Resizing and rescaling in Photoshop

[Sarithus]

Hello Guild. Just a quick thread here asking how you deal with sizes in PS.

First off, I'm not always sure about what size I want to work at and how that size will affect the final image. I've got somewhat of a process for determining size, but it's a bit odd. I'd love to hear how you guys approach resolution and scaling.

1. I decide what size I want to final map to be. Let's say the final map will be 800x1000.
2. I open that document and then increase the size (with constrain proportions) by something like 200% since obviously working at a larger res allows you to add more detail.

3. Here's the part I don't imagine everyone does. I use the 50% zoom (or another %) as what the final image will look like. So I zoom in closer, add details and then zoom out to see how it will look at 50%.
4. When I've finished the map I'll save the image, open it and then resize the jpeg by 50% and use bilinear so it looks the same as the 50% zoom. (Bicubic automatic often results in a sharper image, bilinear looks almost identicle)

So basically I choose (depending on the size) either 25%, 33.33% or 50% zoom as the final size and work on that. It's a bit sketchy though as some zooms aren't a clear representation of what you're working on. (jaggy edges and so on)

Sometimes it just seems somewhat tricky determine how the scaled image will look.

Thanks for any responses.

[Azélor]

I base the size and scale according to the smallest elements on the map. It must look OK: we can see what it is/we can read without too much difficulty. It should still look acceptable when zoomed in but I realize this is something most people will not notice. Why would someone zoom at the trees at 200% ? Only to study the technique but at that size, it should look blurry/pixelated since most trees are only a couple of pixels wide.

[Gamerprinter]

Of course since I use vector instead of an image editor, I don't think in pixels per inch, since pixels (in vector) mean nothing until export. However, I think my perspective still might be helpful even to image editor users. I think about how many inches (vertically and horizontally) do I want the map, and is the map intended to be printed or used in a virtual tabletop application. If a map is intended for print, most large format printers are limited to 36 inches wide, though a 42 inch wide inkjet printer is common. Larger printers exist, but then it gets really expensive to print at larger than 42 inches. So once my inch scale exceeds normal printers, than it is only practical to use a virtual tabletop application for map display. This thinking largely applies to encounter scale map (1 inch = 5 foot scale).

When designing a map that is not intended for encounter scale, such as a city, regional, continental or world map where 1 inch might equal 10, 20, 100, 1000 feet, or miles per inches, then I limit to the largest normal and practical print size which might be 24 x 36 or 36 x 36 inches.

So if using Photoshop or other image editor, using my line of thinking, once you determine how a map will be displayed (print or VT-ready) and what is the maximum practical display size in total inches, then and only then what resolution you prefer to use becomes the issue. If you intend to print, 300 ppi is the goal, then of course, you need to calculate 300 times the maximum inches you need to display your map. If creating a map intended for virtual tabletop, which one becomes an issue as Roll20 (for example) allows you to display up to 72 x 72 inches at 72 ppi, whereas many other VT apps are limited to 24 x 24, or even VT apps for phones might be limited to 2048 x 2048 pixels total.

Just for clarification, the primary difference between image editors and vector is that once I complete my map design in a vector application, I can export the file to whatever resolution I want. The same map design can be exported at 72 ppi or 100 pppi (intended for VT-use) or 300 dpi intended for print. Using image editor software, you are forced to determine max resolution per inch at the beginning of your design, but knowing how you intend to display your map still matters in helping you determine scale and resolution.

[johnvanvliet]

for small maps i use a starting point of 4096 x 2048 pixels
most things are done using a image of 8192 x 4096 PX

time wise a 4096x2048 image is rather quick to edit
a 8192x4096 image will take 4 times longer ( 4 times the square area )

i normally "zoom" in
( this is JUST!!! how it is displayed on the screen -- NO RESIZING involved )
to about 200% to 1200% to work in just 3 to 9 to 16 pixels at a time

'll save the image, open it and then resize the jpeg by 50%

NOT a good idea !!!

the jpg format is what is called a "LOUSY " format
in that is tosses into the TRASH A LOT of the original information and ADDS!!! jpg compression artifacts
-- these are VERY BAD
and the original information is PERMANENTLY LOST


use the DEFAULT image format for YOUR program to save work

For the Gimp that is *.xcf
For Photoshop that is *.psd


if you MUST use jpg then only use it as the VERY LAST step

most editors will default to tossing out 60% of the image DATA/INFORMATION
-- this normally will not look all that good

use just 5% or 10% compression this will still make a small image( in MiB or KiB) but also REDUCE the artifacts

[Sarithus]

NOT a good idea !!!

I'm aware of Lossy compression. I'm not entirely sure what you think I do but I don't save over any psd's. I save the jpeg separately and then resize. As for jpeg or png, I personally don't see much different (if any) and I prefer a smaller file size.

I appreciate the replies but they weren't exactly what I was after. I think I just didn't explain it right.

[Midgardsormr]

I think the nitpick there is that you should save a lossless file, then rescale that and save as jpg rather than saving jpg twice. You'll get a slightly better result, particularly in areas with solid colors and/or sharp lines. My inclination would be to save a flattened psd, resize that and export to jpg, leaving the original layered file intact, of course.

Anyway, your process is fine, but you should be aware that Photoshop's zoom tool won't use the same filtering algorithm as your final resizing does, so your results will not be pixel-perfect. I'm not sure what filter PS uses in the zoom tool, actually, and a quick, lazy Google search didn't immediately turn anything up. A zoom of exactly 50% or 25% will give you the best results since you won't be dealing with as much sub-pixel sampling.

As regards my personal workflow, I prefer to work at my final desired resolution and view at 100%. I'll zoom out to get a look at the composition or color, and I'll zoom in to 200 - 400% for fine-detail paint work, but most of the time I prefer to see the image at the best final display quality.

[Azélor]

So, the whole question is about the zoom level used ?