Where to begin? It’s a huge thrill for me to have an interview with an artist whose work I have found so entertaining as well as inspirational. For the unintiated, Mike Morasky is best known for his work composing Valve’s Left 4 Dead and Portal series. Most recently he has just completed work on Counter-Strike: Global Offensive as well as Team Fortress 2: Mann vs. Machine. In addition to that, Portal 2: Songs to Test By (Collectors Edition) was just released which includes a four-disc set containing music from both Portal games. Please join me below as I get to talk with one of the most thoughtful and interesting game composers in the industry!
OSV: What originally made you decide to make the move from doing visual effects for some very high profile films like Lord of the Rings, The Matrix, and Pirates of the Caribbean to getting into composing for games?
Morasky: In retrospect, I don’t think it was really a binary decision or even necessarily an actual decision at all. I have multiple parallel creative lives going at all times and the main thing that changes is regarding which aspect gets the most attention and/or public exposure. The really dramatic move I made was that of coming to work for Valve. Not only did I find myself surrounded by other “multi-disciplinarians” but I also found myself in an environment where that sort of multi-modal exploration of one’s interests are encouraged. At the time I knew I wanted to dedicate more time to my audio and music work than I had been but I wasn’t quite sure to what degree. After starting at Valve I ended up not only focusing on the disciplines that I’d been wanting to but also in other areas that I hadn’t foreseen at all.
OSV: You have the responsibility of making music that suits a specific world that the designers are creating but does challenging yourself with your own previous work still come into play?
Morasky: Absolutely, but each project is different and therefore requires not only moving beyond previous accomplishments but more importantly moving towards something that is appropriate for the work at hand. One of the great things about the people I work with is that nobody ever seems to be satisfied with what they know or are capable of. Everybody, including myself, is always looking to learn or do something challenging and new. Obviously there’s also a continuum there that one must always be aware of whether it means taking advantage of some particular skill or familiarity that you may already have or avoiding some unfortunate or overused tendency that you may too easily turn to. If you’re speaking specifically about Valve, I am always expanding a set of motifs and techniques belonging to the “Valve Universe.” I sometimes play with them for a subtle sense of continuity or to give the fans paying close attention Easter eggs for their efforts. While I am always trying to challenge myself and improve my work in all regards, it’s also crucially important to keep the goals of the product as well as the resulting effect on the player/audience at the forefront as to not end up too far out on some personal exploration, speaking a language that only you and a small, niche group may understand.
OSV: How often do you get to play games other than your own? Have you played any recently that have served as inspiration?
Morasky: I definitely wish I had time to play more games. I do tend to make a point to play the games that I’ve heard have great audio but I quite purposefully try to absorb the conceptual successes of those games and then forget the specifics if possible. Basically it’s so I can concern myself primarily with what’s right for the project at hand and not with what other projects are doing or have done.
OSV: Did you encounter any challenges incorporating your music into the overall mix of Left 4 Dead 2 given how loud the weapons and action can be? If so, how did you overcome this issue?
Morasky: We spent a lot of time on the first Left 4 Dead coming to the determination that the game actually should be way too loud, which ultimately made mixing a challenge in general. Every time we tried limiting the number of sounds or reducing the volume levels, the game noticeably lost its intensity and people complained. So instead of turning it down we developed a mix layer system similar to animation layering, a complex prioritized music management engine and an AI based “audio director” to massage and manage volume levels and file playback to keep the game just at the edge of complete distortion. The effect was such that the audio feels like the game is meant to, which is just a bit out of control. Of course to do something like this requires a ton of tweaking and time playing the game, to first find the right parameter values and then keep them in tune with the game as it changes with development. There are several techniques at play in L4D, sometimes all at once, to make the music work with the mix and vice versa. Since the chaotic nature of the audio design had been there more or less from the beginning though, it meant that the music design was also conceived of, written and implemented explicitly in that context. This provided a relatively stable framework within which to explore what ultimately became its rather open procedural style.
OSV: I was always kind of curious whether the choral cues for enemies like the Spitter was influenced by the Stargate scene in 2001: A Space Odyssey?
Morasky: When we first sat down to discuss music for Left 4 Dead it seemed that since there were mobs of “infected” in the game that choral pieces would be a great choice for integral representations of certain aspects of both the monsters and their disease. Ligeti, being both a master of 20th century choral writing as well as an essential part of the cinematic vocabulary was indeed very influential. I read a great deal of his work with the goal of absorbing as much conceptual technique as possible and then to apply some of those ideas to the structures that we’d develop in the game.
This early focus on choral music was also how we ended up asking Mike Patton to do the voices for the “common infected.” Many of us are fans of Mike’s and the idea of having him, and potentially other experimental performers like him, as our choir seemed like a great idea. It was only a matter of time before we realized that if he did all of the infected mob voices, we would essentially have a choir of Mike Patton(s). It’s by far one of my favorite aspects of that game.
OSV: When beginning development on Portal 2, what suggestions or adjectives were given as far as what type of mood the music should be? Also, how did the mood change during the course of development, if at all?
