site banner

Friday Fun Thread for May 19, 2023

Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.

3
Jump in the discussion.

No email address required.

I designed a proper cover for my ongoing web serial, Ex Nihilo, Nihil Supernum using a combination of AI image generation on-and-offline, Photoshop and Canva!

Certainly looks a lot better than I could have ever hoped to make in the past without hiring someone, especially since I peaked at drawing stick figures and isometric houses, Graphic Design Is My Passion, and I know just enough Photoshop to hang myself with haha.

What do you guys think? All I'm hoping for is to standout compared to the frankly amateurish art that most Royal Road writers commission, which isn't that high a bar to meet.

It's not bad, but could be better. Since you're asking for critique:

  • The title font seems like an odd choice. a quick search of sci-fi book covers shows a lot of examples of sans-serif, generally trending toward punctuated angularity. While one can always buck the trend, it's good to do so for a solid reason. The sort of psuedo-cursive you've got reminds me of a neon sign or something, which seems a weird choice for a hard sci-fi story.

  • The title positioning is likewise odd. tilted at an angle, cutting over the window frame, and the second line doesn't contrast well with the lower edge of the window frame. people do the boring stuff because it works.

  • Colors are not great. Overall, the colors come out as almost black and white, due to the floor, walls, rockets and the kid. The colors you do have are soft, sorta pastel, and a bit muddy. This combines with the title white and green to make the color scheme just sorta messy.

  • the lighting is wonky. the walls are brightly lit, the floor seems to be too, but the kid is very dark. The light he does get isn't very directional, so it's not even a silhouette or a rim-light, so it looks rather broken.

  • the composition is very stable, very static, all straight up-and-down lines, no real action or motion. the window frame and the way it's presented flat to the camera, the kid's posture, and the rockets all contribute to this. compare:

angle of the clouds, road, and the rockets in this shot

perspective on the smoke and clouds implying motion here

same here, with the smoke and buildings

...Are you generating the background, and then photoshopping the foreground? If so, there's some simple photoshop tricks that can get you better results. Any details you can give on your process?

Definitely going to give the story a try, looks pretty interesting!

The image is pretty much entirely AI generated, any PS work was limited to fixing some obvious congenital deformities in the silhouetted kid, and removing annoying watermarks.

Are those images AI generated themselves? I could do with a better prompt, this is one I thought up waaay back when SD was in alpha, and eventually found a variant that worked for the purposes I was going for.

As for the overall style, I wanted it to be idiosyncratic, something that stood out in a field of works that think throwing a spaceship in front of a nebula is good enough!

I appreciate the comments on the typography, at the moment I'm locked into the current version, but it's something to consider for the future!

Are those images AI generated themselves? I could do with a better prompt, this is one I thought up waaay back when SD was in alpha, and eventually found a variant that worked for the purposes I was going for.

I got the images by searching Pinterest for "rocket launch". Have you tried the generate-from-sketch options? Seems like it'd be easier to composite together an image, and then have the AI chew it over to unify the style... I really need to play with SD more.

I've dabbled with Controlnet and img2img, but it can be a PITA at times.

Bing has one benefit of having a better encoder, it understands subtle nuance in prompts and semantics better, though you tradeoff stylistic flexibility.

Cool story, I'll start reading it.

I've been struggling getting together some art for my still unborn webnovel. Turns out my stablediffusion loves giving robots huge pauldrons, even when that's specifically what I don't want, even when I prompt against it.

AI proofreading however, that's a godsend.

Are you running it locally? Are you familiar with in painting or more advanced techniques like controlnet?

Alternatively you can try Bing Image Creator that runs on the latest version of DALLE, it's quite good, and understands context much better than SD.

Well, you've given me the impulse to make it happen. I bought an expensive GPU for AI and I suppose I should get my money's worth.

I like how sometimes SD just throws up a real gem of its own accord without any manipulation but spontaneity isn't cutting it.

If you want very specific and controllable results, by all means use controlnet. You can inherit silhouettes and poses from reference images, such as a humanoid silhouette that you can make into a robot without the issue of pauldrons.

That, or deleting the offending section and inpainting over and over again till it's fixed.

While SD isn't the most advanced model by itself, it's the extensions that let it put up a fight against Midjourney and the like!

Got it working now. It's fascinating how much the open source community can do, yet the heavyweights can just overwhelm it with computational firepower. Google's high-parameter models could produce readable text on demand in photorealistic images before Stable Diffusion even existed.

https://twitter.com/JeffDean/status/1539743557476663300

We're also seeing advanced tooling start to come out.

Personally, there's so much else to work on that I'm waiting on Drag your GAN models to seriously get into the art side of things. (Those should be able to let you edit away those pesky pauldrons with ease)

Drag your GAN really is going viral. I saw it on Kruel's newsletter first, then 4chan (and not even /g/), now here. All in a single day. Guess that's the speed of progress these days.

I'm barely half into the first chapter, but man, I already fucking love this! I can sense that this story is going to scratch an ungodly number of literary itches for me...

Thank you! In fact, when I began the story, it was with the goal of creating the kind of fiction that I desperately craved but couldn't get anywhere.

Fuck it, I'll write my own book, with blackjack and hookers!

I know zero theory behind grafic design, but looks pretty cool. I'd definitely skim it if I saw it in a book store.

Glad to hear it! It just needs to catch eyeballs, I'm counting on the prose to keep 'em.