somewhat of a handle on it.

Jlitz... feel free to elaborate more

Thanks guys...

Ron


Designer limits the resolution of most operations to 1/128" when saving the project to the memory card, such as X-Y placement, depth, drill size, etc. The machine itself is capable of more.

Z resolution on raster carving is closer to 0.000488" (11bits / inch). However, since Designer only supports 8-bit depth maps the actual accuracy depends upon how deep the raster is carved and obviously the shape of the bit and type of material.

[liquidguitars]

Could be that the true test is to use a 8 bit Gif rendered image and then try a 32bit PNG rendered image and see.. The 8 bit information looks not so good and the 32 bit grayscale looks sweet. Designer reads 8,24,32 bit images you can make the choice.

I only like to use 32 bit PNGs on my projects if i can and the end result is what counts nice and smooth.
32 bit PNG file will give you a extra alpha channel if you need to remove the background 2.

LG

[ChrisAlb]

I guess it depends on what you're working with to begin with i.e. picture, photo etc. but, when I create patterns in PSP from scratch, or convert color images to gray, I only use 8 bit and they look great. By the way, there are only 255 shades of gray.

0 = nothing, no background. So when I want no background, or need to remove one, I just use 0. Works fine.

[Jeff_Birt]

Could be that the true test is to use a 8 bit Gif rendered image and then try a 32bit PNG rendered image and see..

LG, I think the difference there would be between the formats, not color depth. The GIF format like to dither pixels which is a really bad thing for our purposes. On the other hand, most programs that will save as PNGs will not dither unless you get fancy and try to make one palleted.

From working on CPBE I found it was easier to work with 32 bit PNGs but each color channel was set the same, yielding 255 shades of gray, with no worries about things going wonky if you used Designer to make a pattern from it. (Because I had NO idea how Designer did color to gray scale conversion.)

In the past I have also posted a formula that will solve for the carving depth of a given shade of gray based on your height and depth settings in designer.

[liquidguitars]

(Because I had NO idea how Designer did color to gray scale conversion.)

Designer needs a 24 bit rendered image for a nicer sub pixel defermation before converting into the project. Adaptive sampling occurs in Designer for overlays like Photoshop i think. This show the programs runs in a 24 bit space right? if not you could not have more than one 8 bit layer.

LG

[Jeff_Birt]

No, internally (currently) Designer is all 8 bit as far as patterns are concerned.

[mtylerfl]

No, internally (currently) Designer is all 8 bit as far as patterns are concerned.

Yep, that's what Chris and Joe Lovchik told me a few months ago when I was inquiring about supporting higher bit graphics
(Chris is the inventor of the machine and Joe is his brother - LHR's "graphics guy").

[hotpop]

OK! You guys got me throughly confused.

I use the paint.net program for my graphics. It has RGB colors and HSV (not sure what HSV is). I use the "V" scale to set my grayscale of 0 to 100 (black to white). It also has "Transparency - Alpha" which also has a range of 0 to 255. I haven't got a clue what to do with it.

Is each pixel a bit?

I've been resizing my patterns to around 500 pixels by whatever. I don't know if it this makes any difference either.

I save my finished work in PNG format because my erased areas remain empty when I bring the file into Designer. I don't understand why but that's what you guys on the forum said it should be.

I have Corel Paint Shop Pro X. Would there be any advantage for me to put in the effort and go through another learning curve.

[ChrisAlb]

OK! You guys got me throughly confused.

I use the paint.net program for my graphics. It has RGB colors and HSV (not sure what HSV is). I use the "V" scale to set my grayscale of 0 to 100 (black to white). It also has "Transparency - Alpha" which also has a range of 0 to 255. I haven't got a clue what to do with it.

Is each pixel a bit?

I've been resizing my patterns to around 500 pixels by whatever. I don't know if it this makes any difference either.

I save my finished work in PNG format because my erased areas remain empty when I bring the file into Designer. I don't understand why but that's what you guys on the forum said it should be.

I have Corel Paint Shop Pro X. Would there be any advantage for me to put in the effort and go through another learning curve.

Hey Mel,
No, a pixel is not a bit and for all intent purposes, you can forget about bits. Pixels are the little squares you see when you zoom way in. More DPI (dots per inch) will give you more pixels per inch so it can help in smoothing curves better. (Why I wish pixels were ROUND) ..lol

Resizing patterns by the pixel or the inch matters not. It just changes the patterns size and whatever size you're comfortable working with is fine. The only exception to this is...if you have a LARGE pattern with a lot of small details, when you shrink it on the board you'll loose detail. If you have a small pattern with a lot of details, it looks good small AND large. Most of my patterns I make so it fits on a 1x6 "as is". That way, folks can shrink it down or make it bigger without loosing detail.

I use Corel photo X2 for EVERYTHING. SO you've got my vote for that. Very fexible and capable but there IS a learning curve involved.

Hope that helps buddy

[hotpop]

Hey Chris,

Thanks for the comeback on my pixel concerns.

Just wondering if you have your computer in bed with you as you seems you are on the forum 24/7.:p

It's great that guys like you contribute so much time and knowledge to helping others on the forum.

Thanks again!:D

[ChrisAlb]

BED??? What be that??? If you're referring to this HARD wooden chair I sit in 24/7...then YES!! LOL :rolleyes:

And how come all you guys can get those cute little icons to show up in your messages and I can't??

Edit: Oh sure, NOW it worked....figures...lol

[amanda0101]

I appreciate all of your help! I will repost once I create my new lithopanes.

[liquidguitars]

Yep, that's what Chris and Joe Lovchik told me a few months ago when I was inquiring about supporting higher bit graphics

just try running 2ea 8 bit 256 color images in Photoshop at the same time
and you will see that is not the way it works.. could we get the software people from LHR to respond? :confused: :confused:

LG

[dcalvin4]

Amanda ,, I do vidio editing for fun and needed to come up with a vidio for a aniversary . I got the photos from the familly and scanned and imported them to the movie addded music and done.. Now I reviewed the dvd on the tv only to see black spots on most of the pics so now it was do over time for me..The hole thing in the trash..These black spots turned out to be specks of dust.This is getting to be more of a story so Ill cut to the chase [whatever that means] the best way that Ive found to reproduce a photo and have no specks and and no reflections from the front of the photo is to back lite the photo and take a camera picture of the photo, the lite shining thru the photo will inhance it and I think this will help you..Then dnload it as you always do..Denny