About the Showoff Spot category

This category is for users to share their works with each other.

8 Likes

Basketball photography powered by Topaz Photo AI

More @ https://www.instagram.com/roy.kfir/

4 Likes

July 1 is Canada day and Ottawa usually has decent fireworks. So I picked a spot on the roof of a friend’s apt. building where you get a view of Parliament and the cathedral and the fireworks and shot it with a Z8 using a 120 mm F4 lens.

The basic footage was great. But of course the dark areas had a lot of noise predictibly. A problem with the Z8 in general actually.

After running it through Nyx, not the fast but the regular one, I got this.

I should add i shot it in N raw and graded it on DaVinci before it went into Topaz.

Watch in 4K

I have to say I’m pretty happy.

1 Like

I am so glad you decide to share! Great work!

1 Like

Nice shots! I shoot concerts and have only shot a handful of sports. Mostly hockey. Great job on them.

After

Before
Before Giga and after using the new redifine with prompt

1 Like

Is top Before & bottom After?

Love the face on the bottom one! Top has more defined details.

top after bottom before

1 Like

Just out of curiosity (b/c you produce cool images) is there a way to combine the details you got on the After result (minus the face) and somehow do just a face recovery on the bottom one (to sharpen it up a bit but retain its character w/out Redefine) then put those 2 results together. That bottom face has so much character vs being more of a (for lack of a better analogy
) pinup girl. Yes, I know she’s “Topaz Girl”
 :wink:

If we were able to mask or select an area in Gigapixel itself(I already put in a feature request on this) then yes
I could try a negative prompt about the face we will see


1 Like

That makes sense. Yes, it would be an important feature to do things like that.

That’s why I work with the Ps plugin. To improvise like that with compositing or blending images.

when you use a mask or selection in photoshop does the mask or selection show up in GigaPixel
?

I can send a masked portion of an image from a Ps layer into PAI or GAI plugin. But I won’t see the mask that isolated that portion in the Topaz apps.

When I save an output result from PAI or GAI UI’s back to my Ps layer I create the mask to hide or reveal whatever content I want there.

The mask created in either app does not flow back or forth to the other app (as the products are designed as plugins now).

Using your image with the more ‘evil’ moody face as an example (vs the sweeter fairytale Topaz Girl face)
 I could process the image as you did and save it back to my Ps layer. Then I could mask out (aka, hide) the new Topaz Girl face Redefine created and reveal the original face. What I’d do then is create a ‘stamp layer’ (essentially a merge of the prior layers and then send that image back into either PAI (just for sharpening or SuperFocusing or Face Recovery) or GAI (if upscale desired too) and I’d use the Custom masking in either of those to tell the app to only sharpen (or, Recover or SuperFocus) the original girl’s face in the image (so not oversharpening all the already Redefine detailed other content). With that done, I’d send/save back to Ps. Hope that makes sense.

ok thanks
Wish we had the mask or selection option in Giga


That would be great.

I’ve been using Gigapixel in my workflow to produce illustrated typefaces and related promotional marketing mock-ups. 30 typefaces just published as a font book vol. 1. PDF proof here: ART-TYPEmobile.pdf - Google Drive

2 Likes

How do you use it? Sounds interesting. Is it primarily for upscaling, or what? Do you design the full type designs in a different software?

I have several different workflows/ways of design process. I start with free AI pulling images—many many images. My next step is to scale them in Gigapixel. Sometimes those images become the basis of my type design and I work them in Photoshop and Illustrator (see Cicada). Sometimes the AI becomes components of the type design that I assemble and work in Adobe (see Color Blocks). Another way, is to use AI elements to comp my type design in Photoshop then draw it as vector art in Illustrator (see Script). Lately I am using Fontlab to assemble my individual letter art files into a typeface file I can use as a font. For big graphic pieces I build the type words in Adobe.

1 Like

That sounds cool! Thx for explaining your different workflows & where GAI fits in along with other apps.

I looked at the PDF link you posted. Really like the 3D quality of the different fonts you have there. Reminds me of the feature Adobe had in Web Firefly for a while that let us gen fonts based on prompts (like, “loaves of bread” or “fur” or whatever)
they may have moved that feature to Adobe Express. I forget. But it was fun to play with.


interesting how the model gave the raccoon human like eyes.

1 Like