[PD] Creating auidioengines for games using PD

[Andy Farnell]

On Thu, 27 Dec 2007 13:13:19 +0100
Martin Ahnelöv <operagasten at gmail.com> wrote:

> > 
> > So, maybe we're crazy trying to advocate disruptive new technology
> > into that situation, but someone has to do it ;)
> > 
> 
> We're not crazy. They are just too narrow-minded that they see the
> current situation more clear than the future. 

'Narrow minded' needs qualifying I think Martin. Having studied this
area from more than just a technical angle, asking content producers and
games developers about the how and why of static data model dominance,
it's as you would expect with many industrial processes, when you already 
have your factories set up a certain way and all your staff trained to 
think a certain way, then it's difficult to switch mid-stream.

To me the question is really, why have developers ignored the inevitable
technology change for so long? Kees Van den Doel and Perry Cook started
this ball rolling almost a decade ago, but it's been largely ignored
until now. The main reason is we didn't have the CPU capacity in 2000, now
we have a potential excess of CPU cycles. R&D into procedural media has
been inhibited because nobody ever saw a need for it while Moores law
applies as much to data storage as it does to computational resources.
Now we've hit a situation where there are solutions running around looking
for problems and execs asking "What are we going to use all this powerful
technology for?" To get there it's more than just saying, "Okay we'll use
procedural content in the next game", you need to develop toolchains and
train content producers to use them as part of the investment.

The other reason is much like the oil companies contribution to avoidance of
sustainable energy, and the record companies reckless denial of changing
production and distribution models  - change doesn't happen until it hurts.
The only way it's ever going to hurt in entertainment business is when your
competitor can deliver a product at a fraction of what it costs you. 
That's not going to be the clincher for procedural content in the short term
as it will initially cost more to create it than traditional data content.

The catalyst is when your competitors can produce things that you simply cannot.

Many new directions in gaming with open-ended user driven content, maleable
game worlds and assets etc, they cannot be served by the data model. This is
what drives procedural media now, the move away from the linear narrative
game experience.

We could add here that most games producers have always fundamentally 
misunderstood the nature of VR and online games by attempting to shoehorn
them into a 'filmic' form, but that is changing because of user demand for 
richer experiences.


> 
> I think we should show how it should be done by example. I think we
> should develop a game that Does the Right Thing.

Never forget that entertainment is user driven. Despite the Hollywood
"push model" the fact remains that people decide what they want, so the
Right Thing (TM) isn't dictated by us technologists. Probably the thing to
hope for is a Better Way (TM) of doing more of the same, faster, cheaper.

So, Spore may have the 'design decision' of using procedural media, but
unless it is a good game nobody will care about how it works. The same
will apply to proc audio, the best to hope for is that the audience simply
don't even notice the technology has changed from data model to synthetic
production. Then we can transparrently move from games with a few thousand
sounds to games with a few trillion (or effectively infinite variation)
without anybody noticing.

Andy

-- 
Use the source