University of Portsmouth - Animation Yr. 3 - Major Project Animation
The post production stage has changed the feel of the movie a lot, in a good way. Initially, after putting my renders together for the first time, I noticed that there were more inconsistencies than I thought there would be (some actions were not matching, the spatial positioning of the characters was slightly off etc.). Because of that, I used quite a few tricks to achieve a believable and readable movie: camera shake, fading transitions, flashes of light and creative sound design.
The sound plays an indispensable role in this film. I have previously mentioned that I decided to use a temp song (Guitar Guy's songs from Mad Max Fury Road) to get my overall feel and timing for the movie. I have collaborated with a student from the sound engineering course (Robert Moisuc), and asked him to use the temp song as reference and create a similar piece for my movie (which worked great).
As for the other sound effects, they have come from different places, like games, sound packs and videos. Because everything was taking place inside a tunnel, I had to modify the sounds in Audacity, add Echo and Reverb and play with the levels. In some instances, I did my own Foley. The voice of the scientist, the character's movement and some other sounds are my own performances, which have then edited to fit the film. For the voice acting, I actually used a real, old, gas mask that I own, which gave me the realistic-sounding breathing and distortion of the speech. The clip has then been taken into Audacity, where I changed the pitch, to make my voice sound more feminine.
The following videos show the first version of the film, without any post production work and the final edit.
University of Portsmouth - Animation Yr. 3 - Major Project Animation
Having 3 characters on the screen, almost at all times, was obviously going to be challenging to animate. In order to simplify my work, I used a lot of cycles. The variation of their movements has been achieved with animation layers.
Many of my references included Youtube videos and tutorials, my own performance in the mirror and reference pose sheets and filmed footage, for the run cycles. The final scenes mostly followed the animatic, with only a few exceptions (extra scenes, different camera angles etc.)
While animating, I realized that my rig wasn't complex enough and because of that it was limiting my animation (the hips could not rotate independently, the palm rotation was not blended with the forearm, the shoulders were quite stiff etc.). I did manage to manually compensate for some of those issues, and counteranimate some of the movements, but they are still far from perfect. I also played around with different camera angles and lighting setups in order to make those mistakes less apparent..
University of Portsmouth - Animation Yr. 3 - Major Project Animation
Since the previous iteration of the film, the backgrounds have suffered the most changes. The action happens inside an implied military research facility. The aesthetics, on the other hand, are still deeply rooted in the previous idea: the concept of a dark environment, only lit by the headlamps and red (warning) lights. The confined, tubular structure of the compound also helped with the storytelling: the movie has become more linear, both spatially and story-wise, making the position of the characters in space clear to the audience, at all times.
The shape of the tunnels and the props inside them have been inspired by some of the following photos from Google:
Some of the main elements that have been taken from these photos are the semicircular shapes of the walls, arched beams, textures, pipes and cabling system and the shape of the entrance of the bunker.
Even from the beginning I planned to make the background from modular pieces. There are two reasons behind this decision: 1) I could reuse most of them as needed 2) Wanted to approach this project with a game artist mindset, as this is one of the fields that I am very interested in at the moment.
In the end, I ended up with only two different backgrounds for the whole movie, which have been modifiedaccordingly for each shot.
Since multiple objects were going to use the same materials, like the walls, ground, metal pipes etc., I realized that I was going to need tileable/seamless textures. In order to achieve them, I tried 2 methods. The first one involved using Mudbox and painting and sculpting on a 'Tiling plane' and then extracting the diffuse and normal maps. The second method involved using the offset filter in Photoshop and then using Crazybump to bake the normal maps. Both methods worked great, and returned similar results; the second method proved to be much faster, though less customizable.
As mentioned in a previous post, some of the scenes make use of smoke. The effect was achieved by using some smoke cards that I have previously made in After Effects. Initially, the plan was to used these smoke cards in 3D space and render them with Arnold. After rendering a few scenes, I decided to not use them anymore, as they were exponentially increasing my render times (note: Looking back, I think the biggest impact was caused by the ray tracing, since the transparency was so complicated. Maybe I could have turned the ray tracing off altogether).
As lighting goes, before deciding to do this the 'hard-way', I tried to make an art test, in which the lights of the head lamps were added in After Effects.
