January 17, 2006


by Chris Randall

Speaking as someone who makes a not-insignificant percentage of his yearly income providing music for console games, I read this article, The Rise and Fall of Game Audio by Matt Barton, with interest. I'll save you the trouble of reading the first section, which is painfully ill-informed, by saying the following: skip the first section. Don't believe me? Well, don't say I didn't warn you.

Anyways, the author goes on to list reasons his opinion is more credible, because he listens to MODs or something, and then basically provides a lengthy argument as to why game music should sound like Jean Michel Jarre. While I strongly disagree with many of Barton's points (which seem to be, for the most part, paraphrasing of George Sanger's omnipresent arguements viz. game music), the article, in and of itself, makes some interesting assertions.

Let me just say this, though. When the Fatman was making his early hits, music in games was an afterthought, and relegated to the "oh, shit, we gotta get someone to do this" file, so his opinion is more valid in that context. Now, with games costing as much or more to make (and grossing far more) than most Hollywood feature films, the composer is beholden to the director. He makes what he's paid to make, and that's something that's almost entirely (but not completely) like whatever the director put in the temp, just like in movies. Someone should inform Mr. Barton of this simple fact.




Jan.17.2006 @ 8:46 AM
That was one long-ass article, which you can be sure I merely skimmed. My ill-informed summary is as follows: "Chiptune: Good! New game music: Bad!"

Anywho, I was shocked to find this comment from the author below the article:

"Actually, one of the main complaints I received about the article was that I didn't take enough time to research the great electronic music still being produced. Unfortunately, that was precisely my point--such music, as good as it may be, is NOT well-known to the masses and requires some effort to hunt down. Out of the bands you mentioned, the only one that I have heard before is Aphex Twin, who did a marvelous Powerpill Pacman song. I'll gladly check out these bands and hopefully help spread the word; perhaps it really is up to people like you, Sonance, to help introduce new listeners to great new electronic music."

And this was in reference to electronic acts such as Autechre, Boards of Canada, Squarepusher, Coldcut, Kid Koala, etc..! I listen to very little electronic music, but I am familiar with most of these artists, at the very least by name. These are not unknown artists as the author seems to imply. These are artists that get reviews in Rolling Stone. Sure, these bands might not be recognizable to the average Joe, but I doubt the average Joe could recognize a Devo song other than "Whip It!" The author's argument (if it's not well-known and requires research, then it's not relevant) is simply silly.

Quick question: What is "the temp" to which you refer?


Jan.17.2006 @ 8:56 AM
Chris Randall
When they do the live-action or cut-scene portions of a game, like in a movie, they use what we call a "temp score." The director will drop in pieces of music he thinks are appropriate, usually score material from other movies, and occasionally popular songs. Then there occurs something called the "blocking session" or simply "blocking," where the composer, the music director, and the director sit down together and go over the scenes, and the director tells the composer "make it like this."

The difficult thing about doing music when there is a temp score is that the director can get really hung up on a particular piece, and it's the composer's job to basically recreate that piece without actually kiting it, which isn't as easy as it sounds.



Jan.17.2006 @ 10:28 AM
Hmm... this is an odd google ad:

Gene Machines PolyPlex 2: DNA/RNA synthesis at less than $0.10/base with the PolyPlex 2.

Anyways, I can't really add much to what's basically been said about the article. The author does have a point -- some fantastic music has come out of old videogames and the mod/chiptune scene is vastly underappreciated.

However, he doesn't really realize that games have gotten more sophisticated. A chiptune, midi or MOD score would sound terrible in a game today, especially when put up against an entire symphony orchestra, or the releases of a veteran producer. Even Jesper Kyd moved from being a C64 musician to using the Budapest Symphony Orchestra and Choir. As soon as the technology was availible to stop using midi and chiptunes for games, everyone did. When you're given the choice between the Final Fantasy music produced by the SNES sound chip with 64k of memory, and the version played by a real 72-piece symphony orchestra, it's pretty obvious that you're going to want the latter.

The era for chiptunes is over, and while it's nice music, I don't really find it evocative, nor do I think the general public will. Even bands that use chiptune sounds in thier music are pretty much relegated to the kitschy end of things.

Anyways, I think the next big thing in game scores is going to be more collaborations between established artists and game designers, like with what Amon Tobin did for the Splinter Cell games.


