Dan Hume's Blog

New Title Scene
May 11, 2012, 4:28 am
Filed under: After Effects, Extended Major Project, Video

After much criticism of my 3D composite of the title ‘iSolation’, I’ve changed the shot completely with a 2D version in a different scene. The only negative feedback I had after people had watch the documentary was 3D composite. Even after watching the whole piece, people were still irritated by this really short scene at the beginning. Tony Chance, who guided me through my development of this project, said that it looked out of place and unnecessary.

I initially chose to create a 3D title because I’d seen a really impressive tutorial (mentioned in a previous entry) that showed me how to composite 3D text into a live scene. What impressed me the most was the fact that you could light a 3D object using the pixel data from the shot you want to composite into. I was inclined to do this as I wanted to learn the basic principles of compositing 3D elements into video. It’s something I’ve been struggling to actually do for the last couple of years, but I think I’ve now managed to grasp the concept behind compositing really well. Despite all the efforts of this, the outcome didn’t work in the final piece.

However, I didn’t want to replace the 3D composited shot with something really basic, such as 2D text with a black background. I decided that it maybe better to composite a 2D version of the title into a slightly different scene that convey’s the idea of isolation a little bit better. I found a shot that I recorded at Maida Vale tube station when it was completely empty and I thought this would be a good little scene to have the title appear. I really wanted to make the text become an actual object within the scene to look as if it’s isolated itself.

This lead me to reflecting back to an initial tutorial I looked at, which was to composite some text into a still image. If you look at my header image at the top of the blog page, you’ll see I added some text into the scene of the image using a 2D text object. This was all created in After Effects using it’s 3D workflow.

Here is a video showing the new shot without the composite and then the shot with the 2D composited.

It was initially easy to set up the text into the scene. The first thing I did was created a new white solid and turned it into a 3D layer. Like in many 3D programs, you create a plane to allow you get the perspective right to align the object you want to fix into the scene.

I then turned the plane into a grid by going into the effects menu. This then gave me more flexibility to see how well the plane was positioned to the roughly the same angle as the ground in the shot. I also created a new camera, which I also used to help me make the 3D environment replicate the perspective of the actual scene. The camera is a way of controlling plane and anything else you add to the scene, without have to individually adjust the position of each inanimate object.

The next step was to create the 2d text, which was ‘isolation’. I then switched on the 3D for the text and then it moved position to the perspective of the virtual camera I’d just created. I then added a virtual spot light into the scene, which would be used to shine onto the 2D text to cast a shadow.

It was then a case of adjusting the text position to make it look like it was actually at the tube station when I shot this. I also positioned the light in place where there were actual lights; therefore replicating how a shadow would look if there was an object  or person standing there.

I had to motion track the footage so that I could fix the text into the scene. I was having trouble tracking After Effects as the footage was shaky so I turned to Mocha, which is a stand alone Planar Tracking and rotoscoping utility built into After Effects. I was able to track the entire surface of the platform on the station, as opposed to object points. I then exported the tracking data to after effects and applied it to a null object, which I then parented to the Pre-comp of all the attributes of the 2d text. Although it tracked well, I still didn’t like the shakiness of the footage, so I ended up stabilising the whole video to get a smoother playback. Notice in the video example at the top, the original footage is shaky and the final outcome is much more stable.

I do feel the outcome to this technique is more realistic looking compared to the 3D version. As I said before, I wanted to make the text truly represent the title isolation by placing in a remotely empty scene; therefore it looks isolated.


Leave a Comment so far
Leave a comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: