Tuesday, September 30, 2008

Lessons Learned

You know, it's been difficult at times to forge ahead and hone my craft when I am, in essence, a freelancer with hardly any offsiders to refer to. And without other VFX artists around to share ideas with, well, sometimes things feel a) lonely, and b) stagnant. So when a learning opportunity rears its head, you take it.

As such, I want to share some things that I learned from my work on Cupid, things that seem pretty obvious when you think about it, but that nevertheless may not be so apparent when in the thick of a one-man VFX project. So here goes.

Get involved very early on
This is a tough one, because it's down to a good director and/or producer to understand when to bring the effects people. Ideally, that time is well before much of anything else is done.

With Cupid, I was brought in after principle photography. This is not necessarily a bad thing, but it leads to a lot of missed opportunities. The main downfall of not starting the VFX process at the beginning, is that the effects people lose the ability to have input into the shots. So for Cupid, I was basically in a position of reverse-engineering the director's thought process so that I could understand how to approach the shot, which therefore made my job that little bit harder.

Be on set whilst shooting the VFX shots and/or plates
This feeds in from the previous point, in that it's something I couldn't control, since I was brought in to do the work once shooting had wrapped. Which is a pity, because had I been onset, I would have gleaned a lot of important insights, and also may have had the opportunity to point out things I may have missed.

For example, one shot required me to create a reveal of a character who had previously been invisible. This means having a shot with the actor doing his thing, and another shot of the same scene without the actor (a 'blank plate'). When I received the footage, it quickly became apparent that lighting conditions had changed and that the camera had moved the tiniest little bit. Had I been onset, there's a possibility that I may have been able to notice these small things, since everyone else onset would have been looking at other details. Ultimately, the 2 plates were so different that a 'shockwave' effect was added to hide the transition from 1 plate to the other. But it seems to have worked, because when I scrutinise the final shot, I can't see that seam.

The other major factor is that I could have taken lots of photos. There's nothing like being there on the day of the shoot for finding all manner of reference materials. For example, I requested a floorplan showing the lighting of the shot, and was pleased to get a fairly concise image with all the information I needed. But it was a poor substitute for physically seeing the lights, and being told that said light was elevated by 2m still doesn't give me the complete picture. What that means is that I had to study the footage for lighting information (a good practise anyway) and try to discern the info I needed. But one photo of the actor standing next to said light would have told me so much more than "2m elevation".

Work side-by-side with the director as much as possible
Perhaps the hardest lesson learned on this project. Because of geographical constraints, not to mention busy schedules, Melvin and I worked mostly by proxy. To break it down, I would work for a bit, render stuff out, send my work to Melvin via the internet, he would watch it, form his opinions, call me with feedback, and the process would repeat.

This is bad.

Fast forward to the last 2 possible days, in which 2 deadlines had already been broken and things had to be finalised by the end of that final day, no ifs or buts. Melvin was concerned because particle effects that were to be seen in every shot weren't working, so he canceled his entire weekend to come work alongside me, and we quickly covered a lot of ground. In fact, render times and fatigue ate up most of that weekend. But we got it done, and the results were so much the better for it.

What happened on that weekend is that we got immediate feedback, and the evidence was visual, not verbal. Melvin could take the mouse and try things, if he wanted to. Or I could quickly spend 2 minutes mocking up an idea, instead of explaining it over the phone and taking the risk of Melvin misunderstanding. Most important of all, though, was that Melvin could simply communicate his ideas to me visually, such as something on Google Images, or a quick sketch. So much so that within an hour of working side-by-side with him, I knew exactly what he wanted out of the effect. The rest of the time was spent fine-tuning and rendering.

The last minute will always be utilised
How long is a piece of string?

That kind of rhetoric gets thrown about a lot in this industry, usually in answer to 'how long will it take to do?' But what I'm getting at is that this is an industry that always does a mad rush at the eleventh hour; work that is at its most feverish at the last minute. That final weekend was very worrying in some respects, and it was downright painful to keep pushing into the small hours, especially after so many late nights. And I had to remind myself and Melvin that we were doing the last-minute rush in the same manner that any big-budget feature film might do. It was exactly the same.

Of course, things already mentioned fed into the need for that last-minute push, but that's not the important part. Because even if things had been "completed" two weeks in advance, that merely creates the opportunity for adding extra detail; you start finding out that there's more string than you could see.

Your gut feelings are always right -- well, some of the time
Broadly speaking, your initial reaction to something is the one that will give the most perspective on how you should proceed - not just in this industry, but life in general. But other times, that reaction can be deceiving.

Sometimes, something that doesn't work for you the first time will grow on you if you immerse yourself in it. Possibly, this is why they tell you that the only way to appreciate a painting is to sit in front of it for 10 minutes and take it all in. Regardless, an effect or piece of footage may seem abhorrent the first time you see it, but if you loop the video, you may find that it starts to feel right. And the reverse is true: just because it seems right the first time doesn't mean that it will look good after that.