Jan.17.2006 @ 11:17 AM
Chris Randall
"Anyways, I think the next big thing in game scores is going to be more collaborations between established artists and game designers"

I agree. I also think that we'll see a lot more conditional composition, where the musician sets up the initial conditions, and an aleatoric engine actually provides the music based upon game event-driven action. Now that consoles have so much horsepower, the composer doesn't have to go begging for CPU cycles like back in the day, and an aleatoric engine is a reasonable alternative to just firing samples at appropriate times.

This really appeals to my sense of asthetics viz. composing for a game; I'd love to collaborate with a game company doing an aleatoric piece, because it's more sound design and timing and less getting something to sit with a cue. All his talk about Super Mario made me think of that, because the sound cues (e.g. jumping, landing on a mushroom, hitting a coin, whatever) kind of went in with the music, and created a new piece every time you played. The next step would be to have the sound engine actually deal with what sort of music played, how it played, and actually compose it itself, given conditions X, Y, and Z.



Jan.17.2006 @ 2:38 PM
I don't think the technology is as much an issue as the songwriting. Those memorable video game tunes were, well, memorable. Even though played on cheesy sounding "synths", they were well-written and catchy.

I think if you look back at a lot of old games, it's not a forgone conclusion that they all had awesome music. Some did and we remember them. Do fewer games today have awesome music? I dunno, maybe. The only game I've played in the last 15 years that I thought had remotely memorable music was Katamari Darmacy (sp?). (I'm not counting games with major label bands' music in them, e.g. Tony Hawk series) I think you can write a really catchy tune with a full orchestra, but that doesn't always happen.

I do think that the increase in available options for game music provides an easier way to make acceptable but forgettable tunes, though.

As for dynamic music changing with what's happening in the game, it has to be done right. I'm playing this XMen game on my PSP and if one bad guy shows up, the orchestra starts going apeshit with the fight music, and by the time they've ramped up, I've offed the guy, and I spend the next minute or so picking up stuff while the music is blaring...


Jan.17.2006 @ 11:42 PM
It seems like game music is heading the wrong direction. Back in the day when there was no memory for large, CD quality music and every console had dedicated synth hardware, plus more programmers than dedicated musicians/content people, there was thought about using the game to dynamically write the soundtrack. Now days, video games get caught up in these big high fidelity prerecorded, and very static pieces. At best we get awkward fades when certain things occur.

The only game I've played in recent times that has impressed me in terms of game/music interaction was Rez for PS2. All the game sound effects were in sync with the music, and as you did better new parts would be introduced to the audio. It was pretty much like using game events as midi/automation data for a sequencing program.


Jan.17.2006 @ 11:57 PM
the harvestman
I completely agree with the author that the era of chip-based sound generation made for an era of "pure composition" unlike much else, especially in Japan where perspectives on Western counterpoint were already "unique". From a decade of similarly crippled sound devices, we've established a vocabulary of "VGM-like" harmonic cues that still retain their function when used with more advanced instrumentation. I think the biggest thing responsible for the detour from pure composition was the capability to stream full-length PCM audio tracks during gameplay... suddenly, it's not so much about maintaining an accomplished in-house composer than the maintenance of music license of pre-recorded material. This happened right around the time that the "business" became an "industry" -- some particularly flagrant offenders operate from the model of movie-license shovelware, ignorant of the concept of diminishing returns. Just look at the gross misuse of Rob Hubbard at EA - after joining the company and completing his best work ever (skate or die series, immortal, arcade classics), he was quickly placed in an administrative position where they got to

