Wednesday, 19 January 2011

Post 28 - Material problem... and solution

Local mountain terrain                                                                    World mountain terrain

One problem that I came across, unrelated to animation is the pre-built scene. A displacement map and atmospheric effect has been placed on the scene to make the terrain covered in snow. During the creation process, it was necessary to increase the size of the environment. This affected the atmospheric settings and resulted in the terrain being simply brown. After a lot of wading through material settings and sheets of script, I discovered that simply changing the material properties assigned to the terrain from 'local' to 'world'.

This did the trick, fixed the atmosphere effect and created a much better scene. Several early videos can be seen to have this brown terrain effect. This was a mistake and has been fixed.

Post 27 - Lever Shot

Here is the next sequence, which shows the main character running towards the lever and pulling it. Attention was made to the change of emotion (facial expression), the hesitation before he pulls the lever and the reaction once he has done so. I wanted to create a sense of desperation and fear.


The first video is the first run through and the second is the final, polished version with atmospherics and trajectory polishing. The positioning of the human character just before the bullet hits him has also been altered.



I have recently been using the trajectory tool, as shown to me by Darren Wall.This is a brilliant tool and essentially draws a red line in between key frames in order to show where objects etc are moving. I shall be using this tool (or option) extensively during the final stages of polishing and tweaking.

Post 26 - Robot Walk tweaked

Here is the robot sequence as a result of improving the key frames, camera position and timing. It depicts the robot walking, aiming and aiming it's rifle. For this video, I have added the atmospherics back in (snow, fog etc).


I feel that this is much smoother and works well as a result of knee anchor manipulation and after following the 'foot guide' which I devised. I tried to pay close attention to the weight of the motions involved in stopping and kneeling down (also subtle movements in the aiming of the rifle).
 
Looking back at the initial research I conducted in creating a walk cycle, I quickly jotted down the following. It is a frame by frame action for the footing of the walk / run cycle (whole cycle of one foot from frame 1-14). It is basic, but variations of the same instruction has been used throughout the animation:

Walk cycle in 14 Frames

First Foot

Frame 1 - DIRECTION POSITION KEYED IN.
 
Frame 2 - Toe roll (40-45% approx).

Frame 7 - Toe roll (0%), foot rotate (35%), raise leg (at which point the waist is also slightly higher).

Frame 12 - foot rotate (-45%), plus positioned so that the heal is touching the ground.

Frame 14 - foot rotate (10%) - DIRECTION POSITION KEYED OUT (waist as far in front before the knee joint of the other leg can no longer be seen to have a influence over the leg. Waist also lowered slightly).

NOTES

- Once the walk cycle has started, Foot B should start its cycle two frames before the end of Foot A. With one foot off the ground before the other, this would hopefully create the sense of haste).

- The waist should move in a forward direction and end as forward it can be, whilst both knee anchors still have influence over the legs (too much in one direction would result in an unrealistic motion as the shin and thigh rarely are the same angle when walking).

Tuesday, 18 January 2011

Post 25 - Initial Test Renders

(First clips and notes)


Here are the first few renders I have created for the process of continual study and improvement.




This clip is an early 'roughed out' first half of the animation. In it, some movements have been blocked in and not animated. This was to test the camera positions and timing. It is of a low quality and the atmospherics have been switched off in order to speed up render times. From this, it became apparent that I need to slow down a lot of my motions. I am working at approximatley 30 frames per second and I am used to 24-25 fps. I will learn to adapt to this speed as altering the speed at this stage will render any key frames out of sync. Also, a lot of the camera motion is a tad clunky and will be improved on once more animations have been created.


After looking at this, I started to animate the robot and created a series of walk cycle test clips. This one in particular was made to focus on the hop, or small jump due to a dip in the terrain.



Then, added some camera motion. This rotates and is intended to rise and show more of the scene at the end of this sequence.



Problems noted include:

- Too many key frames, making the motion jumpy
- Some steps too slow
- Position of knee hierarchy too wide apart (John Wayne syndrome)
- Key frames from left steps to right steps separate and not smooth enough.

These will be fixed and reported on.


I have so far been creating approximately 8-15 clips of each sequence in order to study and progress. The next few posts will include a selected few, showing my development.

