Many of the projects I work on have strict requirements for content, technical constraints, and a limited time / budget to complete them.
As I had some time available, I decided to work on something that inspired me, and take the time to explore new work flows and software, with out too many of the normal constraints (though they did crop up eventually!). I also wanted to ensure that with final visual quality took president over everything else, which turned out to be a great test for my hardware and patience.
I chose to use the new(ish) open sub div modifier in 3ds max to do most of the modelling, and use substance painter 2 to create materials, (and final renders, though I didn't plan for this).
If you are interested to see my workflow, it can be seen here in my blog, I recorded quite a few time-lapse videos as I went.
My main point of influence was this lego model I found on the internet, as well as concept art by Ian Mc Que.
Using Open Sub Div led me to a slightly different workflow, I found it fast to quickly create complex hard surface shapes, but also quick to add unnecessary triangles. On the plus side there were almost no destructive processes, which led me to be able to clean up and reuse the models very efficiently. I was hoping not to, but ultimately had to fall back on some more traditional subdivision techniques, such as adding support loops and sticking to quads.
It was the first time I had used painter 2 to create texture sets, and I found it surprisingly stable (I don't think it crashed once) the only downside to it was as my model used over 2 million triangles and 5x 2k texture sets (25+ textures) it really started to slow down towards the end, taking 5 minutes or so to do anything. I also realised that many of the details I had modelled in could have been added to the height maps, for not much visual loss (and a mass time and efficiency saving).
By far the best, and almost undocumented workflow to come out of using painter was the ability to paint and edit in height maps, then quickly convert them to the base normal / ID maps for use in procedural material creation. This saved me hours of unnecessary high poly modelling and baking, and is definitely something I will use in the future.
Ultimately I was very happy with the end result, though next time will take the time to model lower poly, to save time waiting for loading bars later down the line, and enable the end model to be used in real time games engines.