project was my first attempt at vehicle modelling, the goal was
simply to create a realistic car inspired by the Aston Martin DB9,
keeping it proportional and including as much detail as possible. Due
to different constraints and issues it was decided that the interior
would be left out of the model.
I was happy with the attempt as a starting point for myself and feel it is proportionally accurate.
Another thing was the different techniques that I was able to pick up in creating the shape of the model. I feel I am a more competent artist and I have more confidence in my own abilities after completing this project.
I was reasonably happy with the outcome of the project in terms of
proportions and what I had learned, there was still a lot more to
learn and a lot to improve on.
I feel the main issue was a lack of references, I may have been a little too keen to get going on this project and just downloaded the blueprints, loaded them as image planes and just got cracking. That's not to say that no references were ever looked at, just that not enough were taken into account early on. I'd actually created most of the body work before realising that there were several details that I hadn't taken into account. Some of this was able to be added but due to the fact that I had already split the body into different parts, it made for a rather awkward process in adding certain others, due to the fact that it would directly effect the seams. I eventually decided to leave out the details that proved more complicated and learn from the experience for the future.
Another thing that made for some awkward viewing was the thickness of the different segments. While extruding the different segments I originally just increased the thickness freely and later began to manually set a thickness, since I had essentially guessed the value for certain segments I later came to realise that it made a huge difference in the different parts of the car. This is most prominent on the windows since that's the only real place that it's particularly visible given the fact that it's the only real open area. However if it was to be rigged and animate as I plan to do in future projects, it would create some undesirable complications.
On to topology, in general I'm typically quite confident in my command of topology, however I feel it may have been a little neglected when creating the rear of the car, particularly in the corners. Building out certain areas, particularly when connecting 2 separate flows to one another, proved to be more difficult than I had anticipated. I also fear that I may have rushed things a little at this point and the resulting topology, although true to the shape of the car, ended up being forced, resulting in an uneven surface that became exaggerated when attempting to add more detail such as the rear lights and the seams for the boot. Unfortunately, once again, I noticed this too late and fixing it would require completely remastering a large segment of the body, so I again decided to move on and learn from the experience for next time.
One last thing I will mention in regards to topology is the different pinches that can be found in various parts of the model particularly around the door handle. These artefacts could be fixed by taking a little more time to look at the topology and space out the vertices evenly.
The next issue was the windows, I didn't spend enough time considering the shape of the window arch, it was only when I began testing lights and creating basic renders that I noticed the odd shape it created, the window appears to curve back on itself towards the rear, creating, as it appears from the front, the effect that it suddenly angles back part way around. Again fixing this at the time it was found would have disrupted the entire shape of the car so it was best to keep note and bare this in mind for future reference.
A smaller issue is the tyres, I intentionally left off the tyre tread, as the work station I was using had a painfully low spec and was getting fairly old. Using this work station for any kind of higher poly work caused constant crashes which became became abundantly clear when creating the grill at the front of the car. It took around four different crashes to create something that would work as a grill, so I decided to leave the tyre tread as it would end up costing far more time than it would be worth.
The shaders were pretty straight forward, once again my work station was dying out and creating low quality test renders were still taking a long time, so not a whole lot of testing was done. I did try a few things with the car paint but it seemed to work pretty well with minor adjustments to Renderman's LMMetal shader. I also feel that maybe the wheel rims are little too high in specularity, I've since noticed that most cars seem to have quite dull rims rather than solid chrome.
The lighting was pretty straight forward, I used a 3 point lighting system to light the model itself with a couple of fill lights illuminating the studio and causing some minor bounce light.
My main take away from this project is that I need to organise my time better and be more patient when creating the topology. I also need to take a lot more reference into account and do a lot more testing.
Another big thing is that I needed a new work station if I wanted to create some higher quality work in the future.
Do you have any constructive critisism of your own? Have any questions? Feel free to contact me on Twitter or by email.