There is a reason for me not posting every day. That summary i said i was doing so many days ago? It's been taking longer than expected. Who knew my thoughts could be so scrambled. So it's ended up that I've been using that still incomplete summary to create my prezi for Crit week. I'm going to give a write up of everything i talk about/ want to talk about right now so if you're in my crit group and don't want the advanced knowledge of what I'm going to talk about this is your
SPOILER ALERT.
Slide 1:
As i'm in a group with people who didn't see my pitch i thought I'd have a quick slide which briefly sums it up. It shows that by the end of pitch week I'd pretty much managed to narrow down a topic to look into deeper but that was about it.
Slide 2:
This is a video of my initial research into GTA4 NPC animations, i chose GTA4 over the others I'd looked into as it actually gave you the opportunity to interact with NPCs outside of cut scenes that limit your options.
I spent time pushing NPC's from different directions and at different speeds, i pointed a gun at groups of people, i carried a gun around and walked around by a police officer to see if he or any others noticed and i spent some time stalking one NPC to see if he reacted in a believable manner.
Slide 3:
After watching the footage back and taking notes these were the four things that kept reoccurring and so i decided to create an animation that addressed those issues. I decided that i would create an animation of a gun being pointed at a group of NPCs and suggest what their reactions might be.
Slide 4:
I didn't want to assume any reactions so i started by doing a bit of research into defensive body language, from there i could make storyboards which reflected the findings. The top storyboard was a more generic reaction where one NPC reacts and the other notices the firsts reaction and then reacts them self so there is a slight delay. The second adds more personality where one of the NPCs might try to protect the other due to a close relationship between them or maybe the one NPC has a sort of heroic disposition. I chose the top more generic situation as i thought the personalities added a level of complexity that wasn't needed in this initial test. I also took out the third NPC as i didn't think it added anything extra that the two already showed. My main observation from this test was that yes it was more realistic but because i hadn't researched into existing limitations i didn't know how useful a suggestion it really was.
Slide 5:
Looking into Limitations i found they tended to cluster together so in an effort to make my findings easier to explain i decided to use an example.
Slide 6:
The limitations i found generally boiled down to the same things Time, Money and CPU power. I decided to start with Time and Money and come back to CPU power.
So the example i decided to use was a game called MechWarrior i found a reasonably detailed account of their animation process which i found quite useful when it came to researching limitations. When planning the animations for the game they discovered that they would need 152 animations for each of the 20 mechs the game required. This is clearly a massive amount of animating which they didnt have the time or money to create individually.
Slide 7:
This would lead to a Limited library of animations but as the project couldn't really sacrifice any animations due to the visual style they wanted to achieve they decided to go down another route.
Slide 8:
They managed to cut down the amount of animations to 30 vital ones and then blended them in varying combinations to achieve the full 152. This of course shortened the time it took to create the animations initially but blending has it's own problems.
Slide 9:
Using blending meant that the animators had less control over the performance Whereas maths can easily bring a characters leg to the right position. The computer doesn't understand how to give the movement a sense of weight or timing that a skilled animators expertise can. This boils down to the revelation that earlier pioneers of animation found when forming the exaggeration principle and is nicely summed up by the bottom quote "Many times, looking real is not actually being real" and can leave the movements looking "unnatural and clunky"
Slide 10:
This meant that the Animators then had to check all 2440 animations that they hadn't created themselves and due to the sheer volume by the time they discovered a problem with the landing animations they had pretty much ran out of time and the game ended up being shipped with the problem still intact in all 20 Mechs.
Slide 11:
Which leads us nicely back to the whole Time and Money problem. They had managed to produce all the animations in time but due to time they couldn't fix all of the problems they encountered.
Slide 12:
Moving on to CPU power i found a general opinion that if the game is causing too much strain on the CPU power available, the animations are one of the first things to go. This seemed to be because generally game play was thought of as being more important. As with before this leads to the possibility of a smaller library of animation.
Slide 13:
A smaller library can result in repetitive motions. This problem seemed more relevant to my project as at it's core i'm trying to find a way to make NPCs into more believable characters.
For now i'm choosing to ignore these limitations in regards to my research as i think both these problems can be solved by technological advances. The blending resulting in stiffness could possibly be resolved by giving the animator more control over how the animations blend together and the CPU power will no doubt become less of an issue as computing advances over time.
Slide 14:
Continuing with the idea that CPU limitations can lead to the limiting of animations. I attended a talk that Lynn organised between Sean Laverty who works at Blizzard and Fraser Maclean who i know off hand as the author of setting the scene. During the talk Fraser made a comment which i could have misinterpreted but it sounded like he thought characters weren't very important in games. I didn't agree with this at all as in my opinion story driven games at least seem to be heading towards being more like playable films where characters play a vital role in story telling. It seemed that Robin Sloan's Journal shared a similar view but i wanted to have more of a reason why so i dug a little deeper and found the second quote referencing the idea that NPCs actions can help define the players opinion of their character.
Slide 15:
I thought this was an interesting concept and formed my own opinion that the NPCs animations should reflect both their own personality and that of the player they are interacting with. The second could be split into two catagories though and i thought the best example would be to combine the interactions from skyrim and fable. In fable the NPCs react to you as an individual based on your actions and reputation within the world, they recognise you on sight and the more extreme your actions the more extreme their reactions. In skyrim i found the NPCs reacted to me based on my race. I played as a Kahjit who have the reputation of being theives, sneaky and deceitful and the comments they made reflected that however they never reacted to me as an indivudual. The blacksmith in Whiterun who i interacted with an inordinate amount of times never changed their reaction towards me. I would have thought it would be better if depending on my actions around them and how much i saw them they grew warmer or colder towards me.