Morasky: Initially we didn’t really discuss the music in terms of adjectives as such. At Valve, although we do a great deal of collaborative design work up front, we often tend to look to the work others are doing and let their successes be the “adjectives” that drive where we take our own work. As with many of our games, the absolutely phenomenal artwork and early level designs acted as unspoken guides. There were, however, definitely a few of us who discussed at length what we liked about the music in the first Portal and what types of music we thought might work as reference points going forward. Ultimately though, it really came down to identifying goals through larger conceptual structures overlaying the game as a whole, trying to execute on those structures as made sense in the context of ongoing level work and then judging how well the results succeeded … or didn’t. Many elements of Portal 2 were in flux right up to the very end and so flexibility was crucial to keeping the various musical moods in line with the successes we were identifying while quickly abandoning the failures.
There was an overall musical design of several large movements of “mood” as the story and locations unfold. But there were also musical elements that needed to change during development as they were not working with the other aspects of the game. An apt anecdote to that effect happened early on when I was using some Thomas Newman’esque modes and rhythm patterns, some semi-lite fair. The writers came to me and very respectfully said that although they liked the sketches I’d done, the music seemed dangerously close to “funny”. They requested that I not play to the humor of Wheatley, which I admittedly had been doing, but more to the seriousness of the situation… to the danger. This was early enough in the process that I then shifted the early “natural” environments from “Wheatley’s beautiful, funny, garden home” to something naturally un-organized and threatening… hence the “modes of limited transposition” progressing through minor to diminished, etc. This was also critical in developing the idea for the music to often come from the environment and characters as opposed to commenting on them.
OSV: One description that does come to my mind is one of coldness, similar to what I found with old Metroid games, did the tone of those games enter your mind at any point?
Morasky: I loved those games but they only really consciously came to mind while discussing them with Zoid and our common fondness for Metroid and music of the ilk. So far I’ve tended toward filling my head with non-game related references. Coldness, however, as an adjective, a feeling, was certainly on my mind as the large amount of time and space that had passed with Chell asleep seemed like it would result in a sense of “cold” on many levels.
OSV: Did you write chronologically or did you start at different areas in the timeline of the game?
Morasky: The first thing that was finished enough to work on was the opening sequence, which I must have done a half dozen or more sketches for and didn’t really finish until the game was almost complete. We tend to work out rough coverage as much as possible, first tending to the bigger structures then fine tuning the details once the big motions are done. I followed that same approach as much as possible. The actual final layout and art of the maps and levels were in a pretty constant flux as well. So as the big picture unfolded, I worked with each team to get some representative pieces in the game as quickly as possible then moving on to whatever other areas needed the most attention. I only focused on really finalizing each piece once I had verified the maps were in a pretty stable place. Luckily the game was built relatively chronologically so final work and implementation was done somewhat from start to finish.
OSV: It strikes me that there was a lot of attention paid to the pacing of Portal 2, particularly with the character interaction and the story progression. How did this come into play in regards to choosing where to put music?
Morasky: You’re right. A great deal of effort was put into the pacing of Portal 2 including the music. Describing the process however, is a bit tricky. I guess we were all playing and altering the game every day and constantly working towards creating an overall pacing arc that worked in our daily testing with outside players. That said, of course there were major beats that needed musical under-scoring, which we would identify and address as they became priority. Or sometimes we’d discuss needing some form of musical energy or timing in a given section of the game based on observations of outside play-testers.
The more interesting problem though, was that the constantly changing shape of the game as a whole required a sort of flexible vigilance on the part of everyone involved. Sometimes the changes were small and only required simple, local musical changes. Sometimes the changes were mid-sized structural ones that required some rethinking of how the various movements of the game would play against each other but still fit within my top level, conceptual musical schemes. There were a few however, that were large structural alterations that required I rethink my greater musical architecture. In any case, affecting local trial changes as quickly as possible to determine their effectiveness followed by testing and iteration is almost always the way forward.
OSV: The puzzles are such that it could potentially take people quite a while to complete a more difficult level. How did this factor into the decisions you made regarding choice of music and sound?
Morasky: We were extremely conscious of this and it played a big role not only in the style and composition of the music but also how it was conceptually handled and implemented. In general I always try to extend the basic concept of “lead melody” to whatever element deserves the players/spectators focus at any given time. Be it the dialog, a visual element, sound effects or an actual melody. By leaving room for or “harmonizing” with that “lead element”, I hope to best serve the greater purpose of underscoring not just the (sub) plot or gameplay at hand but ultimately the mental and emotional focus of the audience. With Portal 2 we made an attempt to extend that lead element to the mental process of solving a puzzle that could potentially take quite some time. We used several ideas and techniques to hopefully address observed behavior in the play-testers. We tried to ensure that most pieces of the musical puzzle that play for any length of time were always changing in some way, were related to the game play at hand or were somehow controlled or altered by the player’s actions.
Of course there were exceptions but this was the conceptual framework that we were working from. This contributed to such ideas as parts of the music coming from the machines or virtual locations that players chose to approach or play past. Activating individual machines that each play successive elements of a piece of music that is complete (but alterable) upon finishing the puzzle. Many techniques were used to attempt to keep the music appropriate to the probable puzzle solving at hand.
