WEEK 1:
CAMDEN:
Our first class introduced us to the interface of 3DE, we became familiar with the layout of the software.
We used footage of a scene at Camden Locks, in which we had to map out the various layers in the shot as the camera was moving. We firstly used the Track tool to place points in the scene. The method of choosing these areas came down to identifying a high-contrast ‘pattern’ in which the point could be efficiently tracked. A method in which to create a good pattern is to utilise the image controls, for example creating high contrast and saturation to enable a point to track accurately. We did this methodically throughout the scene.
At some points we would track backwards, for example when doing the footpath. This is because it was easier to work backwards when certain views come in and out of shot. We also did this for the background, where the window frames were. This is again because these points come in and out of frame during the video.
Once the clip is successfully tracked throughout, we would select all the tracks and ‘calc all‘. This puts the points into 3D space, which we can analyse through the deviation graph.
Looking at the graph we can spot any erratic points, and therefore fix them. There are multiple ways we could fix this, I could turn off that particular point, move the point in frame or completely re-do the point. Whenever we make a change we must ‘calc all‘ again. The aim is to get the accuracy percentage as high as possible.
WEEK 2:
IRON MAN HELMET: Rotomation and Body Tracking
This week we planned to track footage of a man on the street. We edited the footage, to make it easier to see, and ensured the scene was saved correctly.
Next, we export buffer compression to ensure the clip plays back at full speed and now we can start our track. Firstly, starting with the background, we created two point groups, ‘camera track’ and ‘face track’.
Starting the track, we enabled the image controls to get the correct contrast on our footage, in order to create a sufficient pattern for our tracking points. The camera was stable, so with enough contrast we were able to get a lot done.
Next we tracked the harder parts of the shot, particularly when people are walking across. We just started and ended the track where needed, and adjusted the pattern. When completed, we applied calc all and assessed in the deviation browser, fixing any problem points that we can see. The aim is to get under 1.0 and above 0.0.
In the lens settings we input the correct camera information including lens information and camera constraints, then calculated again. We ‘adjust’ the focal length, in the parameter adjustment tab, using adaptive and calculating all.
We then clicked on the Face Track point group. Adding tracked points where needed, starting at the hairline.
Once all points were tracked on the head, we went through the same steps as before to ensure they were successfully done.
When calculated, we exported our work as a MEL Script, which we then opened in Maya. Also importing a 3D model of an Iron Man helmet and an image plane, with our footage. Then we moved the helmet over the mans head.
Once aligned, we export the helmet in the correct position, then on the Face Track point group we imported the model into 3DE. In the lineup viewport, we selected all our face tracks, then projected the points on the 3D model. Finally, calculating all from scratch.
We ensured that the track was accurate with the deviation browser, turning any erratic points to ‘survey’.
Next we would render out our footage and open it in Nuke.