‘Agile’ Rendering and Accidental Masterpieces

Hey, thanks for taking interest in my humble blog! This is a long-ish post, so grab a cup of coco, start that background render you’ve been putting off, and we’ll begin.

Still with me?

So, for a while now I’ve been thinking about an alternative approach to rendering. If you read my previous blog entry, you’ll see that I’ve been frustrated with the slowness of the whole thing. I’m a perfectionist at heart, and the time it takes to setup a model, the shaders, lighting, background, textures, pose, hair, clothes, etc, etc, takes up WAY too much of my time. Combined with the ridiculous 2-4 hour ‘final’ render, I wonder if it’s even worth it. Even though programs like Poser and DAZ|Studio compare themselves to a photo studio; I see the whole CG rendering approach more alike to painting. Traditionally, it’s slow, and requires a lot of planning, and fine tuning to get the best results.

I’ve been reading a lot about Agile Development and Production, for university, and I’m wondering if there is a way to connect some of the principals of agility to 3d rendering. So, what I want to propose (and I haven’t really figured this out yet) is more of a photographers approach to rendering. I’m not saying photography is easy, or that rendering should be lazy; but I want to try and capture the fluid and more flexible aspects of photography.

For example, with a digital camera you can take loads of photos really quickly, and with any luck some of them will turn out fair, or even pretty good. I saw a documentary recently, and someone commented that there are no ‘accidental masterpieces’ in painting (and art in general), but there are in photography. I wonder if there’s a way to harness some of this in 3d rendering. Looking at my past renders over the years (not online), some of my best ones have been very experimental, or quick ‘test’ renders. None of them were perfect or ‘realistic’; but there was a ‘special something’ about them that brought them above the average. Maybe a sparkle in the eyes, a slight expression, a shadow here or there…

So what I’m proposing is slightly against the norm. It’s about throwing away the pursuit of photo-realism (because, arguably, the end result will be ‘uncanny’, either way); ‘final’ renders that take over an hour (and then some); wasting time tweaking things to endless perfection. Unless you have 2 or more good computers, why tie up resources rendering one big image, when you could be doing several?

So, instead it’s about developing a more ‘agile’ system or approach to 3d rendering, that emphasises creativity, quickness, and flexibility — the idea being that ‘actual renders’ (even unrealistic, small, or otherwise flawed/imperfect) are more valuable than sitting at the screen tweaking morphs, shaders, lights, etc. or waiting for the computer to render a so-so image.

I’m not saying this approach is better than the traditional method of working up a really good render. Nor am I suggesting you should throw away centuries of art theory and practise. All I’m proposing right now, is an experiment in trying to emphasise ‘speed’, ‘creativity’, and ‘imagination’, over ‘perfection’ and ‘realism’.

How you actually go about this, of course is the million dollar question. You might feel a bit short-changed right now – if I knew what the holy grail of rendering was (other than lots of: HARD WORK, PRACTISE, LEARNING, EXPERIMENTATION, and EFFORT) – trust me, I’d be selling it on the marketplace. I’m currently experimenting with speeding up render times and trying different, unconventional approaches to rendering, more akin to fashion photography (like doing several ‘snapshots’ of a model, rather than one perfect render). Whether this works, and how you measure the success of it, isn’t really clear either. However, perhaps one advantage of this approach could be to increase one’s creative potential, allowing you to explore promising ideas and do them ‘properly’ at a later date.

Anyway, if you found this interesting, then I’d love to hear your feedback on this.

Related topics: