Best practice for ML Super Resolution resizing an image?

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

2021-04-08 03:22:15

Hi,

Has anyone done any testing on resizing an image by about a factor of 8?

I mean an image going from 2000 pixels wide to 16000 or so.

Is it better to take multiple smaller steps? or just dive through in one big gulp?

Obviously, I want to minimize any degradation and if the time needed to do multiple passes is justified I will take that time.

I need to do about 100 images that I have in an archive from 2004-2005 so it's a lot.

Any tips/advice is greatly appreciated

Thanks,

M./
User avatar

2021-04-08 09:52:24

I also have a question regarding ML super resolution workflow, as in when is the right time to use super resolution.
If an image is directly imported to Pixelmator, would it be best to apply as the first step after importing a RAW file before any adjustments and denoise, or as the last step after adjustments and denoise has already been applied.

Then my second question is regarding a companion workflow with Lightroom, where the files are imported to Lightroom first, is it best to apply adjustments and denoise and then send as a TIFF to Pixelmator for super resolution. Or would it be better to first apply super resolution before any adjustment each time?

Lastly, does file format affect it in any way, such as a TIFF responding better than a DNG or camera raw files - or do all files respond equally well?

Thank you!
User avatar

2021-04-18 07:51:26

by Jason R. Comins 2021-04-08 08:52:24 I also have a question regarding ML super resolution workflow, as in when is the right time to use super resolution.
If an image is directly imported to Pixelmator, would it be best to apply as the first step after importing a RAW file before any adjustments and denoise, or as the last step after adjustments and denoise has already been applied.

Then my second question is regarding a companion workflow with Lightroom, where the files are imported to Lightroom first, is it best to apply adjustments and denoise and then send as a TIFF to Pixelmator for super resolution. Or would it be better to first apply super resolution before any adjustment each time?

Lastly, does file format affect it in any way, such as a TIFF responding better than a DNG or camera raw files - or do all files respond equally well?

Thank you!
I guess this is down to personal preferences.

I use the feature after I’ve completed my edits. Works for me.
User avatar

2021-04-20 08:40:55

Has anyone done any testing on resizing an image by about a factor of 8?

I mean an image going from 2000 pixels wide to 16000 or so.
I guess the best thing to do would be to apply ML Super Resolution to the image a couple of times. This should get you to 9x of the image size with minimum quality loss. There's also a dedicated action in the Automator app you can use to automate the upscaling process.
If an image is directly imported to Pixelmator, would it be best to apply as the first step after importing a RAW file before any adjustments and denoise, or as the last step after adjustments and denoise has already been applied.
ML Denoise is best applied before ML Super Resolution. It can, in theory, be applied after but it won't have that much of an effect. Besides, depending on an image, ML Super Resolution itself applies some degree of denoise when upscaling images. As for the Color Adjustments — they can be applied whenever it fits your workflow.
Lastly, does file format affect it in any way, such as a TIFF responding better than a DNG or camera raw files - or do all files respond equally well?
Both formats should produce more or less the same result in terms of quality. Of course, we'd love to hear how it performs in real-life workflows!