Audio Content for Diablo and Diablo 2: Tools, Teams and Products

By Matt Uelmen

Diablo and its sequel have established themselves as two of the most commercially successful and influential PC games of the past decade. This presentation will discuss some of the nuts and bolts of one particular aspect of the development of this series: music and sound effects. Relationships in the development world have three facets - business, creative and personal. Despite the obvious fact that none of these three aspects exists in a vacuum, I will attempt to focus on the creative elements of my experience in working on these two titles. Aside from some work as an itinerant musician in my teenage years, my time at Blizzard has been my only real experience in the working world, so I lack the experience to offer much comparison between the environment we have created here against other workplaces. Still, I hope that by focusing on some basic elements of the production of these titles I can help shed some light on whatever "magic formula" it is that has given us our string of #1 titles. I will focus specifically on the tools I used in creating these hits; the individuals who played a great part in helping me get my material in the game, and, most importantly, my relationship with the final product while in the muddy trenches of content creation.



The core of my production arsenal was three tools which I still use today, for better or worse. An Ensoniq ASR-10 keyboard/sampler with 16 megs of RAM was by far the most important tool, and its default library of 3 CDs of samples provided the backbone of my musical pieces as well as a surprising amount of sound effects. I was familiar with the earlier generation of this keyboard, the EPS 16+, which I had spent a great deal of time with in my college days. Though this unit did not necessarily have the best interface or stability of the keyboard/samplers available at the time, the high quality of the default library and resident effects, my familiarity with the model and the wholesale developer discount offered by Ensoniq at the time made this an easy choice. This keyboard was then controlled by a Windows machine running an ancient version of Cakewalk. I have stubbornly continued to use this sequencer despite it being made for Windows 3.0 and have found it to be consistently reliable and containing every bell and whistle I could possibly want. With very few exceptions, the music for Diablo was made by packing up the eight tracks on the ASR-10 with as much as I could get into the 16 megs of memory and then controlling them using the Cakewalk sequencer. Even when doing live material through my $150 AKG microphone, I would generally record it as a sample in the ASR-10 first, and often liberally apply the onboard effects, especially the delay. After making an archival pass through my Sony 59ES DAT machine, the tracks would then go to the third of these tools, also resident on my Windows machine - Sonic Foundry's Sound Forge, an editing program which I have spent a great deal of time staring at in my adult life. This piece of software has proved useful for almost every task I have faced in the editing process, and has frequently proven itself as a great tool for the most basic elements of sound effect and musical sample creation.

The live instrumentation used in the creation of Diablo also deserves a special mention, and I believe it made a great deal of difference in the quality and distinctiveness of the final game. The star of the show was a finger-picked 1994 Seagull acoustic twelve-string, which supplied the main theme for the Tristram shopping experience. The town theme also featured a nasty old Artley flute with a lower foot which would constantly fall off, and a turtle-shaped ocarina which my folks bought for me on one of their trips to Latin America. Also featured was a nice old Slingerland snare drum which was multitracked and panned for the march effect used in the opening theme of the game. A Jackson/Charvel electric six-string was recorded directly into the sampler in almost every tune in the dungeons, with crybaby, mesa/boogie distortion and the onboard effects of the Ensoniq often drenching the signal into something new and strange. This handful of toys was important in my production process, but was more important was what I did not have - no sample libraries, no recording booth, no expensive microphones, no Pro Tools setup, and, most importantly, no preconceptions or rules to cramp the creative decisions the team as a whole often had to make on the spur of the moment. Having less resources can be frustrating, but it can also force you to create something more original than might be found with the latest-and-greatest tools everyone else is using.


The core development group at Condor during the creation of Diablo was roughly a dozen people. Though that size of a group is considerably smaller than is necessary to make a hit title in these days, it is an ideal number for creating an environment where ideas and content can flow with relative speed and ease. In retrospect, being part of that group at that time was a truly unique and special experience which I was very lucky to have - the team was motivated, talented and remarkably free of the politics and ego which can detract from a productive, creative atmosphere. Though composers often have a natural urge to hide themselves away and present their work in completed form, I found that my ability to try tracks in the game and get feedback from people who worked on a variety of tasks was essential in getting the right feel for the game. When I finally had the breakthrough I had been waiting for with the tune that ended up being the theme for the first dungeon under Tristram, I knew instantly that I had stumbled upon the correct formula judging by the response in the office. It is doubtful that I would have had the breakthrough in the first place in an environment which was any less supportive and challenging.

