AI for conceptual aircraft liveries and spotting photos - Printable Version +- North East Buses (https://northeastbuses.co.uk) +-- Forum: Other Forms of Transport (https://northeastbuses.co.uk/forumdisplay.php?fid=11) +--- Forum: Aviation Discussion (https://northeastbuses.co.uk/forumdisplay.php?fid=13) +--- Thread: AI for conceptual aircraft liveries and spotting photos (/showthread.php?tid=4784)

AI for conceptual aircraft liveries and spotting photos - lier0021 - 23 Jun 2026 Following up on the recent discussions about using AI to tweak liveries and clean up spotting photos, I wanted to share another angle I’ve been experimenting with over the last few weeks: bringing our static archive shots to life. Most of us here have hard drives (or old shoeboxes!) full of static photography—whether that’s aircraft on short final, or heritage buses sitting at a local rally. While preserving the original historical record is obviously the top priority, I started wondering how modern generative tools handle adding actual motion to these older, frozen-in-time images. The concept essentially relies on an image to ai video workflow. You take a standard 2D photograph—say, a nicely framed shot of an A350 rotating off the runway, or a classic transit vehicle pulling out of a depot—and instruct the system to generate a few seconds of realistic motion based on that exact frame. I spent the weekend feeding some of my older panning shots into a few different platforms to see how well they understand transport mechanics. The results are a mixed bag, but genuinely fascinating. When it works, the system correctly identifies the subject, keeps the vehicle structurally stable, and animates the background blur, engine heat haze, or moving clouds. It creates a remarkably atmospheric short clip from a totally static file. When it fails, you get classic AI artifacts: landing gear that melts into the tarmac, or destination blinds and tail registrations that suddenly scramble into unreadable symbols as the subject "moves." For those who enjoy tinkering with their media libraries and want to experiment without committing to an expensive subscription, it's worth finding tools that offer generous trial credits. I’ve been testing SoraLum recently, as they offer a highly capable text to video ai free workspace that fits this exact use case. What makes it useful for transport photography is that it works primarily as an ai video generator from image . You simply upload your original spotting photo as the structural foundation, and then use text prompts to dictate the movement. For instance, you can upload a static ramp shot and prompt it with "slow camera push-in, ground crew walking in background, dynamic lighting." I've found that keeping the motion prompts incredibly subtle yields the most realistic results. Asking for massive movements usually breaks the vehicle's geometry, but asking for slight environment animation really brings the old photos to life. Has anyone else on the board gone down this rabbit hole yet? I'd love to know if you've managed to successfully animate any of your historical transit or aviation shots without the AI destroying the mechanical details, and what your workflow looks like.