GDC: Day 4 Notes

Day 4 Notes include the sessions, “Boosting Rendering Performance with Automated High Quality 3D Asset Optimization Using Simplygon,” “The 5 Domains of Play: Applying Psychology’s Big 5 Motivation Domains to Games,” “The Art of Diablo 3,” “The Tricks Up Our Sleeves: A Walkthrough of the Latest Techniques Behind FX of Uncharted 3: Drake’s Deception,” “Re-orienting Development and Accelerating Production to Create Infinity Blade and Infinity Blade II.”

Boosting Rendering Performance with Automated High Quality 3D Asset Optimization Using Simplygon

Things that lower frame rate
More drag than hardware allows

draw calls
Geometry on stage
Shading complexity

Manual creation
Many man years
Visible switching

Somplygon LOD

meshLOD less geo per stage

Switching types:
Alpha to coverage-blends pixels two meshes
Geomorph-morphs geometry of two meshes
Traditional switching-Replacesone mesh with another causing visual pops

Geomorph requires 2nd set of coordinates

Material reduces the amount of materials I the scene sharing1 or more across the level

The 5 Domains of Play: Applying Psychology’s Big 5 Motivation Domains to Games

What is the point?
The game design decision model, player motivation model and the translation between.

The big5 or OCEAN
Dozens of researchers
Copyright protected
Several studies some data

Hundreds of thousand of researchers
Open source

Push the big 5 into the 5 domains of play

OCEAN openess to experience, conscientiousness, extraversion, agreement

Opposite of an achievement player is Acontentment player = money onbothsides

Openness to experience

We tend to plays for the same reasons we live
Game design has techniques for targeting most of human motivation
To reach al large audience target Boeing ends of the spectrums

Players want is …. X
Half of them do and the other half want the opposite

Slides on

The Art of Dear Esther – Building an Environment to tell a Story

Engaging voice over narrative
Detailed history and story

Lack visual and audio detail

The concept
Reconnect story to visual
Used environment as an immersion tank

realism vs immersion?

Engine was SOURCE half life
Looked flat
Lacked realism
Not built for large environments

Changedto Impressionistic

Impressionistic qualities
Sense of lit and atmosphere
Evoked emotion

Braking boundaries of realism allowed relative freedom

Realism isnt the holy grail for immersion…

Scenes didn’t elaborateon on the narrative… Talks about ship wreck in bay, but there was no ship wreck in the scene

Static environment can detrimental to immersion
Add wind effects,swaying effects,clouds over, water move

Indirect and symbolic in natur
Representation of images emotions and thoughts
Help build upon surreal aspets

the rt style cn define your esperience to support goals
Not just about. Psychical landscape but an emotional on
All the details are the glue that binds the story

The Art of Diablo 3

Great art is ageless

Key ideas
Stylization over realism
Strong silhouette
Bold use of color
Dynamic animation

Game play first
Making awesome games not just art
Every artistic decision supports the game play


Respecting our heritage:
Stay true to the mood feel and gameplay

Cores question: what makes diablo diablo?

Core features we support:
Fast paced action RPGs
Dark gothic fantasy
Isometric camera
Classic globe UI

Following these principals
Painterly styl
Stylization o er realism
Game is the canvas

supports visual clarity
Allows for stronger setyliton
Visual identapitu that creates s long lasting appeal

Color variety supports feeling of progression in the game

2.5D geometry
Supports the camera
High fidelity look
Low poly costs

every screen has:background, mid ground, foreground

Background has: environment, props, interactives

Lower contraset soft feel
No perfect black or white
Requires clear pathing

Serves as backdrop to mid and fore ground elements

Mid ground: Heroes
Full contrast
Heroes ar lit differently to make them clearly readable

Hero classes: range, melee, caster, barbarian

Mid ground monsters:
Monsters easily seen
Targeting is easy due to highlighting

Mid ground readability
Big shapes first
Irradiance map

Selection outline
Clear and easy to read
Groups with the UI

Skill fx:
Sell the fantasy
Most saturated thing on screen
High graphical impact

user interface framed the composition
Strong contrast and high legibility
Style reinforces identity and gothic themes

The Tricks Up Our Sleeves: A Walkthrough of the Latest Techniques Behind FX of Uncharted 3: Drake’s Deception

Evolution of tools
Fx handwritten in lisp
Shaders hand written in hlsl

Make it artist friendly
Feeds and power
Follow industry trends

Entire pipeline is in maya

Uncharted 2
Evolution of fire:
Static materials
Flip books
Custom shaders the state of the industry

