The Call of Duty series is perhaps the most widely acclaimed WW2 franchise in video games. Developer Amaze Entertainment recently partnered with Activision to bring the series to the Sony Playstation Portable system and provide players-on-the-go with a taste of first person, WW2 combat. Several recent reviews cite the audio, in particular, as one of the strongest aspects of the title. One reviewer said, “..this game sounds every bit as good as Call of Duty 3.”
CDM recently had the opportunity to speak with some of the audio designers for Call of Duty: Roads to Victory for PSP. Mark Yeend is the Audio and Music Manager for Amaze Entertainment, Drew Cady was the lead sound designer, and composer Noel Gabriel, whose scores for Amaze have been recently compared to (confused with?) Hans Zimmer.
(Disclosure: Though I did not work on this title, these guys are my coworkers.)
Create Digital Music: Can you please familiarize our readers with the platform specifics of the PSP?
Mark Yeend: In terms of making games, it’s a lot like the PS2. You’ve got resident per-level-load RAM for sounds, and you also stream data off the UMD disc during runtime, which can be used for a music stream or for buffering sound data. The PSP speakers are bright and loud, and the steaming music format is Sony’s ATRAC3 codec, which sounds fantastic. Aside from gaming, the PSP can store, play and manage your music library, link to your PC, and play movies on the go (which look and sound great). I have really been impressed working on the PSP for the last few years.
Drew Cady: 32 channels of audio can be played at the same time, and believe me we use all of them in this game. Sampling rates can be relatively high, but at the cost of memory. The onboard speakers roll off below 500hz (-33db at this point), but I still try to keep the low end frequencies for headphone listeners.
CDM: Can you describe for our readers some of the primary constraints and limitations you had to work with in making this particular title for the PSP?
DC: RAM is the biggest constraint. The first thing I do on a project is lock down how much memory I will receive for audio. This helps me scale each section of audio. For example – level ambience, level effects, weapons, player character, non-player characters, global sounds, dialog and vehicles. For Call of Duty, the audio memory budget for each level was around 2.8MB. To help juggle memory I can downsample any file independently. Distant sounds can use a lower sample rate of 17k or 14k while first person gunshots are sampled higher at 28k. After creating a sound, I usually listen to it downsampled. For instance, a chain sound effect needs to be tweaked so they can be heard in game. If you downsample to 22k much of the high frequency detail is lost. If you raise the sample rate to 28k the high frequencies are not cut off and you can hear the jingle. Alternatively, you can also lower the pitch by two half steps and keep the sample rate at 22k. It takes a bit of experimentation to really find the sweet spot.
MY: As Drew noted, the two biggest limitations were the frequency response of the PSP speakers, and the audio RAM budget in each level. How do you create an earth-shattering explosion or the rumble of a Tiger tank on quarter-inch speakers? You can’t. We’re counting on people using headphones, but the deep sounds have to be audible without headphones too, because they often affect gameplay. For example, they’ll let you know when and from where the enemy tanks are coming. Our typical solution was to add a bit of mid-frequency content to the deep sounds; just enough to make them present on speakers without changing the character and realism of the sound. For RAM, despite our producers and programmers acknowledging how critical audio is for Call of Duty, we ended up with about 60% of the RAM originally allocated for audio. It was a really tight squeeze, and it’s a symptom of art and geometry and code taking up space to make a strong, robust game. This is just a reality of game development — setting team priorities and negotiating your RAM budget. We found some creative solutions in sharing sounds and using custom downsampling for just about every sound in the game, balancing size versus quality. Even though I’m an audio guy, I’m a gamer too, so I would rather play a fun game that sounds OK than a lame game that sounds great. Fortunately, I think this team did both.
CDM: What were some of the biggest challenges in working on this title in particular?
MY: The big challenge was re-creating the chaos and epic scale of WWII. We had a limited number of sounds loaded at any given time, and we had to get very creative about how we used and re-used the loaded sounds. Drew was especially innovative here. He made an elaborate template for ambient sound beds that stitches together all the gunshots and explosions we have loaded, playing them in random sequences, random volumes, and random pitches, so you never hear the same thing twice. He also made versions of this for inside and outside, so the ambience will dynamically change when the player goes in or out of a bunker. Additionally, he made varying intensity levels for these ambience templates, from which the level designers could choose, using just a number in a text file. It was a lot of work, but the results are realistic and exciting, and ultimately help make it feel like you’re in the war.
DC: The console and PC versions of Call of Duty have many non-player characters that add gunshots and explosions. These games can also use long, pre-mixed loops that stream from disk, and in most cases have larger memory budgets. For PSP, we were limited to 6 players on screen at one time, so I could not rely on a large sounding war from realtime game events. Instead, I had to create a dynamic ambient track using pre-existing and pre-loaded sounds. I did this by creating a preset engine that controlled a wide range of single shot and looped sounds and varied many parameters in real time. Several explosions, guns, rumble and vehicles are orchestrated on the fly. I can change how the sounds are triggered by changing trigger speed, probability, length, pan, volume and pitch. There are four presets per level (low, medium, high, mute) and they are selected according to the situation.
CDM: The Call of Duty series is known for its accurate representations of weapons and vehicles. Can you tell our readers a bit about how you started out with this project, in terms of acquiring your source material and defining the palette you’d use for sound design and music?
MY: For starters, we inherited a whole bunch of the awesome weapon and vehicle sounds that Activision used in previous Call of Duty games. That was a crucial starting point for us. These are real recordings of real weapons and vehicles we’re using. We designed the sound behavior of tanks and jeeps by working closely with developers, but honestly, vehicles are not a huge part of our game design, so the files were super small and re-used often. We bought a couple of CD libraries, but we really didn’t use them all that much. We did original recording for Foley sounds, which Ian* and Drew scripted for hundreds of animations. We also created some special effects from scratch, like tinnitus or heart beat when you have low health. *Sound Designer Ian Rodia also worked on Call of Duty: Roads to Victory, but was unavailable for this interview.