Crucial help from the company which was in the process of acquiring us, Blizzard, also played a great role in making the game what it was. The team in Irvine gave us excellent ideas and feedback in a variety of areas, and was especially helpful in areas in which we had relatively little experience, ranging from technology and design to actual content like the opening cinematic and the unforgettable performances of the actors who populated Tristram and parts below. The experience that Blizzard had in putting together a polished PC title proved absolutely invaluable, and their creativity and perfectionism proved to be a great model for our team in Redwood City. A great part of the success of Diablo was due to the constructive relationship that was established between our two teams at that time, which was remarkably free of the adversarial posturing and suspicion which too often characterizes the developer/publisher relationship and sabotages what might otherwise be successful productions. Glenn Stafford, director of audio in Irvine, deserves particular thanks for establishing a good part of the Diablo sound design universe. Work done by Glenn on monsters such as the Scavenger Demon, Death Knight and the actual player characters not only provided great content which helped create the distinctive atmosphere of the game, but also provided a good model for me in my own development as a sound designer. Glenn's exceptional talent and greater experience gave his creations a degree of polish and personality which I was only beginning to approach at the time.

In December of 1994, I caught the Nethack bug from David Brevik, the man who can claim to be the creator of Diablo with more justification than anyone else. Before working at Condor, it had been since my early high school years that a PC game had obsessed me to the degree that Moria did for the month or so it held its spell over me. Though the game had absolutely nonexistent graphic and sound design, the fundamentals of the play mechanic were largely unchanged in Diablo, with the crucial difference being that we made the transition from step-time to real-time somewhat early in the development cycle. As evidenced by the high scores (under the name of "Diablo") on the copy of Moria which Dave passed along, the game had been an obsession of his for quite some time. For me, the development of Diablo began the day that I asked Dave if it was possible to stick some basic sound effects into Moria. Though the game was built on "open source" code long before such an idea became a common practice, building a sound engine was significantly more work than was feasible with a couple of console titles needing to get out the door. The implications were fairly obvious - we would just have to make our own version of the damn game.

The process of getting work into the game was actually quite fun - the game was almost a perfect balance between models which we had enjoyed in favorite games in the past with raw experimentation. Quite a few misfires in both sound effects and music were stuck in the game before the formula came to where we wanted it. If any general conclusions could be drawn from my experience in developing Diablo, it would be that in-house production was the best strategy to use. Working with an out-of-house contractor is often the best option for many developers for a variety of a reasons, but in this case, I do not believe we could have achieved something with the originality and style of the final product without the attention that you can only give with an in-house relationship.

More than anything else, the creation of Diablo was a victory for the pleasure principle - if it isn't fun, why bother? Though many might find this attitude towards life and work exasperating or immature, I see no other way to create any kind of entertainment product. If you don't have fun creating it, how can you reasonably expect the consumer to enjoy it?

