Fwoosh postmortem

"I did a fwoosh postmortem already, didn't I", I kept thinking, but upon checking, it became apparent that indeed I had not yet completed the postmortem for our latest, perhaps greatest production. Because it was done by me (and not Tim), and I'm doing the postmortem as well, some may think I'm being biased, since "the production we shall not name" is shown to be total crap and this "fwoosh" postmortem is glowing with praise. This, of course, is patently false, and anyone who claims otherwise is a dirty rotten scoundrel.

So let's begin by saying that a lot of what we learnt was in fact from "the production we shall not name", and these same mistakes were largely avoided. Taking care in these areas, of course, left problems in other areas, which is what will be discussed here.

What we did right

Listening to my advice from "the production we shall not name", payed off, since the problems encountered in that production did not occur. The production was made with a view to remove a lot of the variables. Black and white with no sound made it easy to concentrate on story and lighting. Not that everything should be black and white and have no sound, but things like colour balance were never even brought up as issues. We could talk it out and give direction as we were filming, and since the film was going to be sped up, there was little problems with short time stuffups.

The main achievement of the piece was that lighting was carefully considered. Manual exposure settings, even without light metering, worked well to give relatively even lighting throughout the piece (one exception which we'll discuss later). The focus was auto, but that wasn't such a bad thing, since we weren't shooting anything wacky focus-wise anyway. Indoor scenes were illuminated with plenty of lighting, and outdoor scenes had different exposure settings. If anything bad could be said for it, it was that we filmed in the middle of the day, a bad time for shadows, and it shows.

You know that the action is well carried when glaring irregularities go unnoticed. Everyone watching (except us) didn't seem to realise that the car changed sides or that Nathan got in the passenger side and ended up on the driver side. This is because the action was well written and carefully carried. Point of action was not abruptly moving around, lighting didn't change between shots, and sequential shooting meant only the things that were meant to go wrong went wrong. We stopped the script from getting too specific to allow us to take full advantage of opportunities that came by. The handshake and the car bit was pretty much made up on the spot.

Establishing shots. There was just one, but it was enough to give the viewer spacial awareness of the entire piece. There were no cutaways, but that didn't really matter, because our improvised cutaways worked well. Enough said about that.

What we did wrong

The time of day (and the resulting shadows) have already been mentioned. We should've waited for better lighting, but I don't mind so much, and we don't have those giant mirror thingies, so it can be excused. The thing that we did really wrong was that we shouldn't have done mixed indoor / outdoor shots. Not only did we have the character moving from indoors to outdoors, but we have him actually going outside from the inside, which is a mistake, especially considering the time of day. To make things worse, my pergola provided sufficient shade to make the establishing shot confusing, and make the lighting confusing (even though the lighting was even, one shot was in shadow and one was not).

A minor problem, but we shouldn't have been so stingy on takes. Everything was done in one or two takes. This was great for the most part, but it made mixing stuff up hard. The appearing broom shot is... let's say you can tell. Still, it wasn't hard, and improvising in post is what made it funny, in some ways.

Another minor issue is that we need someone saying cut before leaving character. Specifically, there's this bit where Michael "dies", but gets up immediately after hitting the ground. This makes the shot really hard to work with in post, since it's sped up. Nathan just stays in character ad infinitum.

For something that small, post was difficult, time consuming, and HDD space consuming. We need to figure out a workflow, and either split up future projects between people or do something to make stuff more manageable (rough cuts at one person's house, then split up clean cuts, etc). I ended up repeatedly re-rendering stuff, because of problems that kept cropping up. Anyway, that's a lesson for another day.

That's it. The placecard effect didn't look too great because it wasn't jiggling around. I don't know what lessons can be learnt from that though.