Beginning with v 5.2.0 we are offering three AI models to help you get the best possible quality.
Here is a brief overview of what each is intended to do.
This model is our original model and was previously called "natural." This model is likely the best choice for images that don't fit into either of the newer models.
This model was introduced prior to v 5.2.0 as "man made." This created a little confusion with users as to what "man made" meant. It is primarily intended for cityscapes, images of buildings, with special attention given to preserving clean straight lines.
This model is brand new beginning in v 5.2.0. We have trained it to remove compression artifacts and overall improve old and/or low quality images.
This model was added in v 5.3.0. It has been trained on non photographic images such as animation, graphic design, and paintings. (If you would like your image featured here as an example, let us know in a support ticket!).