I think it's probably horses for courses. It depends how much repetition you need and whether your composition is complex/innovative enough that the listener can hear something new and different (or think they have) at every repeat.


At a micro rather than macro level (i.e. talking about thematic progression rather than repeating whole pieces/cuea), I'm still a fan of the baroque/classical/romantic "repeat it twice, the third time can be similar, but at that point you have to move it on". It gives a chance for the phrase to have maimum impact, but the ear should not generally get tired of it.

On the other hand, if you were to ask Phil Glass, he'd probably give you a totally different answer... [H]

If I play a scene and there's only one track for this one mood availible and music is playing all times, that's definately too much repetition. Ideally as a musician I'd love to still experience new things after playing the game 7 days in a row [;)] But that's almost impossible, I know, I know, it would also cost the producer a fortune and a half. But actually setting up more tracks i.e. 5+ for the same mood if it is arising more often is needed for an average length (2-3min) cue in my opinion. The shorter the cue and the more arising a mood is the more tracks you need. One alternate should be there for everything, maybe except the mobile ring. [[:D]] But then again, I rarely play games so much that I'd consider this a highly valueable comment [[:D]]

One thing I always thought about for little variations, but I don't think is technically easily do-able, is to setup a second, maye third or other varying, additioning cue to the major theme cut make a memorizable theme, while still preventing the repetition effect. E.g. one time a theme by a horn, with strings in background, the other time the same horn theme but instead of strings you hear brass accompanied. Drums could also vary but usually we are not that picky on repeating drum-loops. Kinda like a music construction kit building on the fly.

All the best,
PolarBear

Well the really cool thing to do is do stackable cells that change with the game status in addition to having more there one version of each "mood" or location piece. It requires some help from the programers obviously.

Indeed, the endless repetition in game music is a problem inherent to the nature of the format...
I actually find myself turning off the sound in games after a few days (the few I have time to play!), not because the music isn't 'good' or suitable - but because the same motive comes up again and again to the point that I simply can't stand to hear it any more. As others have already intimated, the solution is to simply have more music. Yes, there will still be cycling of the same bits, but at least some variety would be introduced. Whether such a thing will actually occur... I know not! [[;)]]

Actually, the solution would be the introduction of algorithmic composition into computer games. This would of course mean a completly new paradigm in approaching the act of composing, of course, and even more its acoustic realisation.

Our chief programmer, Chris Teuscher aka. "Golem" was into this a lot - but I'm afraid that he won't find the time to contribute to this topic in depth right now ... [H]

Thats just to smart for its own good. The bottom line of course is if it sounds good. does it?

@jc5 said:

Indeed, the endless repetition in game music is a problem inherent to the nature of the format...
I actually find myself turning off the sound in games after a few days (the few I have time to play!), not because the music isn't 'good' or suitable - but because the same motive comes up again and again to the point that I simply can't stand to hear it any more.

I do the same. Sad, because I'm sympathetic to the game composer. But after a few loops, I try to shoot the music and not the monster.

Wow, thank you all for the response. I think if one thing is clear from this dialogue, it is that at least some repetition is acceptable if handled correctly, but too much makes you want to turn the music off, which is a bad thing for composers.

Considering the fact that a game player will play any one game from 20 to100 hours, could it be that you will have repetition no matter what you do? There are other options such as adaptive music which have the potential to at least disguise repetition.

Take an action game for example. An adaptive score might follow the action with the appropriate music if your character is sneaking around unseen, winning, losing, or in an intense battle etc.

But the execution of this adaptiveness is what is truly the key. Any thoughts on this process?

Trying to get a kind of "modular" system for the music. Early games had problems with the size of files and still could create an enourmous amount of music using general midi or embedded tracker music. With big amounts of space and advancing compression methods we are able to deliver a lot of good produced music now. But it's also kind of a cost factor then, because every additional track will rise the costs. And I don't think the game industry is willing to spend a fortune and not having a real orchestra. And recording a 10 hour soundtrack with an orchestra isn't really discussable I guess...

So the composers and programmers have to get together and discuss the possibilities. The composer providing several tracks being able to represent the adpativness with a certain length that can be looped (hearing one track 3 times in a row and have the next time another track is better than hearing one track each time you have e.g. a battle scene IMO). You need a small portion of transition samples leading from one style into the other or connecting two pieces of the same style - this is also to introduce tempo changes, for example. If you combine that with the individual tracks being modular themselve, as I explained above, with the melody and accompaniment changing, the work for the composer should stay in the same range as if he just did one piece. A bit more perhaps. The challenge lies in the adaptiveness still I think.

Just my 2 cents,
PolarBear

Surely the emphasis should be on quality and not quantity? so in that regard, I will take an orchestra and a skilled composer over an algorithm and midi template.

What I am interested in doing however is creating a game score that is adaptive by using elements of my score to do so. I mean, if you code a playback engine that will allow you to stream and layer different tracks and cut and snip it at certain intervals in order to keep up with the situation on screen then you can begin to get something that is a good middle ground.

For the most part, I’d have quite a bit of ambience any way… you know, nothing too thematic when the player is just dawdling around and outside of pivotal moments. Then when a monster jumps out, you can trigger some stings and orchestral hits. You could do quite a few variations of these short stings because they would only be a few seconds long for the most part.

There is also this kind of thing...
http://www.directsong.com/direct_song_main.htm
though I'm not sure about how well it will work.

Of course quality is always assumed. And I hope every composer does his best to crap out some good if not his best stuff when on a project [;)] Actually I'd turn off a bad-quality soundtrack as soon as it is repeating the first time I guess...

James, your link seems to lead to a blank black page - is there a plugin needed to view it?

PolarBear

Hmm, yes you might need flash for it actually.
But basically it's just a place to download (at a cost) new music for your games, so for example they have Jeremy Soule composing a bunch of new tracks for Guild Wars.

Ah, today it's loading just fine... hmmmm....