Uncharted 3
Remove programmer dependency
Improv workflow efficiency

Got node based tool for particles

FX Challenges for Uncharted 3
Scrolling uv distortion… Prcedural noise distorts the uvs
Smoke self shadowedbasedoff of scrolling normal maps

Re-orienting Development and Accelerating Production to Create Infinity Blade and Infinity Blade II

The challenge:
Change to mobile from console and PC
ship by end of year…5 months

The situation:
No team member did mobile
Engine not complete

iOS Advantges
Huge, growing install base
Well established distribution
In-app purchases
3d games emerging at this time
Apple was interested in helping
Dev hardware was accessible
Certification speed is fast

R-orient to mobile
what we had to learn
iOS mobile market
The competition
The consumer
Technical limits

What we had to maintain

What to improve apon
Fast decision making
Outsourcing and external pipeline

What to jettison
Rigid adherence to global milestones
Throw out hesitation and doubt

Playing to strengths
3 programmers
2 znimators
1 designer director

Game must be made specifically for device
Not a port or knock off
Touch interface was key

Game to be tailored ego mobile audience
Not hardcore gamers
Intuitive and instantly fun
Sort meaningful game play under 2 minutes
Unique high quality visuals

As accelerating production

Hybrid production style due to rapid dev cycle and small team
Waterfall content production
Iterative design
Scrum style communication

scoped out entire project early
Task estimates based on vertical slice
Time-boxed all features and content
Forced team to think holistically

Simplified schedule and tracking

Maintained tried and true design process
Adapted existing chair design
Team was a invested from day one

Modular features
Proved core gameplay first
Scalable features and prune able content

Scoped for device
Maximum replay for content

Rapid prototyping and iteration
Team was already experienced with unreal engine
Had running within 3weeks
Art style locked down in 4 weeks
Frequent play testing every other day for 10-30 mins

Speeding up the process
Solved technical issues early
Pushed forward – cut features for later releases
Ramped up outsourcing quickly
Smoothed and simplified pipeline
Empowered leads to make key decisions
Kept everyone in the loop

Communication and feedback

Established trust early
Identified key stakeholders
Regular and frequent lead feedback
Aggressively removed roadblocks
Facilitated diet relationships

Towards the end
Weekly swarm sessions meetings between QA, dev, art
Weekly department meet-ups
weekly PR marketing meetings

Daily bug trending
Daily memory performance
3x daily bug triage – I.e. fix now or wait for an update

Had to test all iOS devices each with different versions of iOS

Lessons learned
Took a year to create but was released in 5 months
Design mobile games for mobile customers
In-app purchases critical
Title updates critical to long tail success
Meta-game keeps players coming back. Achievement based keeps people playing
Traditional production practices still apply. If content is not complete push it out
Leverages outsourcing but needs a contingency plan
mobile multiplayer is very different
Top app list is critical to sales
Use customer feedback to refine gameplay

Infinity blade 2
Demonstrate new os capabilities
Bigger and long than original
Deeper story
More variation, less repetition
More robust social component

Used IB1’s analytics for IB2s target hardware… Hardware was the same for both games

For in-app purchases they analyzed what players tended to buy the most
We the Eakins vs enemy difficulty balanced throughout the game…

Shifted focus from how do we do mobile to how can we make it better
Content pipeline and estimates were known quantities
Tech and tools was familiar
Problem areas and features from IB1 were front loaded

Updates vs sequels
Diminishing returns – people stopped asking for updates and started asking for the sequel
Objective milestone acid test
Analytics vital to understanding your customer
Learned the difference between beta vs final tech
The vocal minority damaged popularity of IB during the switch to final release of iOS 5

Lessons learned
Did 5.5 months to release and 6 months of updates
Take advantage of the small agile team in mobile
Design for the device and customer only
Use familiar tech when working on titles on s small timeline
Create content and features in a modular fashion… “kill your babies” meaning remove features in game if can’t hit dates
Structure content for multiple devices
Cut early, cut deep
Understand meta game, what is it really about
Play test play test play test
Use updates and Analytics

Do not assume your personal mobile experience is typical for an average customer… Yu may be hardcore but they probably aren’t
Avoid shoehorning console into mobile
Perfectionism- focus on most critical visible issues
Over adherence to traditional design methods
Overlooking tried and true production practices
Over-compartmentalism of team members… Tam members should be able to do more than their specialty
Putting outsourcing on critical path without a plan b
Treat it as a AAA title


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: