How realistic is our vision about the 'realism' of sounds?
99% of us base our expectation on youtube clips. How close to reality are those clips? And, what a mic records in a car, how close is that to what a driver hears with a helmet on?
A sound might not have to be necessary 1-1 realistic to be convincing or have depth and subtleties to fool the brain into perceiving it to be "good enough"
Unless you were prepared to strap 2kw speakers to your ears then you are not going to get the volume of a real car and what makes up a large component of a real car engine sound so its always going to be abstract to the real thing in many ways.
But it will still be possible to have something that has many aspects of depth and subtaltiy of a real engine sound and be more analogous than what we typical here now in games.
some of the best quality you tube clips though recorded with standard microphones do a good job of encapsulating that excitement and feel of what a real engine sounds like even though the you tube clips don't really match what a real car sounds like.
I think with the current methods that sounds are captured and played back for driving games the only way to improve the quality would be to increase the number of samples and use different methods to capture the sound , its already very time consuming though and I'm not sure how many samples you can have playing in real time before you start to put a heavy load on the pc/ run into file size issues.
From what I gather with AC its current sound is done using a similar method to how Direct wave VST's work in audio software in that you have a bunch of wave files that then are distorted/tone bent into each other with maybe 1-12 channels playing at any one time for the basic engin + some channels for things like brakes , rumble strips , wind and specific pops from the engine.
Unless the number of samples are increased to the point that you have a couple of GB of well captured sound for each car from multiple microphones all mixed and combined perfectly I think its always going to sound quite artificial.
Thats why I believe it would be best to simply move to a synthesizer solution but the work and experimentation in that would be insane and is a R&D task on a par with the tire models these companies develop. ( and we all know how long they take to get to a good standard ;( )
Still with the current technology and the samples provided I think Kunos have done well , and for now I'm far more interested in the utility and function of the sound rather than the immersive potential of it.
