Gama Article on Adaptive Music
I thought this was pretty interesting: Defining Adaptive Music:
I love the points made and this brings up a very good question.
This is partially put to the test in some games today, but I dont think it can be perfectly inserted. As stated, gaming is not a movie. You cannot create perfect chords and ballads to something that is always changing. You can set the mood with a certain song, but in my opinion I dont think it can be adopted.
I was thinking along the lines of adding more in depth music, something that when you here it, you remember those times killing the last terrorist and saving the world, but that can already be done.
How can we add music without breaking the concentration of the user?
The best music transitions Ive heard so far are probably in final fantasy or Halo, but its just setup so when the player reaches a certain point the new music kicks in.
"Most importantly: how can a game composer score a scene intelligently and compellingly, when she doesn’t know what is going to happen, when?"
I think the composers should work closesly with the level designers etc and just make sure they have the right music for the right parts. I love it when you get into a tight jam and there are alot of baddies waiting for my to peak my head out. The music plays such an important roll in the dramatic tension and creates more depth and mood for the player. It is impossible to know when the player is going to run or when the player is going to hide, all we can do is just set the scene the way we think it will effect the player the most, cross our fingers and hope for the best. :)
I, being mainly a audio programmer, can agree with this whole heartedly. There a number of resources out there but in general it's mostly the transition work that it needed to be completed by the composer. This would be for the most fluid transition anyway.
The programmer could code a algorithm that will fade one song into another but that could get off beat or they could pitch it up and distort it then play a new track (kinda simulating a record skipping).
In general the best way to get results is to have the composer know what system they will be working with and what limitations and advantages it possesses. If it automates transitions then spend more time on the sounds themselves. Basically knowing what parts of the job the audio manager will automate will greatly help with time management. Also knowing the basics behind how it works will help in implementation of said audio. eg do I need a 1second blank space at the ed of my songs so that my transition matrix won't distort the sounds.
About the adaptability.. I think its probably impossible to fully achieve, but you can come close.. Music is not something thats random there are rules and I guarantee there are certain things people want to hear (Johnny Cash was so successful by utilizing the 1-4-5 combo.. its the most pleasant combination of sounds to the human ear)...
There are all sorts of rules and styles of mood setting music:
if you want sad - play a slow tempo minor arpegio or diminshed chord progression with a quiet downbeat solo over top, a bassline that utilizes a lots of "pull-offs" and "slide downs" in minor scales and have a really off-time drum beat thats very light
happy - play aggressive major chords with standard walking bassline and timed drums
Theres tuns of them out there.. using these rules you could easily program a game to create the music for you, just by sending it the mood, and key you want.. you'd have to use a control structure to analyze every possible scenario and the mood associated with it..
Using the specific examples:
Is the player winning?
if user.getHP() > enemy.getHP() then playMusic(Happy, G)
else playMusic(Desperate, Eb)
If you have more hp then the enemy, play a happy tune in G, if the enemy has more then you, play a desperate tune in E flat.
How many orcs are left?
if enemy.getCurrentCount() < enemy.getMaxCount * 0.6 then playMusic(Epic, A)
else if enemy.getCurrentCount() = enemy.getMaxCount then playMusic(BloodPumping, E)
else playMusic(Happy, B)
Thats the closest way I can conceive to experiment with adaptability..
I never realized it before but a great example, probably the best I could find is Lumines for the PSP.
The music goes along with the game, if only there was a way to tap into peoples minds and play the music to their own emotions. Of course music can change those emotions by speeding up the tempo and what not.
As a composer/audio engineer, I feel that fully adaptive music, although amazing and possible, sounds too much when integrated within the game all the time.
If you watch movies you will see that music is not apparent all the way through, its the use of both music and audio (and its effects).
The skill is using it sparingly but still being interactive, this can easily be done on 'DirectMusic' using triggers and some classical music techniques.
I feel that more interaction between on screen visuals and audio (not just music) is the way forward.
Bioshock is a good example, It scares the s**t out of me some times, the enhancement of voices and other elements through EQ and reverb etc, using subtle changes (like in real life) really works for me. And is far more realistic, tricking your mind into a false sense of reality (Isn't this what games are created for)
probably the key to how to push audio forward in terms of intigration with the dynamics of games, is to reconsider what IS music in relation to the games world.
in the world of music, the sounds are broken into beats and bars, keys and chords. in comparison to the flow of games, this way of thinking is too rigid and restrictive.
instead, why are games sound designers not thinking more along the lines of "what does that leaky steam pipe sound like in relation to the engine hum over in that corner, mixed with the sound of the rain outside? does it mesh well with the sound of the players footfall on this surface?"
in games worlds, you have at least 64 tracks of simultanious audio that can be triggered and mixed in real time, in a true 3d space (thats not even taking into account the spacial alogorythms from EAX and the like) a good sound designer should make the environment itself, the theme music. if you approach a project like this, you will find your old "chords and keys" ideas will quickly seem redundant as the main meat of your sound structure, and instead can be relied on to do more subtle work.
for an example of how to make the environment itself the instrument, see doom3 (terrible game, but great sound work)
for an example of how NOT to do it, see half-life 2 (great game, terrible sound work)
(to qualify the last statement) half life 2 features single "showpieces" of triggered music (which is good enough music, and more or less fits the action)
but the music plays until it either ends, or even worse, stops abruptly when you leave the current area and start loading the next. it is jarring, and lends little to the immersive experiance of the game.
this is what i think anyway on the subject of adaptive sound in games.