OSV: Although it doesn’t have to do with composing per se, I have to briefly ask about Stephen Merchant, as I think it’s one of the greatest voice over performances of all time. Do you know much about the recording of it? It seems like to really get the most out of his particular comedic style, a lot of the dialogue would have to be left open for him to improvise.
Morasky: I wasn’t present for the Merchant sessions but from what I understand he was given as much leeway as possible to improvise. That said, the writers worked tirelessly to achieve the story line that we ended up with so obviously there had to be constraints. Having watched the story progress through all the temp dialog and whatnot, I’d say that they really nailed the right proportions of “low frequency” story arcs and humor to “higher frequency” improvised humor, performance and personality.
OSV: When moving through the various decades of Aperture Science, it didn’t seem like there was much of an attempt to match time periods with the music, was that something you had considered and decided against?
Morasky: There were early plans to really play the music off that aspect of the game but the story line that supported it changed. Then there were other plans to hint at period music for each area but actual production time became scarce. Ultimately, there are some hints of it with the quasi-orchestral spheres / astronaut welcome and then the sort of minimalist electric piano progressing into the harsher sort of electronic stuff as you work your way out through time. In that regard though, time as a theme is subordinate to the larger overarching analog->electronic march of progress from natural to artificial, as Cave goes from healthy to sick, Caroline to GlaDOS, etc.
OSV: How was the main vocal melody for the turret opera processed?
Morasky: Bill Van Buren has always done the GlaDOS processing on Ellen’s voice. There’s a recipe he’s developed in Melodyne that just sounds fantastic. That in combination with Ellen’s voice really is GlaDOS, but much of it involves hands on, performance specific manipulation. Time was so incredibly tight though, right at the end that Bill ran his more generic recipes on the different takes, which all reacted to the process in different ways. Then I went through, made selects and edited together the various parts that worked best.
OSV: Now that it’s been over a year since the release of Portal 2, is there anything about the work you did on it that you would change?
Morasky: Well, there never seems to be enough time so there are always things that I’d hoped to do or planned to make “better” but for years now, whenever ongoing updates are not part of the design process, I’ve been in the practice of considering done, DONE. If someday, for some bizarre reason, we were to do the anniversary “infinite effort” edition, at that point I’m sure I would analyze what we’d done and I’d come up with a plan on how to change it. However, the way I’d change it now is sure to be different from how I’d change it in the future so I try pretty hard to not think about it, it could make you crazy.
OSV: You mentioned that that almost all of Portal 2 was recorded ‘in the box,’ whereas Left 4 Dead 2 sounds like it was almost all live musicians. Is this accurate and which method is more fun for you?
Morasky: Indeed it is true. They both have their pluses and minus’. We try to get working material in the game as quickly as possible. As we all know, the artists, developers and even I get what’s known as “temp love”. The material that goes in the game earliest gets the most tweaking done to it and has the longest shelf life so people get very accustomed to it. If it’s electronic and you don’t have to take time away from the game to go to the studio and change it, even if for better, then you don’t have to deal with the inevitable fallout of everyone, including yourself, questioning the changes. The other beauty of electronic work is that the performances are generally all yours. You don’t need to contend with players misunderstanding the intended nuances, etc.
On the other hand, the performances are only yours. I love working with other players who are generally going to be much better performers than I can pretend to be, especially if the writing is intended to be interpreted in some way that only a performer with a live instrument can accomplish. I had my first midi studio in Tokyo in the early ‘80s and have spent a gajillion hours in the studio playing and recording live players so I have a deep appreciation for both. Ideally the choice/preference will be dictated by the design decisions being made for the project. Portal 2 didn’t have to be electronic. For example, if it were a film or 3rd person perspective we’d probably be playing much more to the story of Chell, in which case more emphasis on acoustic instruments, potentially even cinematic orchestra might have been the call. In the case of the L4D(s) the story is intrinsically one of broken humans, so live, largely acoustic instruments, sometimes played sickly, sometimes heroically, was for the most part, the design.
OSV: You had mentioned around the time Portal 2 was released that you had still yet to find the perfect DAW for your needs. Have you come any closer to finding it yet?
Morasky: Well, I really don’t want to be a software whiner or flame any tech companies. I use Nuendo, think it’s the best sounding DAW I’ve ever heard and handles very heavy loads as transparently as one could hope. It’s also got a fantastic mixer that I, as someone who used to design mixing consoles, can really appreciate. On the other hand, I spent a long time working in the visual effects business and the types of software tools available in that craft are far ahead of where I see current audio software. A system focused both on production level quality / stability as well as heavily customizable extensibility is what ultimately what I’m hoping to see. It seems that Reaper is headed in the right direction with their script bindings.
OSV: On a related note – If there were any audio middleware, hardware or software that you could design and have it be an industry standard, what kind of tools would you like out there and what type of purpose would they fulfill?
Morasky: There are really just too many to detail. I guess you’ll just have to stay tuned for this one.
OSV: Thanks a lot for your time, it has been an honor!
Morasky: Thank YOU!
Interviews, Left 4 Dead, Left4Dead, Mike Morasky, Morasky, Portal, Valve