I recently came across a website called magnific.ai . I was extremely impressed by the demonstrations showcased on Youtube. The way it can extrapolate details and enhance the dynamic range is truly remarkable.
Considering the innovations being made in image enhancement, I would like to propose the integration of similar feature into Photo AI. In certain situations, prioritizing detail fidelity, imaginative enhancements, and an overall aesthetic appeal can often be more crucial than achieving a simple upscale and 1:1 replication of the input image.
I would appreciate it if you could consider this feature request, as I believe such a model has the potential to significantly enhance the user experience and broaden the creative possibilities within Photo AI.
yes, Topaz Photo AI does a good job of enhancing the quality of the images but sometimes there is not enough fine detail for the wow effect. With the latest updates I noticed that the program has learned to add grass, sand etc where needed, I suggest going further. We have a number of textures that are often found on images and at low resolution these textures disappear, in fact when you zoom in on images through Topaz such details are not restored. I’m talking about brick texture, rows of windows, fences and roof tiles (very often Topaz does not render such things well), fabric texture, buttons, locks, clasps, tree and foliage texture (it would be great to get not a solid green texture but also leaf veining), snow texture and so on.
We’re working on improving our Remove Noise models to not remove finer details in the textures. In the meantime, try turning down the strength for the filter (or even off depending on the amount of noise there is) and turn on Upscale. You can set the scale to 1x if you don’t want to enlarge the photo but it’ll enhance resolution and do minor noise reduction.