The initial result was promising, but then I realized it only looked good because it was a static image. It didn't take long to rule out this method, considering the range of motion that the characters were going to have and the necessary masking that was going to follow. The manual labor time would have increased, and the results would have also been worse than using 'atmosphere volumes'.
In the beginning I have also used the standard lights in Maya, achieving decent results. At the recommendation of a friend, I researched the Photometric lights and implemented them in my project, towards the end of the production. Since I didn't want to create a discrepancy between the shots, I have added them in places where the impact wouldn't have been that noticeable (like the flashlight inside the storage room).
The following image proves the difference between the lights: (from the left: point light, spot light, Photometric lights). The IES Light profiles have been downloaded from http://www.derekjenson.com/3d-blog/ies-light-profiles
Rendered frames (beauty passes):
The rendered frames include various passes, but which will not be used for the hand-in due to the time limit. I am planning, though, to make heavy use of the Z-Depth to create focal points and the ID pass, to add smoke effects behind the characters.
University of Portsmouth - Animation Yr. 3 - Major Project Animation
CHARACTERS:
The 3D models of the characters are generally based on the concepts that I have previously made.
The first model that I began making was the one of the scientist. I have used a 3D turnaround of a character from H1Z1 as reference for the proportions. I chose to use this one in particular because the character is both realistic and slightly stylized at the same time.
The following images present the modelling progression:
To save some production time, I reused this model as a base for the soldier. Since they belonged to the same faction, it was not a problem if they had some similarities, design-wise. Even though the similarities are obvious (helmet, hazmat suit, gloves), each character can be easily distinguished because of the different silhouettes, and heavily militarized equipment of the soldier.
Both models have been modeled in Maya and textured in Mudbox. Since I didn't plan to use any bump, normal or displacement maps on the characters, I baked in some lighting into the diffuse map.
Initially I was planning to have a lot of performance which made use of the character's faces. Again, I have modeled the head of the scientist, which I later reused to make the one of the soldier.
The rigging process has been fairly straight forward. Both meshes received the same rig, but had to be paint-weighted differently. A BSc student, Mirela, helped me make some customized controls for the fingers. This proved to be a really helpful feature, as I had to do a great deal of posing for the soldiers, who were holding guns.
An easy way to get the desired movement, without having to animate both hands, was to constrain the left hand and the gun, to the right hand. This allowed them to move simultaneously and without any undesired sliding, especially while aiming and running.
Even though the monster is the 'big reveal' of the movie, I decided to make it simple, as it doesn't have a lot of screen time. The design has been influenced by its functionality, by it's power to control humans. Therefore, the monster is basically a mutant brain with tentacles. In many ways, it has been influenced by the 'facehuggers' from Alien and 'headcrabs' from Half Life.
Again, since the monster wasn't going to be a central piece for the movie, I have rigged it hastily, just enough to get the desired motion out of it.
WEAPON:
The weapon that the soldiers are using has been inspired by the Russian-made Saiga-12, automatic shotgun. I chose it to be a shotgun, since the soldiers were going to fight in a close quarters environment.
The design was based on various modernized versions of the platform.
In order to better understand the intricate shapes of the gun, I borrowed an airsoft replica of an AK-74 from a friend, which helped me model the receiver, dust cover and many other parts.
University of Portsmouth - Animation Yr. 3 - Major Project Animation
As the first semester came to an end, and I still didn't have a satisfactory story, I made sure to prioritize my animatic and storyboard. After watching my animatic over and over again, I came to an odd conclusion: the space was too eccentric, and that was affecting my story. The reason behind this statement is simple: I had to explain that that an alien space ship has crashed in the middle of nowhere, broke through the surface of the Earth and ended up in a cave system and the army was there to destroy it. Right off the bat, this was way too much to explain and took too much effort to make clear to the audience. The motivation of the characters to go there in person was also unclear. I believe that these are some of the main reasons why the first attempts were unsuccessful.
