Your Music Is the Game Designing PHASE, the Other Game at Harmonix by Chris Foster, Senior Designer
About Me Formerly of Turbine (LOTRO), Impressions Involved in various non-Rock Band initiatives Designed final version of Phase, programmed bits of it
What Is Phase? Inspired by Frequency/Amplitude Play using your own music, on your iPod Tap and glide Single songs and marathons 6-15 people, ~1 year (after external R&D, demo) For more info, see http://www.phasegame.com http://www.phasegame.com
Overview First Prototype Reacting to Apple Visual Style Audio Analysis Core Gameplay Metagame
The Inspiration for Phase After Frequency/Amplitude, how do we make music games successful? If how much you enjoy playing music depends on how much you enjoy the music itself… … then playing your own music is a good idea! One inspiration: Vib Ribbon
“Play your own Music” Can tech make notes as well as people? External R&D: Tech can do it well enough, some of the time. Let’s Prototype it!
What Was Learned Analysis worked for some songs Data needed to be filtered somehow “Building a world” wasn’t central Drum prototype was heating up – Enter Rock Band! Project stopped – team dispersed
Phase on the iPod? Great: The music is there Unique controller It does games! Challenging: Constrained tech Unique controller Fundamentally different audience
The iPod Audience, and What the Game Isn’t Not necessarily gamers Won’t “invest in fun” Listeners first, Players second Complexity is not welcome We entertain them Game sessions aren’t super-long We have some thinking to do… and a brand new team!
New Team, New Idea: The Visualizer iPod couldn’t handle world-building Gameplay wasn’t adding up anyway Personify your music Album Art, genre iconography Your own photos? (Eww…) “Listen first, play second” Listen in Phase, Play from the OS
Visuals and Accessibility Image access broke the visualizer Re-assess: We’d ignored accessibility! We considered people, too Monster Rancher = your music comes to life Wait! Our game is traveling down a road! No, not actually in a car, Chris What’s the style of our journeys? Must be compelling, but technically simple (Note: Hand-drawn Chris Foster originals have been known to increase in value among speculators aged 8 and younger.)
Aaron Stewart “Understandable & personable” became “fun” Set a tone for entire team… including design An external “rally point” that energized us
Raw Data Transients: Gameplay Notes Low/Mid/High – Left/Center/Right Loudness Beats: visual effects (for starters) Estimated tempo track A “confidence” measure that wasn’t. As raw data, this stuff worked well enough… for “normal” songs.
Outlier Songs Songs with long silences Ambient music Super-irregular rhythms Feedback and noise Björk
Quest for Musicality Patterns should get the same gems Verses/Choruses should repeat the same gems When the detected beat breaks or shifts, we should spackle in repairs Emphasize Instruments over Drums over Vocals? Low/Mid/High does this a bit
The Dangers of Smartness Which songs have which problems? Fix one song, make others worse More rules, more bugs Already porting from Matlab to C++ 200 formal test songs vs. “millions” Still flushing out catastrophic failures
Keeping It Simple “More Musical” = Beats x volume Easier difficulties get only most musical notes Still limited: Easy got more notes Breaking up “sameness” of a song Loud/soft sections = denser/sparser notes A little intent went a long way Squeezing out more notes More sensitivity, more errant notes But worth it
… and this is simple. Transients Time Energy (volume) Lanes Beats Time Confidence Transient Proximity to Beat Consistent Beats Volume RMS per Time-slice RMS Volume per Beat Loud and Soft Sections “Spackled” Visual Beats Notes Expert Hard Medium Easy
Taking a Break We still have gaps without notes! Stupid musicians! We can’t just make more notes We need a gameplay solution…
Core Gameplay No guitars, no plate- spinning… no problem!
Re-inventing (for) the Wheel “Gliding” feels good and natural! Can “slaloming” feel musical? Let’s fill in gaps… with sweeps!
A Sweeping Success (sorry) Better than expected! So: short sweeps Musical placement From “rails” to “coins” A sweeping-only game?
A New Input Device First: Tapping Back to Natural Mapping PSP team had creative controls, now largely impossible Without iPods, how do we test controls?
Are We Fun Yet? Songs were getting boring Old techniques were gone No plate-spinning No plastic guitar PSP team cast a wide net for alternate gameplay Too complex for iPod (and our schedule)
Adding Tension Not enough push back Accuracy too narrow “Stars” include streaks, provides broader range of success/failure Mid-song goals Who needs old-style Checkpoints? We do! We show star progress during a song Rock Band stole it!
A Lesson in Simplicity Vs.Vs. Strikes (last known photo) HeartsStars The Problem: Gain one good thing: Stars Gain one bad thing: Strikes Result: Confusion The Solution: Gain one thing: Stars Gain one thing: Stars Lose another thing: Hearts
Speaking of Simplicity… Suggestion: “Choosing difficulty levels is too ‘hardcore’ ” Proposal: Hit a phrase/bubble to survive, hit all gems in bubble to thrive Deciphering “gem bubbles” would have been even harder
The Metagame The Gameplay that Happens Between Songs
Grand Schemes “Reward players for exploring their music” Play every song from artists, artists, albums, playlists Collectibles, pegboards, auto-generated “maps of your music” Easter eggs in random songs, artists etc. Maps in iTunes too? Sure!
Schedule-induced Elegance We have a contract! We have a ship date! It’s cut our schedule in half! (BTW, Campaigns aren’t fun yet)
Back to Basics Classic Arcade Design Building a multi- song arc Forcing the difficulty issue Searching for the round peg
Discarding Grand Ideas Much of the old ideas could have worked There wasn’t time to find out The open-ended game didn’t push back It would have… eventually Lessons: Fun wasn’t central to the design Humble ideas can be better than complex ones
Embrace your platform’s limitations and audience If gameplay is abstract, find a way to personalize it For analysis-driven procedural content, simple approaches reduce risk Be wary of complex solutions, even when schedule allows them Make sure everyone can understand the design “Design is making bad decisions in pursuit of good ones.” You have to be willing to be wrong to find what’s right!
Q&A Thanks for your time! Special Thanks: Dan Teasdale, Rob Kay, Kasson Crooker, Josh Randall, and the Phase Team