Aston Martin Mid-Engine

I have now achieved 50 posts and counting.

I started this project with a design direction that evolved into something entirely different. I had no brand in mind with the initial sketch or that it would be mid-engine. As you can see, I was imagining a more utilitarian concept which I’ll probably revisit in a future post. I put softer forms in these early sketches that transition into creases. The softer forms look sexier but I found it challenging to create a contemporary design with them. I ended up choosing more classic proportions than my original sketches by extending the deck length. I feel my final design direction is attractive but more derivative and safe compared to what I was striving for.

I found the Vizcom ‘palette’ or rendering style that yielded the best result was ‘Technicolor’. It produced the most realisticreflections. My rough sketches now had a solid look. I refined my design further by using a reference image in Vizcom to give it more punch. Note that using a reference image overrides the selection of a palette style. The reflections and backgrounds were greatly improved. My design wasn’t always interpreted as I had intended, but in some cases it was improved. Editing them with Photoshop is crucial to getting the results I desire. I combined two Vizcom renditions to get the featured image.

I decided to brand this post as an Aston Martin due to the classic proportions and the softer forms which are reminiscent of the marque. The layout and air intake imply a mid-engine layout.

I generated very satisfying videos with both ChatGPT Sora and Runway ML. I find that you can train the AI to get what you want by repeatedly generating versions with slight variations in the prompt. Each attempt seems to learn from the last one. The Runway videos are cleaner and less erratic than the Sora ones. The Sora videos with the car driving tended to cartoonishly shorten the wheelbase and enlarge the wheels. Runway produced more accurate proportions and a natural movement of the car while driving. My favorite video is the one with the car driving off into the woods. I would rather it drive on a road but this is how the AI responded to my prompt. At least it didn’t run into a tree!

The last image in this post was created by Photoshopping a still image from a Sora video. I like the dramatic lighting and view. I took this image back into Runway to create the last snippet of the video.

I played around with Topaz Labs Bloom for image upscaling to enhance and sharpen my renderings. They have a free trial for 10 images. I found that the highest creativity setting produced too much unwanted and frivolous detail. I couldn’t tell the difference in the lowest setting. As with all AI, there’s a learning curve.