In a new Gamasutra
feature audio postmortem, Rob Bridgett discusses what went right and wrong during the creation of audio for the Swordfish-developed, hyper-stylized hiphop action game
50 Cent: Blood On The Sand.
The third-person action sequel to
50 Cent: Bulletproof was developed with the Unreal Engine, using a combination of third-party and proprietary audio technology. It features voice-overs from Curtis "50 Cent" Jackson himself as well as G-Unit members Tony Yayo, Lloyd Banks, and DJ Whookid, as well as a host of exclusive music tracks and hits, an original in-game hip-hop score written by producer Swizz Beatz.
According to Bridgett, it also required a huge amount of over-the-top cinematic SFX work and a two-week off-site post production mix.
The project's direction was established in early production, as Bridgett explains:
The game's overall direction was also not to take the subject matter too seriously, and to have more fun with the license in order to get away from hip-hop -- which is all too often is unable to make fun of itself. The dialogue, particularly the inclusion of a taunt button, was designed to fit this brief, and to create a feeling of an over-the-top hip-hop arcade experience. It also introduced a lot of fun and humor into the gameplay -- an ingredient often sorely missing from other hip-hop licenses.
The team had a complex dialogue system left over from Swordfish's previous project,
Cold Winter. But it needed a rewrite to meet the design requirements for
Blood on the Sand. To save time, Bridgett recommended replacing it with a new technology developed at Radical for the
Scarface, Crash, and
Prototype projects in Vancouver.
Bridgett elaborates:
The dialogue tool, called UDO (Universal Dialogue Organizer), is a piece of technology originally developed by audio coder Robert Sparks and myself (as well as the audio team at Radical) off the back of the Scarface game.
It is a pipeline-agnostic, stand-alone dialogue database that not only contains all the spoken content for the game in text form, it also allows you to organize, search and flag characters and lines with values for implementation and re-appropriation in the game. It is essentially a dialogue engine and dialogue database in one. It turned out to be fairly simple to integrate this into the Unreal pipeline at Swordfish.
A new audio team member at Swordfish, Andrew Green, took largely sole responsibility for implementing complicated build steps and implementation code for the entire system:
This involved taking the data from UDO and introducing a build step which bundled the data into Unreal engine packages. The loading system also relied on these packages being of small enough size to load and unload as required, so things like dialogue variations were bundled into packages of around five to 10 lines each -- which for things like 50 Cent's taunts, a category consisting of a total of 99 variations, allowed us to load small variation packages which would be unloaded and replaced with new packages once they had been used up.
The total number of in-game dialogue lines was around 10,500, the majority of those belonging to player and co-op characters (50 Cent and G-Unit). Each of the twenty or so foreign enemy characters had around 120 lines each.
You can now
read the full in-depth audio postmortem, with details on process, working with dialogue from hip hop stars, and more (no registration required, please feel free to link to this feature from other websites).
