Hello! Thank you for the kind words about our work, and for taking the time to read the documentation.
Regarding your question about rendering:
- 16x16 to 32x32: It is possible to use PixelAfterAll for an upscaling task like rendering 16x16 to 32x32. However, for a small scale-up like this, the result will mostly be a simple enlargement of the original pixels.
- Resolution Lowering: PixelAfterAll is generally more effective when decreasing resolution (e.g., 128x128 down to 16x16) as it excels at capturing the essence and style of the original high-resolution image and distilling it into a lower-resolution format.
- Manual Cleaning: Whether manual cleanup is needed depends entirely on the conversion result for your specific image. Even if some manual correction is required, PixelAfterAll should significantly streamline the process and get you closer to the final product.
For example, the dog in this X post (linked below) only required a small manual adjustment to the eyes after the conversion:
https://x.com/masuone_/status/1988547225929101804
I hope this clarifies things! Let me know if you have any other questions as you start experimenting with the tool.