I agree that dynamic music is the way to go. Soldier of Fortune 2 had a really primitive setup for changing music cues, but the neat thing was that the composer, Z. Quarles, actually knew his shit and made some enjoyable loops for the project that actually fit the atmosphere of the game (using little more than a bank of K2000 presets.) That's not to say that all redbook projects are shit, things like Interplay's little-heard Descent 1 CD soundtrack are to me what modern game music should be: an extrapolation of the rhythmic and harmonic cues from old chip tunes into more advanced instrumentation. VGM was at its core a prime example of what's referred to "utilitarian music" in the scientific study of music cognition, adding a depth to a multimedia assembly meant as a thickening agent, while entrenching certain expectations in the listener who ends up playing way too many games composed by those anonymous Japanese composers.
The technical limitations of chip music caused its audience to develop a unique skill that's been forgotten these days: the ability to perceive a greater musical arrangement than what's physically coming out of the speaker. Much like how the brain is able to resynthesize the absent fundamental and lower harmonics of a tone in a less-than-ideal transmission, the severely limited parametric bandwidth of VGM instruments were able to successfully imply greater instrumental performances with careful composition, something that Hubbard and the anonymous Japanese cast of thousands were able to accomplish quite well. This cognitive ability is one that I'd write about at length, if I had the discipline for a doctoral program. Unfortunately, this history of low-fidelity composition has been completely wasted in the development of new CONTENT DELIVERY METHODS, where we're raising a generation of kids completely ignorant of the rich tonal accomplishments of game music with limited specification. Instead, their cognitive ability is wasted on unwrapping the musical meat from the latest lossy compression scheme, only to find that it's some overproduced dropping from some Hollywood shithead, not even intended for game delivery in the first place. Much like well-designed movie scoring, game audio is endangered if the consumers of degenerate music continue to get their way, demanding less and less in creative utility while begging for the latest in musical mediocrity: the "soundtrack" is no longer a collection of purpose-built composition, but a simple gathering of prerecorded studio work that does much to distract the player from appreciating the Gestalt structure of a good game design.
Algorithmic composition is a great way to produce new works for game delivery. I once did a simple exercise where I played the jazz intro to Mega Man 3 as a MIDI recording, then analyzed the functional tendencies of each note event as it relates to its neighbors in 2 dimensions, and exported a series of new themes: they all were quite listenable and had the distinct rhythmic and tonal character of the original tune. I still can't stand Lisp, though. davetron5000 is awesome.


Jan.18.2006 @ 12:05 AM
the harvestman
The end of paragraph 1 above should read as follows:

"Just look at the gross misuse of Rob Hubbard at EA - after joining the company and completing his best work ever (skate or die series, immortal, arcade classics), he was quickly placed in an administrative position where he got to deal with evaluating the latest methods of content delivery -- a pretty weird application for someone who's widely considered to be the most accomplished game composer ever."


Jan.18.2006 @ 12:20 AM
Chris Randall
"suddenly, it's not so much about maintaining an accomplished in-house composer than the maintenance of music license of pre-recorded material."

Hey, that's my check you're talking about!!! Microsoft shows me the love with the licensing!

That aside, well said. There are other good examples of current scores for games, including orchestral and semi-orchestral material, though. The last two Myst games, for instance. Well, hell. All of the Myst games. There was a similar puzzle game called "Qin" I believe. It had several sections where the game player had to actually work out tunes on ancient Chinese instruments in order to unlock portions of the game, which I had a great time with. And an _extremely_ spare ambient soundtrack all done on older oriental instruments that was quite nice.

Another one with a semi-orchestral score which is excellent, and has certain algorithmic elements, is the Homeworld series. The cutscenes were kind of over-the-top, but the in-game music was subtle and really went well with what was happening. As you say, many companies overlook the fact that music can accentuate the gameplay. Then again, many movie houses overlook the fact that music is an integral part of the film experience, as well.

I'll tell you what, though. Now that I think about it, most games I play that have song-based soundtracks (like the ones I've got songs on, Amped 2, Project Gotham 3, et al), I turn the music off entirely, because I don't want to hear Limp Bizkit covering "Don't Believe The Hype" 650 million times. I'm a huge Agent Orange fan, but Tony Hawk 3 about ruined "Bloodstains" for me.



Jan.18.2006 @ 12:59 AM
the harvestman
Shit man, I hear you when it comes to Tony Hawk 3. The soundtrack, while it had a really good selection of tracks, got really repetitive after a while with many songs that I've heard elsewhere. Not to mention that Hubbard already made the perfect skating game soundtrack years before, so the musical experience was kind of dull to me other than when the soundtrack was attenuated during periods of high activity in the game.

I just wish the same technical attention to detail was still given to musical implementation as it is when it comes to graphics technology. Fortunately, I think sound effects at the very least are going to receive some long-needed attention - the platforms that run these games now have the computational horsepower needed for a limited implementation of 3D audio (another huge topic of research over here)... once they're able to create the spatial illusion on the poor man's controlled environment (headphones) without doing something retarded like licensing Qsound, I think the standards for sound placement will be made much higher once a well-designed product makes it out there.

We fly into Anaheim in a few hours... time for a fun weekend of plumbing the bowels of the INDUSTRY.




Sorry, commenting is closed for this blog entry.