ML Super Resolution image pixel size and dpi

Talk about Pixelmator Pro, share tips & tricks, tutorials, and other resources.
User avatar

2020-05-17 02:55:29

I wonder what is the purpose of increasing the pixel size and dpi following image transformation using ML Super Resolution? Can't it be achieved on the image with the same pixel size?
User avatar

2020-05-18 12:00:30

Hey there, I replied to your comment on the other thread, so I'll just paste that reply here:
Sometimes, though not usually. If the image is quite small, the amount of detail in it is constrained the number of pixels. For example, if there is a face represented by, say, 48 total pixels (8 x 6), you can't make the edges more natural without adding more pixels to the image, it's physically impossible. If the image is larger and the face is represented by, say, 192 pixels (16 x 12) but it's blocky or blurry because someone previously upsampled it using nearest neighbor/bilinear, then it would theoretically be possible to increase detail while using the same number of pixels, but that would require a different approach. In a case like that, the ML algorithm would probably need to be trained to recognize previously upsampled images and detect the resampling algorithm used in order to undo those changes and then, while keeping the same size, upsample the image again. Also, this isn't a very common problem — usually, the images are just very small, so the pixel size needs to be increased to physically be able to create natural edges and reintroduce detail.
User avatar

2020-05-18 14:14:27

Thank you for your reply. But, honestly, I didn't understand it fully. You mention small pixel size images and the need to increase pixels size of them. However, I tested images of different sizes, including relatively large ones (3000X4000) and they all became x3 in size after image restoration. Also, there is another issue, such as dpi. What is the purpose of increasing dpi?
User avatar

2020-05-18 15:05:42

In a raster image, the DPI setting is meaningless unless you'll be printing the image. When creating new images or resizing, it can have an effect but only because the pixel size is changed automatically if you specify a print size and a PPI (in which case, you don't specify a pixel size, so it has to be calculated automatically).

To put it another way, a raster image that is 1000x1000 pixels with a DPI of 1 is absolutely identical to the same raster image that is 1000x1000 pixels with a DPI of 1000. The only difference is the print size of the first image is 1000 inches by 1000 inches and the second is 1 inch by 1 inch. But on a screen and in terms of the detail they have, they would be identical.

For images that are 3000x4000 pixels, ML Super Resolution probably isn't especially useful unless you'll be printing a poster that will be viewed from a very close distance and you need the absolute best quality possible.