Diablo 2


The success of Diablo made a few things quite clear - the public would want a sequel, and this would be an opportunity to polish and enlarge the world of Diablo which we had originally presented. The low-budget guerilla production tactics of the original game were effective and appropriate for a debut game which hoped to sell 200,000 units at best, but with the title smashing charts worldwide and a sequel positioned for similar numbers there was a need to upgrade our production resources to do justice to our suddenly gigantic and global audience. What was the heart of my Diablo tools, the ASR-10, remained and many of its characteristic sounds, especially the choral and string patches, figured largely in Diablo 2. I knew that just as making the leap from sequencing "blind" on an external box to getting a real visual interface from the PC had helped my composition a few years earlier, so I also had to take the leap from cramming samples into one machine to a real multitrack environment which provided me with visual feedback. My first attempt at this was with a Pro Tools 882 interface running on a hand-me-down Macintosh. Perhaps we should have invested in a newer Macintosh system, as I became much too well acquainted with the cute "bomb" logo in my time attempting to get it consistently working. Regardless, the instability and slowness of the system proved disappointing, and the Macintosh lasted barely a month in my office. Although the Macintosh is something of a standard in the music and video world, I personally have found that the PC standard has been infinitely easier to work with in this environment, largely due to the relative ease of sharing content and technical support when one standard is uniformly used. The two pieces of software that ended up successfully hosting the majority of my soundtrack work were Vegas, the multitrack companion to Sound Forge by Sonic Foundry and Gigasampler by Nemesys. Keeping things completely digital meant keeping things fast, which becomes very addictive. It allows you to concentrate on the fun stuff. Unfortunately, I was too lazy to concentrate on playing with compression on the way in.

During this time, I indulged myself in a nice collection of toys ranging from pedal steel guitar to bass flute, and was consistently happy with the colors they would give me. The game ended up with roughly 80 minutes of music, and I gleefully helped myself to some libraries if the content was good, notably the Spectrasonics' libraries. The asian dulcimers are wonderful, as are the log drums on the Africa library. The tracks where things went most successfully were almost totally based on live source, constructing 8 bar loops out of the work of percussionist Mustafa Waiz, with live Electric Bass and Fender Rhodes piano on top. I listened to a few old Cuban records near the beginning of this period, and definitely cultivated a strange love for the maraca. Most tracks in Diablo 2 were built around a blend of maracas and the human voice whispering or shouting. Favorite sources for this sound were live, using the 808/909 rack emulator, the Ensoniq percussion library, and Spectrasonics' Heart of Africa. My favorite choral voices were from the ASR-10 libraries and the ubiquitous Symphony of Voices. In the sound department, some Lucas source was also used, with the trademark fireball being found in the portal-generation sound. The Diablo 2 skill tree was a nicely sized task for everyone directly involved, and it meant significant thinking through almost one hundred miniature operas. This was some of the toughest stuff in the game to get right, as skills can be used constantly and repeatedly. If something is especially annoying, it can kill the game experience.

Scott Petersen and Jon Stone were as important in the creation of the sound for Diablo 2 as myself, if not much more so, and respectively put in the best performances from a programmer and sound designer that I have ever witnessed while working on an interactive product. Jason Hayes also gave the game his touch with direction of Voice Acting, with help from Tammi Donner. Jason, Glenn Stafford and Tracey Bush also did great work on the cinematics, which were excellent as always. It was really hard to look bad with this particular team, and the anticipation of the title made it more comfortable of a development environment than most groups might experience. The unseen team was important as well - we dug deep into the libraries, and got some additional help from Joseph Lawrence for the 100 or so object sounds in the game. My favorite moments of teamwork in this project were recording the destruction of produce with Scott Petersen and getting our few interactive tunes working with Jon Stone at the end of the project.

Bigger, better, more lush, deeper - those were the orders in creating Diablo 2 and it demanded a systematic approach to content. When staring down a 120 box skill tree and 108 individual levels, the appetite for content is a fierce one, and filling those spaces put a heavy load on many different people, myself included. The breakthroughs which made the game happen were in the organization of content, rather than inspired bursts of the content itself. Matching up sound content to the new "component" based character art system did not always seem glamorous, but it was truly satisfying to listen to some of the final battles. Having the time to record a variety of potential source for "swishies" is exactly the kind of luxury so many products need but never get. Thought Diablo 2 was often more an exercise in perspiration than inspiration, the final product was something I didn't mind putting my name on.

Making a surefire hit with a talented crew should only be a pleasure. When the going got tough I would usually light a little candle under my shrine to Don Simpson, recently departed Hollywood great. If I get frustrated, bored, or simply angry that someone is displeasing the muses, I simply ask myself "what would Don do?" Even though the answer does not always consist of behavior which might be considered healthy in a predictable sense, it generally gets me to the next sound effect and the next tune.

Return to the full version of this article
Copyright © UBM Tech, All rights reserved