I almost dismissed this as I was losing a tonne of detail in a drawing. It applies a lot of smoothing, even when using the photo model - too much smoothing tbh. But then I tried sharpen image and it's impressive. I ended up doing two versions, one with photo and the other with sharpen, went into photoshop and layered the sharpened version ontop of the photo version, and then applied some blending. This allowed the texture from the sharpened upscale to come through which was lost in the photo upscale. It would be nice if they allowed some settings to be tweaked, but an impressive packaged little app!
Amazing tutorials. I am just starting to discover SD features, joined one course, and now found your channel today. Question about the characters in this video. Did you create them using a trained model, or those are just randomly generated via the prompts?
Thank you. I do believe that was a standard model, I’d have to go back and check-been testing about 40 of them recently. So it was prompt engineering. There are some amazing things you can do with some hours of iterations and refining descriptions to tease out stunning visuals. I’ll be posting a video soon on how to do that and show off some of what I’ve done. I’m a poet and novelist, so editing words down to get the most complete visual that expresses my inner intent is part of my normal routine.
Thanks again for this.
I almost dismissed this as I was losing a tonne of detail in a drawing. It applies a lot of smoothing, even when using the photo model - too much smoothing tbh. But then I tried sharpen image and it's impressive. I ended up doing two versions, one with photo and the other with sharpen, went into photoshop and layered the sharpened version ontop of the photo version, and then applied some blending. This allowed the texture from the sharpened upscale to come through which was lost in the photo upscale. It would be nice if they allowed some settings to be tweaked, but an impressive packaged little app!
Amazing tutorials. I am just starting to discover SD features, joined one course, and now found your channel today. Question about the characters in this video. Did you create them using a trained model, or those are just randomly generated via the prompts?
Thank you. I do believe that was a standard model, I’d have to go back and check-been testing about 40 of them recently. So it was prompt engineering. There are some amazing things you can do with some hours of iterations and refining descriptions to tease out stunning visuals. I’ll be posting a video soon on how to do that and show off some of what I’ve done. I’m a poet and novelist, so editing words down to get the most complete visual that expresses my inner intent is part of my normal routine.