Now is the time to really get going.

Hello again.

A week ago I went out and recorded some nice footage outside my house. The focus of the project has shifted from indoor shots with people to outdoor shots of the plant pots coming to life and running towards my house. Here's two example shots to show how well they matchmove. The first shot has a prototype plant pot character in it, the second shot just uses cubes. The second shot matchmoves slightly less well but I'll get that buffed up real soon, and also there seems to be some missing frames, no idea how that happened.

I'm currently doing the animation on my plant pot characters, and then I'll start dropping them into matchmoved shots, and this should all start coming together. Next week I might take the camera out again and get some more shots, I just need enough shots to make a coherent piece. At this point I've had to scale down the project considerably, but when it's up to scratch I'll consider adding more ideas. Right now it's really about getting something ready to submit which shows I can matchmove.

I'm going to go back to my previous post and add the early test footage which I never uploaded.

April 1st

So i'm quickly realising that the footage i've captured already isn't good enough, and here's why:

It's overexposed: everything is much too bright and a lot of data was lost. I blame this on the inaccurate screen of the camera, but in fairness I should have used the zebra lines display function. The next time I film, i'll put more effort into proper setup.

I didn't think: Turns out that most of my shots are just nodal pans, I need to factor in LOADS more parallax to get more 3D matchmoves. When I next rent a camera, I won't be taking out a tripod at the same time, the tripod is mostly useless for creating parallax. I managed one shot which I should really upload. I'll do that when I get home. I did that by putting the tripod on a tray and getting somebody to drag that along the floor while I held the camera steady. That makeshift solution wasn't very practical at all but it did work. There might be some more clever ways I could use to get the shots really smooth, so maybe I should consider that while planning shots, and take the tripod out if needs be... Then again, it's nice without a tripod because it makes it much easier to transport the equipment, which is a plus for motivation. Still, the quality of the final film should be a bigger focus than whether I can be bothered to carry equipment.

They look really cheap: I think I've aimed my film too high as far as production values are concerned. I don't think I can emulate the kind of thing i'm thinking of in my head. The problem is, I was thinking "epic" for the film, but in order to achieve that i'll need permission to film in university at night, i'll need actors. Combine that with the possible need for multiple shoots and I think we've got a problem...

It might be possible to go to the university when it gets dark and film somebody walking up to the door, to set the scene. They might need to run towards the door so maybe security would be suspicious of that, but the worst that would happen is that we'd be asked to leave, so potentially no biggy.

I was thinking I might change the direction of the project to look more "amateur"... So instead of changing the setting and actors, i'll just put them in a situation with amateur looking footage and ad-libbed reactions. That way it would look like a home movie, kinda like Cloverfield I guess. I'm as tired of Cloverfield clones as everybody else is, but I've gotta get a project together right?

That's the lazy option, but with a bit of clever planning I might be able to pull together something a bit more pro looking. I'll definitely try that.

Also I REALLY need a much more solid idea of exactly what shots will go where in the film. For this, i'll need a proper storyboard, shot list, which shots will use VFX, how long that will take, see if any adjustments need to be made. I should have done that from the beginning I guess, but I didn't, so I've got to do it now.

I've also got the option of maybe taking some Steadycam shots next week... With the possible new direction of the film, that wouldn't work very well at all, but this could factor in very nicely if I wanted to make it look more professional. The outside shots would need to be at night BUT I could film some inside shots with artificial light while using the steadycam. Perhaps I could look into a day to night shot? These are all things i'll be considering over the weekend and updating the blog with my thoughts.

But I've got to go now because the room is being occupied by animation students.

Update 24th march

So I did some filming a while back. I got a test shot which turned out quite well but I didn't have much in the way of USABLE MATERIAL or anything related.

So I filmed some more test footage following the storyline better, but now I think the end project is going to look awful.

But i'll do some stuff with this footage and see if it ends up okay. I might have to change the scope of the project so that it'll look good within my capabilities.

