I know in my last post I suggested these test clips might slow down a little, but I actually made another one already. I said I’d only do them if I had a specific idea for something I wanted to try, rather than doing them for (what feels) like the sake of it. Soon after writing that post, I genuinely did have something I wanted to try, so here we are!
Ever since hearing about ‘Mixamo’- the free Adobe resource (what’s the catch lol) of 3D characters and animations you can use in your own projects, I wanted to utilise it in some way or other. I’ve had a few ideas before, but not really got that far in executing them …until now. I think my other ideas were over complicated hence never carrying them out, though this idea was definitely something I felt I could complete.
So what is it?!
I wanted to try rotoscoping over one of the 3D characters. Why? A couple of reasons. Firstly, to see how it might look and if it would work altogether. I’d been curious about this, just to see how much drawing over a 3D character would differ than drawing over filmed footage of a human. Whether it would pick up on certain nuances, or even look anything like what I was trying to replicate. I think I was curious about trying these things, because I thought it could potentially be a great way of making my next film, rather than using real actors. More on that later.
In terms of the process, that barely differed- it was a quite simple case of importing the footage and redrawing it in a style I felt suited. In fact, because the reference footage was computer generated, it was quite a lot easier to rotoscope as more things with it stayed fixed. Such as hair. Rotoscoping hair on filmed footage is not fun (IMO), especially long hair! The footage was also cleaner and therefore easier to see some of the outlines.
A negative I found though, was the fact that because it was computer generated it was somewhat glitchy. There were occasional artefacts sticking out of the character at various points and sometimes the body/ clothing was not acting especially naturalistic. These things were fairly easily fixed though as I could just re-draw them in a way which appeared more ‘normal’/ natural.
Overall though, the concept seemed to work- or at least for this piece of footage it did. I went for a shot which contained a lot of action/ movement to it. I’m considering maybe if I tried it with a more still shot- how well that would work. Something I’d like to try anyway. It also didn’t look a complete mess (phew) and it didn’t appear too ‘uncanny valley’ either- or at least not in my eyes. Perhaps because I went for a highly stylised black and white finish, the end result came out really nicely. I skipped a lot of the facial features/ detail, which perhaps cut down on any uncanny feeling there might have been, if I went full detail with the face. Who knows…
The second reason I wanted to try it was, as if it did work/ look ‘good’, then I might consider using it in my next short film. As you probably have guessed from reading this blog, I’m a person who likes to experiment with technique and technology, so this felt like the perfect testing ground! So, as stated above, I felt like this technique did work (though needs more experimenting on different clips, I’d say), but if it was something I’d want to use in future productions, I’d have to learn how to animate 3D characters, or at least to a certain extent. I mean this would be if I wanted them to do specific things and act in a certain way. So would there be much point taking a lot of time to learn 3D animation enough so I can use it in a short film, over hiring actors and filming them, to get a look which is similar? At this point, I’d probably say no, but it’s not like there aren’t advantages of using 3D characters instead of real life people. One main one would be that it would cut down the stress/ hassle of being on set, alongside sourcing a decent location to film in at a reasonable price. It would cut down on filming budget, but raise the amount of hours spent, therefore increasing the time budget.
Like I said, I’d want to make a few more of these test clips in the same way, to see whether it works with different shots and characters, rather than just the one I tried. I think this might help me decide if it’s something I want to pursue on a bigger project. I have even thought that it might work in a more abstract project, rather than a narrative film, because I can just use the pre-loaded animations, rather than making my own (so I wouldn’t need to spend hours and hours learning character animation in Blender). Again, I feel more test clips would help me decide here.
So in conclusion, I feel it was successful, but needs more testing. I sound like a scientist in how I approach my animations!
Here’s the clip:
No comments:
Post a Comment