ZBrushCentral

Getting resolution right for t-shirt printing

ZBrush Usage Questions
#1

Hi, I have a question about creating a model that will be used on a t-shirt print.
I have read the documentation on Zbrush and also was advised by the nice chap at pixalogic that I should resize my canvas as per the following:

3780 width
5906 length

This should enable me to print at 300dpi on a print area of 320mm width x 520mm length.

I don’t need to render (though I have tried in an attempt to better the quality) and I use plugin called Ryan’s Tools, which has a nifty antialiasing button that I use. However…

When I use ZApp to take the model into Affinity Designer (alternate to Photoshop) well, it doesn’t look that good (I also change the document in Affinity to 300dpi). The image looks pixalated and just not that crisp.

I’m at a bit of a loss on how to improve the quality and can’t find anything on Youtube etc, though I get that most people are not using Zbrush in this way.

If anyone knows, I would be most grateful to hear from you :smile:

#2

Hello @SharleenC ,

First off, check with the printer on what their needs are for the file. I’d be surprised if something being printed on fabric really needed to be that high res. It’s possible they may require that kind of res for prepress/ color separation work.

Secondly, I’m not certain what you’re describing at every point, but this only applies to image dimensions. It does not apply to an active 3d model. When ZBrush exports a document at a specified image size, it is exporting the entire canvas image. If the model only occupies a small portion of that canvas space, then you will have to zoom in to see it in your image editor, and this may result in pixelation.

So in ZBrush, if you have an extremely large canvas that can’t fit on the screen, you will have to zoom the document out with the Document > Zoom function until it fits on the screen. This is different than scaling a 3d model. This operation only affects the canvas. When the entire canvas fits on the screen, you can see the edge of the exported image. Any pixels the model occupies should fill most of that space, otherwise it will be a tiny part of the eventual image that you will have to zoom in to see.

Images are best viewed at 100% of their size to accurately see the quality of the edges in the image. Images viewed at odd percentages of their full size may look jagged as a result of displaying at a non-standard size.

If the pixelation is apparent in the image when viewed at 100%, you may need to increase the amount of anti-aliasing you are rendering with. The canvas in ZBrush is not anti-aliased unless the 50% AA viewing size is selected in the document palette. Otherwise, the image only has as much anti-aliasing as it was rendered with.

Re: image resolution. Image resolution settings in image editing programs are different ways of describing how large the image is in terms of pixel dimensions. “DPI” can actually mean different things depending on the context, but in digital image editing is often used interchangeably with “PPI” or pixels per inch. This basically means how many pixels the image will have in one direction per x number of inches set for the image output. So an image output at 10 inches wide at 300ppi, it would need to be 3000 pixels wide.

However, you are working in MM. Photoshop has options for Pixels/Inch and Pixels/Centimeter. You may wish to calculate using the latter, but again check with your printer for their preferred output.

A 32 cm x 52 cm image at 300 pixels/cm would be a huge 9600 x 15600 image. This would be an extremely large image, and larger than ZBrush is capable of outputting. I suspect your printer does not actually need an image of this size to print on a low res surface like Tshirt fabric.

Good luck, and I hope I haven’t confused you more!

#3

Hi Syndel

Thanks so much for the response, I really appreciate it.

Perhaps I could just double check what you’ve said as I’m not sure I have understood it all :blush:

I have checked with my printer and they do need 300ppi (think I said dpi) as a png.

When I go to ‘document’ and change the width and height, I’m just changing the size of the canvas – is that right? I didn’t realise that if so.

When I do this and I view the model at 100% (‘actual’ button), yes it does have jagged edges. Are you saying that I must render it in order to get rid of this? I don’t actually want to, as I prefer the model without the shadows created in rendering – do I have to?

If I’ve worked it out correctly (as I’m working in mm at 300ppi) then the size should be

(width) 320mm x 300(ppi) = 96000
(length) 500mm x 300(ppi) = 150000

(this is what you said but I just calculated in mm). So should I in theory set my canvas (‘document’ then set width & height) to this? I appreciate you said that Zbrush can’t handle this which may be why it crashes when I try it! Plus it looks like more than the maximum will allow anyway (8192 x 8192?).