Slide 16:
This is what i'm doing now. I'm developing a test using the Likert Scale on the reaction to initially seeing the player. I'm looking further into body language to help dictate the reactions that might occur. Then i'll give the resulting animations to a selection of people in a random order and ask them to place them on the Likert scale as to how warmly they feel the character reacts. This should hopefully prove that the animations give them an impression of the characters opinions.
Slide 17:
Depending on the results i will either go back to the drawing board with body language or choose a game character and an NPC study their personalities and traits and devise some sort of animation tree which takes into account their own personality their previous stereotypes of the character and the past possible experiences of the player. I'll have to figure out which of the three is the most important somehow.
Thats the slides i have and hopefully what i'll be saying about them, just so my work isn't lost for all time as i think theres little things i might want to keep in there here is the post i haven't posted yet thats not complete at all...
------------------------------------------------------------------------------------------------------------
Summary
I think i need to summarise what I've actually found and the direction I've been heading in the past couple of weeks since pitch. So kind of make a between pitch pitch, to myself. Because I'm getting confused and with so much information floating around i think i keep going off topic.
What am i looking at? I think i need to look at my aims and objects more when researching. I'm getting away from what i need to be doing with some of my research
Objective: NPCs Role within Story Driven games.
http://vgstorydesign.com/solutions.php
This looks at the role of NPCs and character within story driven games, it argues that its the levels and story elements involving the protagonist changing or growing in some way which are the most memorable and dramatically the best for the game.
http://rpgmaker.net/forums/topics/11519/
"NPCs are indeed crucial to the story as not every character presented can be a playable character ""Pivotal (NPCS) - those constituting the supporting cast of a video game - are essential to the telling of a game story. Particularly in genres that focus on story and characters" (Sloan, 2011)
"Then the article speaks about methods of portraying personality of the main character within the game and one in particular i thought applied to NPC personality. They speak about predefined action as a method of retaining consistency of personality and the example they give was the way characters (NPCs) react towards the protagonist. An NPCs reactions to the protagonists actions or presence will give the player some idea of the protagonist's personality. They use ICO as an example of this, Yorda's reactions to ICO within the game are trusting and obedient which will help to establish ICO's nature as friendly and caring." (My observations of an article on the importance of character)
Note, however, that the player will be affected by how NPCs react to the main character. Unless you want to lower a player's opinion of the main character, NPCs should generally react positively towards it. (Toby Guard, 2000) <-- backing up ICO statement above link to original article below
http://www.gamasutra.com/features/20000720/gard_01.htm
"I then picked out a reason why i thought skyrim had done better than a lot of previous games because when playing as a Kajit some of the NPCs in the world had almost a racist attitude towards me, and comments and reactions fitted with that, but what they didn't do is react to my character as an individual which is where it fell down a little. From this i thought perhaps a mix of skyrims reactions to the stereotype of my character and fables reaction to me as an individual would be the best way to approach a new system. So if you spoke to a character in skyrim often enough they grew either fonder or more aggressive towards you based on your previous actions." (My Final thought)
Objective: Research what technical limitations exist in game animation.
Problem"The high state of rendering fidelity exposes low-animation quality much more than on previous consoles. Characters look great in screenshots, but they can seem unnatural, clunky, and even robotic when in motion."
(Reil, 2006)
Solution
"The way to solve this, and to get in-game animation to the quality that the raw material already supports, is to provide animators with intuitive tools to perform the tasks currently carried out by animation programmers."
(Reil, 2006)Problem
"After playing any game for a little while, users will notice that the animations always look the same. " (Reil, 2006)
Solution
"Rather than pre-producing every clip beforehand, we can use the console's processing power to synthesize animation on the fly, as the game unfolds. This, of course, is not straightforward, as it requires a full (and real-time) simulation of the 3D character (including body, muscles, and motor/nervous system), but it is now possible on the Xbox 360 and PS3."
(reil, 2006)
Problem
Key frame methods can create repetitive movements that break the suspension of disbelief
Solution?
New method of animation called "Mesh Skinning"
"The mesh skinning technique has several distinct advantages compared to the traditional computer game animation techniques such as key framing and rigid articulated body animation. It can produce realistic and smooth character animation in real time by allowing more than one transformation matrix to affect the vertices that form the "skin" of the character. With advances in both hardware and software, it has now become more practical to implement this technique in PCs in real time." (Zhigang Wen, Quasim Mehdi, Norman E Gough, 2002)
Objective: Develop my understanding of human interactions"Making a character aware of its environment has an incredible impact on its believability. If your character examines its surroundings and the other characters in it, it automatically appears to be thinking about what it's looking at." (Toby Guard, 2000) (reffering to awareness in the illusion of life)"Besides just awareness of presence, emotional responses by NPCs toward the main character add immeasurably to its substance and believability" (Toby Guard, 2000)
Practise using the book Force : Life drawing for animators i've been practising adding more movement to my drawings
Dishonored Note: There is a part when you reveal your identity to a character who knew you at the beginning of the game. His animations should really reflect that but they don't.
Dishonored Note: There are times when civilians will wait till you have your back turned then run for help. Nice little bit of self preservation.