I bring up this point, though, because of the final weekend in which Melvin and I rushed madly to finish it all off. As stated above, it was due to the particle effects in the shot, which weren't working. I was very proud of my particle effect, and really wanted to make it work, but it was clear to me that it wasn't working, even before Melvin pointed it out. So after we redesigned the effect, Melvin asked me if I was okay with the change. (I should add here that it was both a personal and professional question, and I addressed it as such.) I told him that what I wanted the effect to be was the right thing that the shot needed, not what I felt should be in the shot.

There's no sense getting precious about your work. Filmmakers have to be prepared to remove their favourite scene, if they feel that scene damages the story they're telling. Same goes for writers, or just about any creative medium. What makes it great is the things you omit. There's no reason why visual effects should be any different.

Take breaks. Often.
Let's face it: you're bound to be doing long hours at ridiculous times of the day. You may work until you crash, then get up 2 hours later and work for 3 more hours before crashing again. Hey, whatever works for you. But inasmuch as you need to concentrate on your work, you shouldn't neglect the needs of your body.

Look away from the screen, often. Not only does it help your eyes, which no doubt have trouble focusing on the images onscreen by this point anyway, but it also helps how you perceive what your eyes are telling you. Any given VFX shot has you staring at the same image over and over again, and although you may alter that image considerably over a few hours, you nevertheless make those alterations one change at a time. Coming back after 10 minutes gives you the opportunity to see the image from a somewhat fresher perspective - very important when you only have a few days or hours to complete the task at hand.

Of course, there are also health concerns. On the Saturday of that monster weekend, Melvin and I worked until 5am, and we only stopped because I stated that we were becoming too tired for our own good. As it is, it's been difficult for me to regain a healthy sleeping pattern, even after 10 days. We also didn't want to start making stupid mistakes that we'd only have to correct the following day, thereby wasting precious time.

The main thing, besides health, is that you need to keep yourself both focused and motivated. And there's nothing like staring at a computer screen to sap both qualities. Taking 5 minutes here for a smoke, 10 minutes there for a snack, really helps to keep up that motivation and focus.


That's about all for now! I'm going to take my own advice and take a break now, because it's getting hard to stay focused. I may post up some more, but for now I think I've covered all the main points I wanted to.

Sunday, September 28, 2008

Cupid

First off, I really have to thank my friend and former work colleague, Melvin, for providing me with this opportunity. As per my previous post, I have spent a lot of time learning about VFX compositing, and he came to me with the offer for a VFX sequence in his short film. The timing was perfect, as even a few months earlier, I wouldn't have felt ready to tackle it (I certainly wouldn't have turned it down though!).

Melvin is presently studying at AFTRS, Australia's premier film school. This is his second year, and he's spent the last several months directing Cupid, a short film about that dude who shoots love arrows (not a euphemism).

I'm not certain if I should be sharing my shots from the film, since, although my shots are completed, the final film is not. But I of course am putting them into my demo reel, so they'll be posted soon enough. So for the interim, here's a test I did for the major effect shot (that would be Shot 7 -- oh, how I hated you!):


Gives nothing away, right? Well, let's just say that it's a reveal shot, obviously of a character. To explain what's going on would be to spoil the film's major conceit, and I'm not a spoiler! At any rate, that was just a quick and dirty test to discover how to do the shot in question. The final result is quite different to what you see above, but tests like this pave the way for the final, more complex composite. From the above shot, Melvin was able to instantly send me feedback on the elements he liked, and the ones he did not, although I must say that the core of what you see in that youtube vid is still intact.

What's changed is the stylistic approach. This is because the design was never really nailed down at any point during the process, instead being an elusive, amorphous blob of potential that manifested itself more and more clearly as time went on. It's a quirk of the manner in which Melvin and I worked together, and the upshot is that we learned some imperative lessons when the fat lady started singing.

So consider this a teaser: I intend to outlay the top lessons I've learned from the experience in a future post. For now, I'll tell you that it damn near wiped me out! It's almost a week since my part in all of this ended, and I'm still not fully recovered. But we literally took it to the last possible second to complete it all. I'm talking about starting the final render 9 hours before it needed to be handed over for grading. That after a long, sleepless night and some sleep-deprived delerium in between short periods of work. But I took comfort in the fact that this is exactly what happens in the industry at large. Livin the dream, man ... livin the dream!

Saturday, September 27, 2008

Almost back from the dead

Oh man, I can't believe it's been 7 months since I last posted.

Oh well, here I am, updating away. I suppose it's fair to say that I haven't really had anything worth saying throughout the year, least not in regards to VFX. But I've been spending the time well, taking any spare moments available to improve my craft and my knowledge.

Surprisingly, I've found that Adobe After Effects is a rather good compositing program. And as of last week, I have completed my first bunch of VFX composites for film work. More on that in a bit. I had been using Combustion primarily until the start of the year, and thinking of After Effects as more of a motion graphics program. But when faced with VFX shots for a film (don't get too excited, it's for a short student film), I chose to work with After Effects. Why?

Well, primarily because it's the program I've been using more often recently, but there are other factors. I find that whilst Combustion offers better results and more comprehensive tools, After Effects facilitates a more streamlined workflow. Naturally, there are features in the one that I miss in the other, but that's just how it is I guess.