If this is right, when I set it to the maximum allowed (‘document’ – width 8192, length 8192) and then use Zapp to go directly into Affinity (PS alternative) it does look small on the canvas in there. When I zoom it, it looks pretty pixalated.

I’m left still not being sure how I fix it? The printer definitely requires 320mm x 500mm at 300ppi (png). When I set the canvas size in Zbrush to the max, use the AA half button to view at 50% and use ZApp it still doesn’t look good! What to do?! :face_with_raised_eyebrow:

Many thanks

#4

That’s fine! It just makes understanding what’s going on a little more difficult, because then you’d have to convert mm to inches first to do the calculation. Using 300 pixels per inch in a file set to MM dimensions is otherwise perfectly fine if this is what your printer wants. I was simply trying to illustrate what was happening, and that is easier to do if the resolution is set to metric measure the same as your specified image dimensions (mm).

Yes, that’s correct. For the purpose of constructing an image in ZBrush, the document dimensions define the size of the canvas, and the exported image. For an especially large image like this, the canvas may not fit on the screen all at once, and you may actually only be working on a very small part of the overall image, like pasting a postage stamp in the middle of an A4 sheet of paper. This doesn’t matter if you’re only going to be working on the mesh in 3d edit mode, but does matter for the purpose of rendering and exporting an image.

Further information on this subject can be found here:

http://docs.pixologic.com/getting-started/basic-concepts/

You only need to render it if you want it to be pretty. :slight_smile: Not only does an actual render greatly improve the quality of the image in terms of shadow quality and detail, anti aliasing is the feature used to blend high contrasting edges together so they don’t exhibit stepping (jagged edges). This happens during the render process, and I’m afraid the render process may take a while for an image of this size, as it would in any program.

Now the caveat here is that if you’re working from a large enough image, the stepping on the edges may not be visible in the final print, as most printing processes themselves tend to soften the lines, especially when going to a rougher surface like fabric. Your printer may be able to advise you on whether you can get away without rendering. It may be the case that you can get away without AA, but your image would definitely look better if you rendered it. It would be safer to render it.

I’m afraid that I’ve confused you here. You were correct originally. A 320mm (32cm) x 500mm (50 cm) image at 300 Pixels Per Inch is approximately 3780 x 5906 pixels large.

If you meant that the image is actually 320mm width x 520mm as you specified in your original post , then that would be approximately 3780 x 6142.

:slightly_smiling_face:

#5

@SharleenC

One other thing I’ll mention that you might be running into. When a mesh is in edit mode and you change the size of the canvas, that mesh will be dropped to the canvas as pixols and look extremely pixelated at the larger size, like blowing up a smaller image too far in an image editor.

You would need to clear the canvas with Ctrl-N, then re-draw the mesh at the full size in the canvas and enter edit mode after re-sizing the canvas. To do this, you will want the full canvas on screen so you can see where the borders are. Remember also that performance burden increases in ZBrush both with the amount of active points, as well as the active pixels. You can expect this process to challenge your system a bit more, and you may want to draw your mesh out at the lowest available level of subdivision, and then switch to a higher level of subdivision after you have saved the file and are ready to render.

Remember to disable WSize in the document palette before saving the file, or your document dimensions will be altered when reloading the file.

Good luck!

#6

Hi again

Thanks once more for taking the time to explain things to me.

So I managed to make it work is doing this: (this is for the benefit of others who have gaps in their Zbrush knowledge like myself)

Set canvas size (‘document’ – 3780 x 5906 for a 320 x 500mm tshirt print area and then ‘resize’ (turn off pro)

Draw out model again and get in edit mode

Look to see how much space model takes up in white rectangular viewer (heck, I’ve no idea what that’s actually called) to the top left of your view port. Resize your model with gizmo until the appropriate size. (I think this was what you were trying to explain in your first email Spyndel but I missed this somehow)

Click on ‘anti-aliasing’ on Ryan’s Tools plugin. (this is not a plug, I just find it a useful button!)

Save / export as you need.

Whilst it isn’t perfect this has made a big difference – I only wished I got this before!!

Thanks again for your advice and patience :blush:

#7

Oh, and I also found this Zbrush video that helped, for anyone else who have been confused by resolution issues:

#AskZBrush: “How can I make a fairly large render in ZBrush?” - I guess this is why I couldn’t find it with previous searches as it is about rendering however the idea is the same whether you are rendering or not