It’s always interesting to hear how someone with such a great job
winds up getting there. Could you briefly outline how you went from
touring in indie rock bands to working at Valve?
Although I’ve been playing in bands professionally since I was a teenager, I’ve always had several parallel creative lives going as well. Computers, animation, visual fx, composing, audio engineering, etc., through all of these various pursuits I have come to know some very interesting and talented people and it’s really through them and the odd skill set that I’d assembled that landed me at Valve. That and a very long interview process.
At the Valve/GANG event this summer, I was struck at the emphasis
the audio team placed on the culture and overall vibe at Valve. Could
you describe more about what the work culture is like there, and how
that affects the quality of the Valve games?
There are probably as many answers to that question as there are Valve employees and each project is different from the previous as the challenges and people involved change. The general concept, however, remains the same. No one at Valve has a proper “title” and as such our structure is essentially flat, pushing the decision making for any given aspect of a game as far out to the “leaves” of the structure as possible. This means that the person responsible for executing some aspect of a production will have been directly involved in determining what that responsibility is and how they are going to execute on it. As individuals we also give input and request feedback from the entire team/company and base much of our decision making on observations of daily playtests. We also focus a lot on working in small, inter-changeable, self-organizing groups to keep the overall priorities and goals of the team and product cohesive.
This means that we can be very flexible and adapt quickly to large scale changes to the product as each person impacted will probably have been part of the decision to make such a change and also be in the best position to know what kind of efforts are required to best implement it. It also means that as an artist wraps up one aspect of a project they may transition to another completely different field but on the same project as they are already fully familiar with the ins and outs of the product and the team working on it. It also endows each individual with a huge sense of pride and ownership in the product as a whole which guides their prioritization and contributions as a self-managed member of the team.
There is another interesting aspect to this methodology regarding how the most interesting or challenging ideas survive the process. Contrary to “design by committee” wherein an idea is whittled down to a shadow of its original self, there is a much more organic, natural selection / evolutionary quality to the propagation of ideas in a Cabal at Valve. When some, but certainly not all, of the most interesting ideas are born into the group they can be small seeds of an idea that get picked up, owned, re-owned, re-worked and re-pitched by various people and will ultimately survive if enough of the executing individuals buy in and choose to own the idea sufficiently for it to come into existence. It may be in exactly the same form as it was originally or it may be completely unrecognizable, but it will generally be much more “fit”. Unlike most workplaces, at Valve it’s very satisfying to hear someone else presenting an idea you may have initially come up with (or did you?) as if it’s their own. It just means the idea is surviving the natural selection process. It’s also partially why we don’t have titles and we try to not speak of particular aspects of a game as “ours”.
It sounds like many of the cues for Left 4 Dead 2 were quite short,
or in some cases, short cues simply made of random, really short
loops. First, why was the decision made to have such short bits of
music, and secondly, did you ever feel a bit unsatisfied with that
requirement? I know some composers who feel like their pieces really
don’t even get moving until the 50-60 sec mark, and might find it
frustrating to do 10-20 second cues.
Most of our decisions tend to come down to what works best for the game itself and any goals we’ve established within that framework. Keeping in mind that L4D is a very fast paced game with relatively unpredictable game-play dynamics and with re-playability being a primary goal, short cues can accomplish several things in that context that would be very difficult for longer cues. They can “respond” quickly to any game play developments, they can function singularly or as a layer and they can become symbols of simple atomic game play data, which combined with repeated exposure can become very powerful emotional triggers. For example, depending on the experience level of the player/listener, the “horde call” motif, the most simple 3 note figure, is probably one of the strongest musically emotional moments in the game, especially when used in conjunction with an event as potentially devastating as a Tank spawning.