FOR EXAMPLE, changing to to a Cloverfield handheld cam type of thing might be cool, and it would mask the terrible "set" I'd have to use, making it look like more of a decision.

Concerns and solutions.

The viewer may be dissapointed - If the viewer is expecting giant robots, how can I avoid dissapointing them? With the fact that it's a comedy piece, and one of the main ideas of the piece is to deceive, it may be fine, BUT just in case, here's my Solution: Have a final shot of the robot's shadow approaching the characters, and have the main character say "aww no, I forgot about them!".


The entire point will be missed - Solution: I'll just have to make it entertaining enough even to people who don't know the point of the piece.


The fake objects will look fake - Obviously i'm trying fool the viewer into thinking that the real objects are 3D, and that the 3D objects are real, so Solution: I'll need to make smart choices about what would be easy to model realistically. For example, a plant would be a bad choice because it would be hard to make the leaves move realistically when thrown.

My idea: Red Herring

So it's fair to say that i've not had good luck with ideas these past few weeks, instead doing vague tests which i'm not sure will apply to the final project. However, this weekend I came up with an idea that i'm very happy with. I'm currently writing it out in full and doing a storyboard and a small script, but i'll outline it here.

It's called Red Herring. It will be about tricking the viewer into thinking something is going to happen, and then something else happens. It's about taking the mickey out of film conventions and throwing the viewer off the scent.

The original idea was "Attack of the inanimate objects!", and that idea is being incorporated into my new idea. I originally imagined a line of objects, some real, some CGI. The viewer could try to guess which objects would move by which ones they thought were CGI, or which ones looked the most threatening.

The main threat in my film will be self-aware plant pots throwing themselves at people's heads, but the viewer will THINK that it's going to be giant robots.

When the film starts, there'll be two people walking down a street, they'll walk past a car with the radio on, and you'll overhear a news report saying "scientists warn about the imminent danger of giant robot attacks". You'll walk past a kid playing on a portable games console, and it'll show for a split second some footage of a robot destroying a city, and his mother might make a comment about these games being too violent, and then the kid might say "but it's not realistic at all!".

Then, the ground will start shaking, and people will look to the parting in the building's expectingly. THEN, a pot hits one of the guys in the head and the shaking stops, and you realise you've been tricked into the fake robot attack story.

I'll write the rest of this later.

Second update.

So I rented a camera on tuesday and i've been doing a bit of filming around the house mainly. I'm going to try and get some characters to walk around on the floor, i'm going to try and replace my desk with my virtual desk.

Of all of the ideas i've had, my best one so far is to have inanimate objects flying and hitting people in the face. How would this take any skill at all? Well it's all about getting the real object up on screen and then making people guess which object it is which will fly at the person's face.

I could actually replace multiple objects with 3D versions just to make people really confused as to which objects will begin flying. I was thinking "attack of the potted plants" but plants might be quite hard to model... So maybe just ATTACK OF THE INANIMATE OBJECTS would be good, saves me having to make a choice about what catagory of objects I want to throw at people... Except that they have to be inanimate...

BUT WHAT IF I WANT THEM TO ANIMATE?!? Tough decisions! Anyway, for now it's test shot time and as for my idea... I guess I ought to start writing a solid plan right about now.

Matchmoving is hard with SD footage.

After learning how to use the software with tutorial footage, I've gone right ahead and recorded some footage of my own. The problem is, my camera isn't HD and I was too lazy to borrow one from the uni.

I have modelled a desk for my previous project, so I decided to record some footage of my actual desk and put the 3D desk on top of it. This is however proving difficult as the points which were automatically selected on the footage are non reliable (grey).

I'm going through and making all of the good ones hard, I am also adjusting and extending a lot of them manually, in an attempt to get the matchmove gradually more accurate. It's been several hours now so I'm starting to think I should just give up, but if I manage to do this it will be a wonderful day in the world of Alex.

Now back to matchmoving I guess.