CDM: Were there any constraints on sound design, in terms of whether or not your title had to meet certain specifications or requirements to match the console titles?
MY: I played all of Call of Duty and Call of Duty 2 on the PC early in production, and I encouraged all the sound team to do the same. Besides what I’ve already mentioned, the constraints were just that the franchise set an incredibly high bar for us, which is a great problem to have. A great aesthetic precedent was a really pleasant constraint.
DC: Having the original sounds made the process smoother since they were already approved by the publisher. Whenever possible, I tried to stay true to the original games and source material, and Activision was pleased with what we created. In terms of authenticity, there were only a few issues popped up. For example, the firing of the M1 Garand, which needed a ‘ping’ when the gun ran out of bullets.
CDM: Were there any particular tools or applications that you found particularly useful in this title?
DC: I am a big fan of impulsing and convolution reverb. Impulsing allowed me to create convincing environments. Digital reverbs can be too perfect and since Call of Duty is in a real environment, impulses tend to work better. SIR (Super Impulse Reverb) is a handy impulse reverb plugin that allows you to slow down the impulse to exaggerate the reverb. When this happens the delays or slaps become more pronounced.
MY: We’re using proprietary tools for implemeting our audio. There’s a custom Lua script for dynamic music triggers and custom Lua scripts for animation sounds. Aside from those tools, being on good terms with your programmers is always a useful tool! And though it’s not sexy, we created some useful, custom Excel spreadsheets to track level memory budgets.
CDM: What level of communication and interaction do you have with the programmers and designers on a title like this?
DC: There are many audio-specific hurdles to get over for a game like this. The producer understood at the start of the project that audio was going to need more attention. As we created the game we discussed our options. I think it is the sound designer’s responsibility to push the envelope and aim high. Once we understood what we needed, the programmers were very helpful in creating the tools. You always need to leave time in the process for revisions – especially when you’re creating new toolsets. Fortunately we were able to revise and tune the tools with the programmers.
MY: At least one person from the audio team was in every team meeting from early on in pre-production, all the way through the project. We evangelized for robust audio in Call of Duty: Roads to Victory at every turn, and found that basically we were preaching to the choir. Everyone knew that the sound design would sell the realism, and the music and voice would sell the emotion. All the developers were extremely helpful, and the producers gave us the resources and support we needed to make it great. Sometimes we needed custom code hooks to support a special case, and we continually pressed for more memory in levels, but it never got really ugly with the programmers.
CDM: In composing the score for Call of Duty: Roads to Victory, were you working in any particular musical idiom, or did you have to follow the particular feel and flavor of previous titles?
Noel Gabriel: I suppose WWII music has become a musical idiom of sorts, with so many games and movies making use of the subject matter. I’d say the classic WWII score is derived from the music of Jerry Goldsmith (“Patton”, “MacArthur”) and Elmer Bernstein (“The Great Escape”, “Bridge at Remagen”) and you hear this influence in much of the score for Call of Duty 3. Without a doubt the most influential score in recent years was written by John Williams for “Saving Private Ryan”. Call of Duty: Roads to Victory pays homage to this in the music for the American campaign with its use of brass chorale and contrapuntal trumpet passages.
CDM: When you’re composing for a game such as Call of Duty: Roads to Victory, can you tell us a bit about how you ‘design’ the music, in terms of creating musical passages, transitions and segments?
NG: We didn’t rely too heavily on our dynamic music system for Call of Duty: Roads to Victory, but in terms of ensuring pieces joining together nicely, I made sure they all started in the same tonality. And I use the term tonality instead of key, because many of the combat pieces were chromatic enough that you couldn’t say they were in any set key.
CDM: What were some of your biggest challenges in composing the music for Call of Duty: Roads to Victory?
NG: Living up to the precedent established by Michael Giacchino and his use of an 80 piece orchestra recorded in a scoring stage in LA was daunting to say the least. But whether you have the luxury of premier players or have to create an entirely midi based score, the important part is to create an emotional connection between the game and the player. Some of the complex emotions evoked by the music (such as battle fatigue or a sense of loss over fallen comrades) were especially challenging with a midi-based orchestra. But I’m pleased with how it turned out. There was one point where I said “I’ve got to have a real trumpet player” to make the chorale I was writing work. Fortunately I was able to get approval to do so. The majority of the music is midi-based, however.
CDM: What tools do you use for composing?
NG: The sample libraries used for the music were a combination of East West Quantum Leap Symphonic Orchestra, East West Quantum Leap Symphonic Choir, and the Vienna Symphonic Library. The solo trumpet work was performed by Pat Murray.
MY: Noel really knocked it out of the park with his Call of Duty: Roads to Victory PSP score. And while I’m not quite the orchestral maestro like Noel, I do compose for our other games. I’m using Cubase SX with Gigastudio 3, Pluggo, some old E-MU hardware synths, and NI’s Battery, Reaktor and Absynth. My tangible instruments include an Epiphone electric guitar, an Alvarez acoustic guitar, Yamaha drums and dozens of percussion toys. When I write for GBA I use ModPlugTracker, and for DS we’re using Nintendo’s Nitro Composer. At home I’ve been using a ProTools LE rig for a few years now.
CDM: What’s next for the Amaze team?
MY: We are looking forward to some games for the XBox360 and the Wii, and I hope we make more games for the PSP. It’s a strong platform.