The other factor is that there are far more tools available on the web for learning After Effects. I have, for instance, started a love affair with Video Copilot, which is an insanely helpful site. I was quite naive to the potential AE affords, but no longer, thanks to Andrew Kramer and the team at said website.

Conversely, Combustion offers very little. I don't know why, but finding anything more than the most basic tutorial for the program is very rare. I still can't wrap my head around Combustion's native motion blur feature, or how to make it useful. Using the velocity channel in an rpf file is very good, but processor intensive and has some annoying limitations.

So long story short, I've learned quite a bit. I've learned how to composite VFX shots, which I guess is a very good skill to have. I've also learned why we go through the processes that we do in VFX compositing, or at least, I've come to some promising conclusions as to why we do it that way.

Now it's time to start talking about this film work....

Friday, February 1, 2008

... if it weren't for the fucking clients....

From Clerks:

RANDALL: This job would be great if it weren't for the fucking customers.

...or something like that.

Man, I hate clients. Well, not all of them. It's just that clients don't know how to be clients. You see, the process should go something like this:

  1. Client approaches you with request for (graphics/video) work to be done
  2. Client provides you with everything you will need (designs, illustrator files, logos, colour schemes, instructions, etc) -- this one is important!
  3. You listen to client's wishes, repeat them to client, and set off to work
  4. You show your preliminary work to the client for feedback
  5. Client provides feedback, requiring that you make some changes and tweak things
  6. You make the changes and tweaks
  7. Repeat steps 5 and 6 until it's done
  8. Client pays you
  9. You buy new game/X360/new graphics card/booze
  10. Pay rent and other overdue bills
I'm having a rant just now because of a client who asked for a design without giving any input. At all. They gave me a copy of their invite which was just blue text on white with a sort of blue stripe feature on 1 of the pages. So I did my best, and came up with a design that I thought, frankly, was crap. But the client liked it, and requested a tweak. That was fine.

AND THEN I was asked to make further changes -- in essence, to throw away the majority of the design in favour of a baffling new one which squishes everything into a third of the screen and leaves the rest for a purdy animation that was really only a background loop.

So, being studious and industrious and all that, I did what the client wanted, because hey, we give em what they want, right? Anyways, I have no doubt that it shall be rejected and they will ask for a further change, but one that is in line with my previous efforts. So when they do that (and they will, I'm 95% sure of it) I will be left, wanting to say 'thanks for wasting my time' but of course, unable to because we need clients, right?

Well, I'm with Randall. I'd love this job... etc, etc

Monday, December 31, 2007

A preview

Ok, so here's that shot I've been talking about. I still haven't finished it, but it's a nice work-in-progress for you to check out:



So the animation is out, and the camera move gets downright weird at one point, but that's okay, because it was all about lighting and compositing, and in that sense, I feel that I've almost nailed it. I'm just gonna add volume lighting and some more particle effects -- and a heat-vapour once the rocket's flame clicks off -- and it should be done.

After that, I'll put together a tutorial on how I did it, but no promises just yet :)

Now then, let's get pished!

Saturday, December 29, 2007

WIP for year's end

Was just impatient to show off a work-in-progress of my recent project. As mooted before, it's just a practice exercise for me to learn with, but I'm happy with what I've got so far. Check out these pics:

Before

After


As you can see, I've dramatically lit the scene by playing in Combustion. I will reveal later on how I did this, once I've got the whole thing sorted out. Note also that the flames on the logs were added later. It was made in the same file in 3D Studio Max, but rendered as a separate pass. I think it can do with a bit of tweaking here and there, but it shows the power of compositing and how, with a bit of smarty-pants trickery, you can get some great atmos out of an otherwise boring shot.

Coming up next will be some volumetric fog and some extra particle effects, plus the full video which will really show off what I'm trying to achieve here. But that will have to wait until 2008!

Friday, December 28, 2007

Tail end of Xmas; a new year; impatient to do work

So the big consumerism day has come and gone, and now we're all gonna get really pissed , set of fireworks and pretend for about 5 minutes that we're much better people than we really are -- really!

To be honest, I want to get that behind me as well. I suppose I've grown past the whole new-years thing, and I'd much rather sink my teeth into some graphics instead.

I've been working on a test shot, which I shall reveal soon, designed to teach myself how to work with lighting and some other elements when compositing a 3D shot. I've spent a few hours here and there over the last 2 weeks putting this shot together, and it's almost time to render everything out and sink my teeth into the exercise proper. But I didn't go into work today (what with there being no actual work to do) and therefore did no extra on this shot. And since I'm not there, I can't even post up a work-in-progress pic. Oh well. I may do some tomorrow, and render it out whilst I get drunk over the new year. Then come back on January 3rd and try with an almighty headache (so I'm presuming) to work the composite.

Well I have the highest hopes of it all working out, because then I will have passed a milestone in terms of creating a visual effects shot. After that, I will only need to perfect my camera tracking and on-set data knowledge, and I might be worth some money to the industry!

By the way, if someone is willing to give of their time to help me learn about on-set data collection, and how to use it, please email me!

End of 2007 -- peace out! Let's not screw up '08, eh?