The original article is one near and dear to me since it was the first real indication that someone else was thinking about these things in context with western art music. There are several posts that I want to comment on I'll just go from the bottom up.
Irishlostboy: Making music sound more like the environment is a great idea. Rather than be used as a standard for all in game cues I would love to see music gradually transition into more noticable and recognizable out of the ambient sounds of a level to accent tension events taking place in the game.
Dannyarmstrong: No real comment other than 100% agreeing with you. Great post!
Aurican33: I actually disagree that the adaptability is out of our reach. I will grant that certain musical styles are harder to use in adaptive music than others however I think ANY musical style can be used adaptively in games.
Let's start out with the easiest of the easy: minimalism. For those who can't immidiately place the musical style think Phillip Glass. Minimalism involves multiple ostinati (repeated patterns) and musical changes fundamentally involve adding, subtracting, and altering them. These changes can be triggered by in game events and depending on the nature of your ostinati you may not even have to bring the changes in on a barline for them to sound convincing. The static nature of purely minimalist writing would allow for more structured melodic lines to be cued in for key events and frequently accent changes in the ostinati make the same line sound quite different.
For those of you who frequent euro clubs my description of minimalism probably sounded very familiar. That's right, techno/trance music is not only wildly popular but also subject to the same additive structure as minimalism. It can be manipulated in the same ways.
Perhaps the next 'simplist' implementation is with pop/rock/funk/(shudder)country/etc. This actually isn't that simple at all compared to the above but my ultimate ideal of "Wagnerian Game Scoring" makes it look like a cakewalk (no offense to cakewalk product users out there, SONAR is great, I'm just a digidesign kinda guy). Here you have enormous creative freedom with a huge price to boot. You have to finish phrases. This means that if you need to have a contrasting event mid phrase you need to do it with dynamics, articulation, orchestration, or lyrics. It also bears mentioning that in these styles you typically have to wait for the next strong beat to have any significant change so there may be a slight delay between the event trigger and the audible cue.
Another big problem with dynamic scoring in this style is that everybody knows and expects to hear these song forms played out. What happens when a player completes the sequence before the cue runs out? What happens if they take longer than the whole cue? There are a lot of creative responses to this ranging from tempo modulation (thank goodness there's software that can do this without sounding like junk now) to the simple fade out. The cuteist solution I've found is used by Yoko Kanno in the "Run Rabbit Junk" cue found in many places throughout the Stand Alone Complex series. She just ends it with a one beat statement wherever the cue needs to end. I know, musically horrible right? Actually it's suprisingly effective because the cue itself is used to accent action and the end marks this action grinding to a halt.
There are a myriad of other musical problems but the point I'm trying to make is that all of them are ones that can be surmounted without a musical revolution.
My personal ideal of the "Wagnarian Game Score" has even more with the problems of coming up with leit motifs that can be played over each other even if offset by a few bars, not to mention having transposed versions for all of the harmonic environments they might have to be played in. Even so, while being both difficult, time consuming, and well beyond the sound budget of any game company in exsistance (actually, this may not be true. I have no idea what the rock band and guitair hero liscensing fees were but they must have been astronomical.).
The technology is there. I know this because I've been able to do all these things in protools for years. I know developers can script in game triggers because... well... they do. Synching these two is definately not beyond the industry's technical capabilities. The problems facing game music advancing in this direction have more to do with where it lies in the priorites of most companies. Heck, half of the games that win awards for their music don't even do it internally. These kind of endeavors are not the sort of thing you can outsource to Jeremey Soule because they require the compositional team to be on board and adapting at every stage of developement.
Finally the biggest barrier is that it would require having larger and more artisticly savvy sound teams and with game sound in it's present state it's a heck of a lot cheaper to produce than any other part of production. It's not all that uncommon to have only one Audio Designer/Composer/Audio Programmer on a title and then some credits for 3rd party contractors who help out with late production. Why jack up production costs for something that a mouth breathing public might not even notice? Everyone can immediately see the higher poly count or sweet volumetric effects but in one cursory playthrough how many players would even notice a well implemented adaptive score?
|All times are GMT -8. The time now is 08:49 PM.|
Powered by vBulletin® Version 3.6.9
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.