But all the work was not in vain. I knew that something was just missing and to make my story clearer, I had to go back to the beginning, to the motivation of the main characters to physically go inside the cave and not just blow it up. One simple reason could have been that there is a hostage that needs to be saved. Stereotypical? Yes. Effective? Also yes. Since the whole movie was going to be short, I did not have the luxury of explaining everything that is happening, so I made a simple 'knight saves princes from dragon' kind of story, but with a twist: the princess is already 'possessed' by the alien.
As for the setting, the cave didn't work anymore, it didn't fit the story. To make it clearer to the audience, I turned it into a research facility - easier to understand as a concept and easier to produce (a few high-tech-looking-location shots should do the trick).
So the story became: -Research 'subject' breaks out and kills everyone.
-The facility is evacuated, but one researcher/scientist is left behind.
-The task force is going to the rescue.
-The task force finds the scientist.
-The scientist has already been taken over by the 'subject' and everything was just a trap to lure the soldiers in.
-The 'subject' attacks the soldiers.
-The lone survivor gets out = alien/soldier (?)
Since I was happy with the preproduction concepts that I had up to that point, I started working on the actual 3D models for the characters, while still working on the story.
I used those work-in-progress models to create a new animatic/previs. This has also helped with generating new ideas, for the setting, lighting and cinematography.
As cinematography goes, I have used some techniques that I found in different places, like framing, high contrasts, various lighting setups, use of shadows and 'shooting through things'.
The last remaining issue was the timing and pacing. To solve it, I resorted a trick that I used before: Instead of creating the pacing myself, I was going to time my shots by following the rhythm of a musical composition. After doing a bit of research, I ended up with two songs to choose from: one that was a bit more action driven (Flamethrower Guitar Guy's suite, from Mad Max https://www.youtube.com/watch?v=MhvPlzelj9I), and another one which was more eerie and unsettling (Area 51 OST from Zero Dark Thirty https://www.youtube.com/watch?v=mFTDNRdjTyE). The downside of this trick was that I could only use this audio as a proxy. I had an ace in the hole though - a friend from the Sound Engineering course, who agreed to produce a similar piece, based on my proxy audio.
These are the resulting animatics:
v4 (long) - Mad Max soundtrack
v4 (short)
v5 (long) - Zero Dark Thirty soundtrack
During a feedback session I have been told that the Mad Max soundtrack works best, so I made up my mind to continue developing that version.
University of Portsmouth - Animation Yr. 3 - Major Project Animation
During the production stage, Mihaela focused on making all of the relevant assets, while I focused on the animation process.
The first animations that I produced were the ink drops, which we have planned to use for some transitions. They were made in After Effects, by overlapping static images, changing their scale and applying a Fractal Noise driven Distortion Map at the end.
To make sure that we are sticking to the set style, I had to modify some of the assets by adding the gritty textures:
The shots themselves went through some modifications as well. Shot progression:
Edit versions:
V1:
V2:
To finish the movie, I had to add sound effects. I wanted them to complement the recording and not to overwhelm the girl's voice, so I chose to use low bass effects and fire whooshes. (make sure to use headphones when you watch the final version!)
University of Portsmouth - Animation Yr. 3 - Major Project Animation
The storyboarding process proved to be quite tricky. While the message in the audio clip was indeed dramatic, we had a hard time coming up with appropriate images to accompany it. We have tried as much as possible to avoid showing exactly what was being narrated, but that put us under more pressure instead, since we had to come up with other symbolic images.
We had decided to both come up with ideas, compare them, and then choose the best options.
My experience from the previous year, in working with ink, has influenced the first few iterations of the storyboard. Because I was confident with my After Effects skills, I wanted to make a movie with mostly silhouettes, double exposure and ink bleeding transitions between the images.
Mihaela's concept pieces:
My animation tests, based on her concept:
Both I and Mihaela participated in the brain storming process, but we decided to split our work load and focus on different things: while I was still working on the sound design, she decided to draw the actual storyboards.
1st version:
Meanwhile, we have decided to focus on the idea of the fire and therefore the animation had two main symbols: the candle and the firefly. Final version:
Since, again, that seemed to be too literal, we tried to take another approach. The answer came after being shown the movie "Boy" (https://youtu.be/SIOUfrlnMx0). While still keeping the silhouetted aspect of it, we have tried to give our animation a more gritty look, which suited the audio.
