To be honest, I'm not greatly impressed by the quality of the enhancements, as they appear rather cartoon-like and have an enormous SJX/ActionGirls logo trying to hide the original ViewPornstars.com logo.
I though I'd give it a try to see if I could do any better enhancing this cracking set. I've deliberately not let the AI go over the top removing noise and blurring so that the outcome is more like a photo than a cartoon.
As you might expect, the enhancements from smaller original photos in the 08x and 09x range have mixed results, but they are all recognisably Silvia.
Anyway, enough rambling. Here are the photos:
set088_001
set088_002
set088_003
set088_004
set088_005
set088_006
set088_007
set088_008
set088_009
set088_010
set088_011
set088_012
set088_013
set088_014
set088_015
set088_016
set088_017
set088_018
set088_019
set088_020
set088_021
set088_022
set088_023
set088_024
set088_025
set088_026
set088_027
set088_028
set088_029
set088_030
set088_031
set088_032
set088_033
set088_034
set088_035
set088_036
set088_037
set088_038
set088_039
set088_040
set088_041
set088_042
set088_043
set088_044
set088_045
set088_046
set088_047
set088_048
set088_049
set088_050
set088_051
set088_052
set088_053
set088_054
set088_055
set088_056
set088_057
set088_058
set088_059
set088_060
set088_061
set088_062
set088_063
set088_064
set088_065
set088_066
set088_067
set088_068
set088_069
set088_070
set088_071
set088_072
set088_073
set088_074
set088_075
set088_076
set088_077
set088_078
set088_079
set088_080
set088_081
set088_082
set088_083
set088_084
set088_085
set088_086
set088_087
set088_088
set088_089
set088_090
set088_091
set088_092
set088_093
set088_094
set088_095
set088_096
Edit: updated with revised images 077-092,096.
-- Edited by spike on Sunday 5th of March 2023 02:16:26 PM
Amazing work Spike. Perfect from 1 to 76. But 77-96 can be enhanced a little bit. I have found a way for my part to upscale small pictures and here is my results with pic77
Yes, as I said, the smaller photos aren't great, so need a bit more work. They were lazily fed through Topaz with the same settings as the larger photos, but actually need more individual attention.
With 077 as the example, I think somewhere in between my version and yours would be a good place.
Yours has maybe gone a little to far on blurring out the detail, whereas mine has brought out too much detail.
They could also do with meeting in the middle in terms of colour balance. Mine's a bit yellow-biased wheras yours is red-biased.
077 was tricky because it came from the LegActionMag series which was darker than the ViewPornstars series, so we have both tried to adjust the colour balance and contrast and come up with different results.
I'll have another go at the smaller photos individually.
That's much better but i don't like at all your 082 and 086 : 086 because Silvia's face is too white and looks cadaveric, and 082 because there is a grey pattern on her body.
Perhaps you could work again on these 2 or play with my upscales :
Yes, I agree these two are a bit ghost-like, as is 088 too. I'll persevere with my own upscales, as yours also have some artifacts, e.g. around Silvia's garter belt in 086, although much nicer colour balance.
My next attempt at upscaling the LegAction.com photos (077-092)
I realised I had gone down a cul-de-sac with the colour balance of my previous efforts, so I went back to the originals and using a different method to try to get the colour balance more in keeping with the ViewPornstars set. I'm a lot happier with the results.
set088_077
set088_078
set088_079
set088_080
set088_081
set088_082
set088_083
set088_084
set088_085
set088_086
set088_087
set088_088
set088_089
set088_090
set088_091
set088_092
P.S. I have updated the original post with these versions.
-- Edited by spike on Sunday 5th of March 2023 02:14:51 PM
Respecto a la ampliación del 77 al 92, se nota bastante que la calidad de las imágenes originales no es suficientemente para que el resultado sea bueno con una resolución tan alta. Las imágenes quedan muy difuminadas, perdiendo demasiada definición.
Sin duda, debemos jugar mucho con el tamaño justo para evitar, sobre todo en la parte de los ojos acabe peridiendo toda la definición.
To be honest, Topaz can do better with the facial detail, but it can also overdo it, making Silvia look strange.
I deliberately set the facial reconstruction at a low value (25%) to keep Silvia's face recognisable, but perhaps I aimed too low.
Here are some tests from 077 to show what I mean. The percentages are the levels of facial reconstruction set in Topaz.
0%
50%
100%
All look pretty much identical in thumbnail form, but you can see obvious differences when you view them at full size.
0% lacks detail as you would expect.
50% is better. Silvia looks natural, but the detail still slightly fuzzy around the eyes.
100% has sharper detail around the eyes, but in my view has over-emphasised Silvia's teeth, making her look a little buck-toothed.
Both 50% and 100% versions have also blended out much of the tonal differences in Silvia's skin. It's that effect I was mainly trying to avoid, as it looks out of place, like a faked photo, which is why I set the target so low.
However, we can use the different levels of facial reconstruction selectively by overlaying the images and adjusting their alpha levels to blend them together.
Here is a composite of all three images done in this way.
The bulk of Silvia's skin is from the 0% image to keep the tonal range, her teeth from the 50% image to reduce the emphasis, and her hair, eyes and nose from the 100% image for maximum detail. These are all fairly subtle differences, but I think the result is an improvement on all three individual photos.
I'll have another go using this composite method, if you think it's worth it.
Or please tell me if I'm barking up the wrong tree...
Your last attempt with pics 077-92 is much better.
Concerning facial optimization, Topaz and Photoshop have the same problem : they focus on eyes or lips and rest of the face, as nose, hair, ears are blurred and you'll never have a complete perfect face.
Of course, you can try to divide the image (or face) in parts to have a better result, but you'll need hours to upscale a single picture. And you'll perhaps become crazy before the end.
We have so many 'easy' sets to upscale.
I suppose the AI models are in their infancy. Human facial recognition is very highly tuned, so tiny facial imperfections are magnified in our perception, whereas similar differences in other areas of the photos are almost imperceptible.
There are only a few small images in this set that require this manual treatment. I've come this far with it, so I think I'll persist with composites.
OK. Here are my final efforts with the 600px/650px photos, using a composite overlay method and a little bit of manual tweaking to correct obvious errors.
For completeness, here are all the upscaled photos, including the new ones numbered 077-096: