I’ve been making photo-heavy collages in Krita and exporting them to Instagram at 1080 x 1080 pixel resolution.
Now I would like to try selling print-on-demand items on RedBubble. To upload a single all-encompassing file the recommended resolution is something like 7362 x 6480 pixels.
I’m a bit new with all this and I just need to know - if I rescale my files, will that give the right quality or would it be better to create them from scratch? If they look fine on the screen does that mean they’ll be fine to upload?
Sorry if this is a stupid Q. I’m struggling to find answers to it online.
No image software can create information from nothing. You can scale it up and it would formally satisfy their requirements but it would look blurry or pixelated. Imagine drawing on a balloon and then inflating it. Krita will try it’s best to fill the gaps but ultimately it can only guess based on what’s there.
Don’t get confused by what you see on your screen. Your screen doesn’t get more pixels so when you look at the full image it is zoomed out (basically resized to be smaller again). Imagine the balloon again but its now some distance away. But when you look at it in 100% zoom you will see the loss in quality.
Thank you, Takiro for cutting through the fog! This was what my logic was telling me but then I saw other people online saying advances in some tech means the software “fills in” what was missing. I wasn’t sure what the case was here with Krita.
I just read earlier that there are apps that can help you do this fill-in process but I’m feeling so paranoid about it I think I will redo them for my own peace of mind.
There are some smart algorithms that can do quite a bit but they have their limit’s. As soon as a a picture was scaled down, or when it was taken at a low resolution from the start, the information “between” the pixels is lost forever and can’t be recovered. When scaling up the algorithm looks what pixels are around the new ones and fills them based on them. However most algorithms have no awareness of what they actually scale (what’s on the photo) so they basically just fill the gaps with a mix of the surrounding pixels (it’s a bit more complicated in reality). Photoshop has some pretty smart algorithms that can even take the content into account, but I have my doubt even PS can handle an increase of over 600%.
In addition some pictures can be scaled up more easily, like pixel art where interpolation is actually not desired.
In case you don’t already know, Redbubble recommend that you produce/design your artwork using in CMYK format and then convert it to the sRGB profile before uploading for use in Redbubble. Notice that they don’t say which CMYK profile - sighs.
This is not a good thing to do in general but I can understand why they give this advice.
It’s to reduce the number of complaints from disgruntled and inexperienced artists who would be disappointed and annoyed that their printed artwork didn’t look the same as it did on their PC screen if they designed it in RGB.
Working in CMYK has many disadvantages but it may be just about ok for simple painting artwork. Some people here would tell you that working in CMYK is always a bad thing to do. However, there is a better way.
You should work in sRGB and use the softproofing facility View → Softproofing (Ctrl+Y) to see what your artwork would/may look like when printed and you can use the Out of Gamut Warnings (Ctrl+Shift+Y) to show you if the artwork colours will get ‘crunched’ when printed with a CMYK process.
Note: @Takiro may tell you that this is all dependent on your monitor being properly calibrated in the first place, which is hardly ever the case
I found this useful comment in their comment section.
I guess RB doesn’t make any RGB to CMYK conversion, most probably they deliver bitmap images to manufacturers directly, or they use some auto-conversion tool that would use different color profiles depending on the location of a buyer and selected manufacturer.
Anyway, here what I understood after 3 days of research, pls, correct me if I’m wrong. The end-result color depends on:
RGB to CMYK conversion, whoever or whatever is doing it. I tried several online conversion tools and they gave me different CMYK values for the same RGB color.
Manufacturer, as far as I know when a customer orders an item, RB choose the manufacturer that’s closer to the buyer location. Those manufacturers use different devices and inks, so the end-result may not be the same. Meaning, the customers from different parts of the world won’t get exactly the same product.
Medium, I read that both material (paper, plastic, textile) and its color may influence the end result.
Ink, as I understood the manufacturers may not use always the same ink.
Additionally, color profiles define a mapping between color subsets, they are device-dependent and there’s no world-wide standard. For paper printers they recommend FOGRA39 v2 in Europe/UK and GRACoL 2006 in US/Canada. So, using a particular color profile makes sense only if you sell your products:
to customers who live in one specific region
RB employs manufacturers in this region
most manufacturers in this region use the same “standard” color profile and you can actually get it (I only saw standard profiles for web and paper prints, but not for textiles, plastics, ceramics, etc.).
Conclusion: use CMYK to design (to get rid of the colors that cannot be printed no matter what), convert to RGB to upload (since both jpeg and png don’t support CMYK), and keep fingers crossed and hope that your colors will be reproduced accurately.
Hey, thanks, Ahab. It is all soooooo confusing to me and so I’ve been flailing about a bit and stressing that I can’t see how my designs are going to look bad. So this is helpful, thank you.
