I bought some cubes to print an image on, it’s for my cousins new child to see and also to play with. However I can’t seem to get the laser to work right. I’m confident enough to play with the setting to make it lighter/darker etc but the issue is uneven burning, does anyone have any tips here?
The focus is right, I did a manual origin for the same reason others are having to do.
The cube was printed in greyscale mode at 10 dots/mm and 90% power.
Hi David; I suspect that the density may be too high for the image. With 90% power you may find it hard to avoid staining the wood with smoke or burning the edges of the image. The density setting may not help because it packs as many dots as it can into the space under the laser beam. It is not quite like digital photography (more pixels equal better resolution) and contrast is tough to set correctly. The image may need to appear far brighter in the software than you think is reasonable.
The smoothest of the dithering patterns appears to be JJN. I have tested the software extensively with a 1.6mW laser and tried to get reasonable printing results.I think you will find it is all trial and error and the way forward is to make as many tests as you can and run through the settings. I did this and only changed one setting at a time. My extended discussions about what the software is trying to do can be found here:
You may find something of use within that discussion but you will have to draw your own conclusions.
Can you share your original file? I am happy to test.
This is my test image which I used for the cube above, you may just about see the image within the burn. It’s part of a much larger image that I want to print over several cubes to give as a gift, however I wisely thought to get a test print right first so bought extra cubes (plus at least you get 6 tries with a test cube).
All suggestions welcome.
I’ll be trying a lower density option later on and will update on the progress.
With this image I had it set to
55-contrast
55-brightness
240-white clip
Greyscale
Atkinson
One hint, quality and contrast of the picture you want to engrave are the most important things for quality.
I think your picture is not that good you need it to be (for a perfect laser picture).
This image is rather low in contrast and does not appear to have very much in the way of dynamic range. That is to say that if one were to look at an image histogram representing shadow to highlight detail, half of the image would not be shown at the highlight end. The implication is that it does not have sufficient dynamic range for Luban to see and process it correctly.
Atkinson is a dithering pattern that can produce a lot of artifacting. In fact all dithering algorithms will produce artifacts but it is a matter of which of them are less objectionable. Floyd-Steinberg patterning is usually more acceptable than Atkinson. I will download your image and show you what Luban will do to it after giving it a more useful spread of highlight and shadow detail.
Some images for you…
This is your original image in an image editing program. The image was 275 x 275 pixels in size and was dowloaded at 72 dpi, which equates to a size of circa 3.8 inches. As I suspected and mentioned earlier, the image does not occupy the normal dynamic range. If you look at the graph next to the image, you will see no graph after the red colour. Extending to the right it is just black. The right hand side of the histogram denotes the highlight area and the left hand side denotes the shadow detail. Much of the highlight area is empty and this is all the light areas of the image, which for the sake of this discussion can be thought of as all of the areas above an 18% grey colour (mid grey).
Part of the issue is not giving the Luban software an image that it can deal with easily. The controls in the Luban software for contrast and white clip do very little. They cannot put detail into an image where there is no detail. In your image, the lighter parts of the image have been suppressed because the exposure has been shifted to the left (shadow area).
I corrected your image insofar as placing the histogram values in an expected normal range…
When you look at the enlarged image you will note that the red colour extends to the right hand edge (the highlight side) of the histogram and the graph also extends as far left (shadow area) as it can. Because this is a jpeg image, the detail is lacking and the headroom to adjust the image is impaired. This can be observed in the comb pattern of the histogram. Where the pixels have been adjusted (spread out more) the image begins to fall apart technically. However you will see that the image looks considerably brighter and more normal in tonal value.
The next image compares the two files without the image editor in the background. My tonally corrected image appears on the right hand side.
Now I want to look at what Luban makes of these two image files…
This above is your original file in Luban but set up for my SM1. There is almost no highlight detail in this image at all. It is likely that Luban cannot adjust this image sufficiently well to make the laser copy the image with any degree of veracity.
The image above has been expanded to occupy the Luban page and the rings denote the artifacting that is frequently seen with Atkinson dithering. The regular herringbone pattern is typical for images that are mainly shadow detail images. If the identical image had more highlight detail then the dithering artifacts will be different.
The processing of your original file in Luban shows that it is unlikely to yield satisfactory results for the laser etching.
I followed the same pathway with the image after I had corrected it tonally as well as I could despite it being a low resolution image. The following image was opened in Luban.
The artifacting was different but may have been more acceptable because it does not look quite as obvious as the first image.
The final image is the processed one which has more detail in the highlight areas than your original file.
Fine grained control of the image will only be available in a formal image editor. I have not tested it but I would probably aim to overexpose images for the laser. If you followed my earlier link to the discussion of the laser and how Luban processes images, you will see that I have yet to get any satisfactory results. I was pointed to a file and some settings but was unable to get the file correctly located in Luban. My requests to understand how to get the file (cat) located correctly so i could attempt to print it were met with silence.
I still do not know why Luban appears to be so complex to use. I suspect that the dithering algorithms and the way Luban determines what is the pattern to follow when image pixels are submitted to the laser aspect of the software are where it falls down. The act of getting a laser beam to replicate dark and light should be known and simple. I am afraid that Luban does not do it well. I have had good success with laser etching .svg files but no photographic images have been satisfactory.
Hope this is useful
Edit: I remember seeing a reference somewhere to there being a number that equated to a set focus point dimension for SM1 (presumably the same for SM2) where the laser beam would always be in focus when the laser beam aperture was adjusted to the smallest it could be. This is one of those things that Luban ought to be able to store as a home position for the ‘Z’ axis so the dance of jogging the laser beam to find focus is unnecessary. The required refinement is that the point of focus home position should automatically take into account the stock/workpiece thickness.
Nice detailed explanation of the process of prepping and the problems with getting results.
One trick I learned at some point (somewhere dealing with printing and halftones at some point in my life or maybe it was film school?) to quickly tell if a picture has sufficient contrast is to sort of squint at it with your eyelids barely open. If you try this with your first image, it all looks the same. If you try it with Jepho’s adjusted image, you see separation between the light and dark.
-S
Thanks S; that is a good point about squinting to assess highlights and shadows. When processing images (at some point in my life, for my sins, I processed graphics and text for pre-press purposes) I was taught that you will understand composition better if you turn the image upside down. I believe the theory was that when an image is unrecognisable to your brain, you are more easily able to assess its composition. I may be wrong but I think the technique may have been handed down by Jaques Lartigue or Henri Cartier Bresson. (Memory lapse 
) Nevertheless, it works really well.
Thanks @jepho, I’ll be going through this in detail and your step by step explanation will make this easier. I’ve got a few personal issues going on at the moment so sorry this reply was delayed.
No worries David; in your own time is fine.