In L4D there was also a very specific effort made to make the music part of game play itself. The idea being that if you were to turn the music down, you would have a harder time surviving in co-op or competing if playing multiplayer. This effort also drove a lot of decisions about how to best design and integrate the music cues in the game. In the case I think you’re describing, the “horde” music, there are a few pieces of data being conveyed and a couple of goals driving how it was ultimately implemented. Specifically in L4D2, on the most basic level there are 3 tracks playing when a large horde attacks: the drums signify the perceivable existence of an attacking horde, the volume of the regional instrument “collage” represents the danger or proximity of a certain density of that horde and the volume of the synth figure responds to the player successfully causing harm to the horde. By keeping these elements and the data simple, we found that a greater number of people understood and responded to the significance of the music. When we tried to make it much more complex, the symbolic meaning of it was generally lost on them.
Secondly, the horde music happens a lot in the game and as it’s designed to be played many hours a week it needed to have a lot of variation to stay compelling. By using several different patterns of drums and the regional “snippets”, each of which containing many different loops triggering in semi-random patterns, the possible combinatorial patterns are pretty enormous. So on the one hand, the actual “pieces” of music are very small but the overall effect is a very long and varied piece. By sticking with simple, short elements we were also able to keep the symbolic data being conveyed familiar enough to be understood and it ultimately had the added benefit of making the familiarity of the regional acoustic instrumentation somewhat unsettling as the resulting collage is unpredictable, chaotic and somewhat broken or “infected” sounding.
I’m not sure I’d characterize the decision to work with short cues a “requirement” as much as the most effective solution to the challenges that L4D presented us with. Overall, I’d say the most satisfying thing about L4D in terms of a “long form” audio experience, is that all aspects of that experience, sfx, dialog and music, when combined with the greater experience of the game-play can often result in a pretty emotionally varied and compelling experience with many of the smaller pieces coming together to create larger “movements” that are, at times, far more dynamic and effective than we could have created as longer “canned” pieces. Far from unsatisfying, I find it very exciting to relinquish the control of those movements to the emergent experience of the 4 players.
On the other hand, we do also have full length “songs” in L4D2 and a complete fictional band to go with them, that’s pretty fun too.
Could you describe a bit of how you arrived at the instrument
combinations/musical styles for both Left 4 Dead games? At the event
you mentioned that you had tried using traditional orchestral cues
(within Left4Dead) which didn’t seem to work, so what were the ‘bread
crumbs’ that lead you to the styles you would up with?
It’s true, early on, in L4D1, we tried some era 2006 Hollywood style cues but they really just didn’t seem to fit with how the game played at the time. I then developed the main theme, which tells a sort of short musical story and had a sort of 70’s t.v. horror vibe. I derived a hybrid scale from that theme and from there we just tried a series of different ideas based on that scale and feel. We tested each one in-game to determine its efficacy. The Tank cue was the first to test really well and between that and the main theme drove the development of the 60-70’s “B” horror film style, which also seemed to work very nicely with the “living through a horror film” mode of story-telling. In both games, we kept all upbeat modern/electronic instrumentation and cues for positive feedback, ie: when you shoot the horde, make it to a safe room or escape.
The use of regionally based, old-timey, “southern gothic” instrumentation in L4D2 was driven largely by the locations, concept art and the fact that we first developed and demoed the “The Parish” campaign, which being set in New Orleans, everyone thought should have a funeral dirge as a theme. Since the other campaigns, such as “Swamp Fever” clearly shouldn’t have traditional dixie orchestration, we conceived that each location would have its own set of instruments as consistent with that location as seemed vaguely plausible and that all the map/location specific cues as well as the cues associated with the “common infected” of that map would get re-worked to incorporate those instruments. The main L4D2 themes also got re-worked with a campaign/instrumentation specific “dark roots” vibe.
The decision to retain the original cues and style of the core “special infected” from L4D1 was made for a few reasons. Primarily, these characters did not change from map to map and it was the regional aspect of the game that was driving the different sound of each set of cues. Secondly, it would only have been a year since releasing the first game and our players would have grown very accustomed to some of these cues and we wanted to keep some feeling of familiarity to the game. Finally, the sheer multiplicative nature of all those characters times the 5 maps and the expected DLCs also helped to make that call.
