My back still isn't, um, "back" to 100% so I can't get to the layout. But I can play around with Google Gemini (an AI program) to try and create some more backdrop images to fill in behind the foreground buildings at the north end of Middletown . . .
I started by asking Gemini for "fronts of factory and mill
buildings, viewed straight on" and got this image that included 4 images.
Not fantastic (when I rotated it, I lost a bit of the character of the varied roofline), but pretty darn good for less than 5 minutes' work.
I'll do the same thing with the other images and then see how they look. But HOLY COW! For someone like me that enjoys creating photo backdrops, Google Gemini is an amazing tool!
Just check out this before-and-after . . .
Here's the image I started with:
 |
| Point A |
And here's what I ended up with:
 |
| Point B |
Still a bit too "peak autumn" looking for my layout (which is set more in early Autumn), so I'll be doing some PhotoShop work to get it into final form, but from the B&W image to the image above took me no more than 7-10 minutes, if that.
Truly amazing!
I can't even imagine how long it would have taken me to get this result by regular PhotoShopping - even if I had all the skills needed. And this is certainly a long way from "just" colorizing a B&W photo (which I've been doing for over 5 years now - here's my original post on that, and the RMC article).
I actually screenshotted the nine separate instructions I used to get from Point A to Point B. So if anyone's interested in seeing those, let me know in the comments or contact me directly. It really is shockingly easy.
I'm not a huge fan of AI in most other contexts, but for making backdrops for your model railroad? I'm finding it pretty hard to beat. It's definitely a remarkable time to be in the hobby.
So give it a try - and if you do, I hope you'll let us know!
Very disappointing blog post. Even putting the legal and ethical issues of artist plagiarism and the unsustainable nature of the data centers used for generative AI, the reliance on a computer to bypass a fairly simple (and fulfulling!) Photoshop job comes off as uninspired and frankly devoid of creative spirit, looking more for results than the joy of the creative process that model railroading is all about. I apologize for my harshness, but as someone working in more than one field that is being actively sabotaged by the base principles of generative AI, I thoroughly reject the idea that it could produce any of the actual fulfillment this hobby can provide.
ReplyDelete>>>>> reliance on a computer to bypass a fairly simple (and fulfulling!) Photoshop job comes off as uninspired
ReplyDeleteI enjoy building models, not working in Photoshop, so from my point of view, letting AI do that work leaves me more time for creative model-making, not less.