Post 24 - Narrative in stills

(Animation sequence in 12 stills)

Here is the roughed out narrative in stills. These are labeled with the same 'shot' names as the previous post. These stills have been taken from the roughed out, early version of the animation. (These can be made full screen by clicking on the images)


At this stage, about 40% of the animation has been roughed out, with the remaining 60% to be established. A lot of the roughed out animation has been created using the timing and pace of the camera I am using in the scene. No that I have got the angles and timing of the camera worked out, my work flow is as follows:

1) Rough out animation, according to where and when the camera in the scene is positioned.
2) Stage one of refining movements and rough playback using the time-line playback function (attention to be made towards facial, secondary limb and joint orientation).
3) Create a short render of the sequence (usually between 5 and 15 seconds long for study).
4) Analyse clip, pick out any flaws and begin stage two of refining / tweaking of key frames / timing.
5) Repeat three and four until I am happy with the outcome (although towards the final stage of polishing and tweaking the animation, several additional stages of tweaking will take place).

Post 23 - Initial Animation: The Narratvie



I have ironed out the basic plot of this animation. 

Shot  One
Essentially the animation starts in the middle of a battle, where a simple last resort objective must be fulfilled. The 'human' character (protagonist) has the objective of reaching a lever labeled 'SELF DESTRUCT' and initiating the process. This is located in the center of the scene and much of the plot revolves around this spot. The camera will pan and rotate, showing the robot (enemy) character on the other side of the war torn environment. The robot will be patrolling the surroundings (possibly in the search for the protagonist). shot one shows both characters, the lever and that they are armed with the same rifle. 

Shot Two
The human runs for the lever. He runs down some steps, where the camera simply pulls out from his facial expression and rotates to see him running for the lever. 

Shot Three
He is spotted by the robot, who then takes several steps, kneels to take aim and fire. 

Shot Four 
A brief shot, featuring the protagonist reaching the lever, pausing for a moment before initiating a 'SELF DESTRUCT' sequence. This will feature several facial expressions from the rushing towards the lever to just before pulling it. 

Shot Five 
Slight pan, then front view of the robot. It fires the gun and the camera follows the bullet across the scene. The bullet hits the human when a slow motion 'death' sequence is shown (hopefully showing an understanding of key framing in relation to timing and movement). This rotates aroun the character until he hits the ground. 

Shot Six 
Shot from a low angle with the dead human character in the foreground and shows the robot in the background throwing it's gun the ground (slight panning to establish the scene. Just as this happens, a projectile falls from sky, hits the planets surface and causes a huge explosion, which engulfs all in the scene. The robot will be shown to acknowledge this with its arms in the air and will be engulfed just before the main character. This will be shown using artificial 'camera shake' and will grow in strength as the explosion wave gets closer to the camera. This is where the scene will end.

In total, the animation will be made up of six short sequences. Two options which will be considered for this are as follows:

1) Uses different cameras for each sequence, render them out in different clips and then edit them in a video/audio editing application. This would allow for several versions of the scene to be saved out and different scenes to be managed in each file. In turn, this would allow for a smaller time-line and a more bite size structure to the work flow.

2) Use the same scene/file and camera and create different sequences by having the end of one sequence at key frame 100 and the beginning of the second sequence at key  frame 101, thus creating the illusion of an edited video. This would mean that a much longer, single time line would need to be worked on, requiring the need to stretch, shrink and scroll the time line.

After looking into both methods, I have decided to opted for option two. Although this means that I will need to navigate across a much longer time-line, I believe that this will make it easier to manage the scene by using one camera, does not require a separate program to edit the clip/s, and means that all movements and assets will be in the same file. This will keep the whole project within 3ds Max.


The next post will show this shot by shot using several screen grab images.

Post 22 - Scene and Costume Construction

Constructed scene (best viewed with the 'full screen' mode)



This is my scene that I shall be using as the backdrop for my animation. As mentioned previously, all elements have been sourced and the buildings have been manipulated / modified in order to create a semi-war torn surrounding. I have also created a simple 'Self Destruct' lever in the middle of the scene, which will drive the narrative. 


Using the skills gained from the previous tasks, I will begin to animate the scene. The scene will have two characters. One will be a regular Dee character, and the other shall be a robot (low res version of the same character). I have created several costumes for the main character, seen in this post, but I have chosen to go with the blue trousers and black t-shirt costume.























So far, I have found the Dee character to be extremely useful, versatile and manipulable. It has also been very easy to change the materials on the character to give it clothing (although I was a fan of the blue default, it looked too much like an alien and I felt a change would be a good decision).