If you didn’t cover this in the question about the “vibe at Valve”,
could you describe your typical day routine? Writing music in the
morning? Implementing in the afternoon? Lots of meetings? Late nights?
And if you do have a strict routine, does that help the creative
I do not have a strict routine and tend to work from a prioritized list that can change almost from minute to minute. If I’m spending my days putting out fires, I tend to get my most creative work done after hours as that’s when the interruptions usually stop and my inner clock wants to work anyway. However, early in a project or when working solely on linear material, I can work pretty un-interrupted all day, which is generally how I like to write but doesn’t happen very often. I do tend to do a lot of concepting early on in a project, which usually entails hours at the piano, usually late at night, and then working fairly quickly in my studio to get something in front of people as quickly as possible. If the project requires “proper” studio time, then that’s scheduled, booked and is usually whole days or weeks at a time. I can’t imagine how I ever did it without my iphone as I’m always tracking my work email.
If I’m working on a cue or series of cues that have some dynamic or interactive elements to them, I consider integration intrinsically part of the authoring process and I go back and forth from the game to my DAW. Even though this can be very time consuming I’ve often found that even if a piece mocks up well in the DAW, the only way to know if it’s going to actually work is to try it in the game and that the _feel_ of the gaming experience best dictates what pieces work interactively. I’m always looking for the best way to integrate an idea into the game itself and not just lay it on top and that’s definitely an iterative process.
Speaking generally, do you have any pet peeves about game scores or
game sound? Any aspects of game audio that you’d like to see get more
Being aware of the various constraints that all game artists work under, makes it hard to be too critical of any particular aspect of the process. I’m pretty grateful to be participating in the art of games at what continues to be such an amazing phase in its history. I guess I’d hope that we keep that phase going as long as possible and continue to see a lot of innovative ideas driving development until the well is truly dry, which I don’t foresee happening anytime soon.
Do you see any correlation or parallel between the indie rock/noise/
weirdo bay area band scene and game audio? Do they attract similar
Valve is certainly populated with some of the most interesting and creative people I’ve come across in my life in “the arts”. The demographics definitely skew differently than say, the Mission District of the late 1980’s early 1990’s but the creative energy is similarly powerful and focused. There is a passion for the art and craft of game making that does remind me very much of the energy present during that particular era of independent music. More to the point, I also seem to know quite a few people from that same era and “scene” that eventually made games their home.
Do you have any favorite games (or even movies etc) that you find
inspiring, or that set the bar at a place that you always try to reach
for in your work?
Half Life 2 is always present in my mind while I’m working, as is much of the wisdom Mr. Bailey has been kind enough to share with me over the years. HL2 is sort of an audio Holy Grail for me.
Super Mario Bros. has a deep, important place in my heart, I still love playing that game just to listen to it.
The Matrix continues to speak to me and in many ways, in my opinion, has yet to be replicated.
I always find myself looking to Morricone and Williams as absolute masters of the motif.
What do you love most about your job? The game biz?
Boy that’s a tough one. I think the adventure of each new project is unbelievable. Things don’t feel pigeon holed or pinned down just yet and I’d say there’s still a lot of room to explore.
Would you have any advice for students looking to get into game
audio? in terms of experience/knowledge they should look for?
Equipment/software they should learn? Or other non-game industry
experience that would serve them well?
1) Always be working on your ear, 10,000 hours is the magic number just to start. You can’t fake a good ear.
2) The computer is your tool _and_ your medium, learn all of it. That goes doubly if you’re in audio. Your DAW is a tool, knowing how to use it is a pre-requisite, not a skill. Understanding what’s going on under its hood or how to write scripts to manage large numbers of files, procedural animation concepts, AI, level design … boy the list just goes on and on, the more of it you know the better.
3) Whatever it is you want to do, get out there and do it. Generally speaking, there’s nothing stopping you from working in the field you want, you just might not get paid for it at first. If you do it and do it well though, someone will want to pay you.
Special thanks to Valve Software for the interview http://valvesoftware.com
Interview written/edited by Dren McDonald exclusively for audiogang.org