PDA

View Full Version : Digital Audio Processing, and "soundtracing"?



cronodragon
02-11-2006, 08:17 PM
I have done several image filters, and I would like to know if there are any good articles about audio filters, and maybe how to do it in DirectX, if I'm not asking for too much.

Also, I'm interested in knowing if there is something similar to raytracing, but for audio. I mean, an algorithm that simulates sounds according to physics of sound (not light), derivated from events like friction, air compression, resonance, etc... Would this technology be called "soundtracing"? Is it already developed? Maybe physics accelerating devices must enhance a few generations to be able to simulate something like that, what do you think?

JernejL
02-11-2006, 08:34 PM
I have done several image filters, and I would like to know if there are any good articles about audio filters, and maybe how to do it in DirectX, if I'm not asking for too much.

Also, I'm interested in knowing if there is something similar to raytracing, but for audio. I mean, an algorithm that simulates sounds according to physics of sound (not light), derivated from events like friction, air compression, resonance, etc... Would this technology be called "soundtracing"? Is it already developed? Maybe physics accelerating devices must enhance a few generations to be able to simulate something like that, what do you think?

You mean the the doppler and EAX effects? usually software library does that.. in software or hardware.

cronodragon
02-11-2006, 08:38 PM
No, I mean if I hit a bottle of glass with a hammer, the lib must simulate the sound of the hammer, the resonance of the bottle before breaking softened by the contained liquid, then the pieces of glass crashing each other, and everything crashing in the floor... not just playing wav files, but creating the sound waves from the physical simulation algortihms. :D

JernejL
02-11-2006, 08:54 PM
No, I mean if I hit a bottle of glass with a hammer, the lib must simulate the sound of the hammer, the resonance of the bottle before breaking softened by the contained liquid, then the pieces of glass crashing each other, and everything crashing in the floor... not just playing wav files, but creating the sound waves from the physical simulation algortihms. :D

It is just an EAX effect.. i doubt it is created on the fly, where did you see that?

cronodragon
02-11-2006, 08:57 PM
EAX? I've never heard about that, interesting.

Well, I was thinking about it, since raytracing can simulate light behavior, then there should exist something to simulate sound waves, if it doesn't then I'll try to make it :D But from what I explained to you, it seems like a lot of physic simulation is required, maybe it will be a future technology after several generations of physics acceleration... :arrow:

JernejL
02-11-2006, 09:26 PM
EAX? I've never heard about that, interesting.

Well, I was thinking about it, since raytracing can simulate light behavior, then there should exist something to simulate sound waves, if it doesn't then I'll try to make it :D But from what I explained to you, it seems like a lot of physic simulation is required, maybe it will be a future technology after several generations of physics acceleration... :arrow:

You can specify EAX effects (echo, environment, etc..) per listener and per source...

WILL
02-11-2006, 10:01 PM
Well this would be something rather innovative in the way of digital sound. The trouble would of course be in figuring out and dissecting what common elements in all the possible sound combinations would be. No small task indeed. ;)

However the idea is somewhat sound. (Forgive the pun. :P) You can would obviously tie this in with 3D audio generation and use the combination of those set sound elements in established sources (objects, engine-wise) from physics information gathered in a 3D rendered world.

Possible Scenario: Take your engine's 3D objects assign the common sound elements specific to each of them and measure in speed, mass and friction of both objects that impact and generate the resulting sound from those interactions.

I think that would become the basic idea of what you are looking to accomplish. Though, doing something as complex as a motor engine would require a TON of processing. But the core concept might herald in the need for new audio hardware to suport such and a new 'audio innovation boom' in computer and game audio would be born. ;)


BTW: I think OpenAL would serve as a great platform for this sort of thing as it already has all the 3D constructs right at setup, all ready to go. Plus with it's common scheme with OpenGL, it would help fasttrack some of the initial concept design work.


Just a few of my own ideas. I like the idea of the technology you propose, I very much would like to know more about your efforts on this. Be sure to keep us in the loop of how it all goes. :thumbup:

cronodragon
02-11-2006, 10:12 PM
Hey thanks! As you said it's a matter of studying every little effect and make combiations using a lot of processing power. But I think it's an interesting challenge. The game engine I'm making is based on doing what other engines don't offer. I think it would be a great addition. Then, I'll start crashing things around the house to find out how they sound :think: ...

I found something about modern audio in games:

http://www.digit-life.com/articles2/sound-technology/index.html

jasonf
02-11-2006, 10:36 PM
WOoooo, this sounds like a massive yet facinating area of research...

Things like EAX and DSP's are the equivalent of 2D Sprite systems with alpha.. quite primative but fund for a while.

Sound or Molecule vibration tracing would be a brilliant field to check out..

Instead of the usual psuedo realistic sound occurs here, gets quieter with distance (imaging early 3D sprite systems for comparison).. then add doppler effects (Basic Sprite rotation) and what do we have? A 2D system from the early 90's on an amiga

In the 3D world of graphics we have a lot because it's harder to fool the eyes than the ears.. or most peoples ears anyway, I say that because some people have fantastic hearing and playing computer games for them must be like a world where the best games are on a Sega Mastersystem.

Creative labs et al need to pull their respecive fingers from their bum holes and get to work on this.. realistic sound pressure simiuation. For a lot of us, we won't notice the difference.. but then maybe we will. Aferall, if I can tell the difference between a MP3 and an OGG Vorbis file of similar size, I'm sure others can.

WILL
02-11-2006, 10:38 PM
I think a key to the whole 'what kind (or category) of sound is this thing going to make' part is probably gonna be by material. Metals vs woods... or plastics vs. stone or rock...

:lol: I think you might even be able to get right down into chemistry almost for this... but I don't know if that is too extreme for what you want.

They even did this with smell though. I think there was 11 different elements? ...and light is obviously broken down into 3 so...

cronodragon
03-11-2006, 01:48 AM
I think light is easier to study because we can see it. It would be easier to study sound if we could be able to hit a glass and see how the energy waves propagate in the material and compress the air. Then we could see amplitude, length, dispersion, duration, decay, etc. Does any of you have some googles to watch this? :D

EDIT: I found a great source of math for sound:

http://hyperphysics.phy-astr.gsu.edu/hbase/sound/imgsou/souconw.gif
http://hyperphysics.phy-astr.gsu.edu/hbase/hframe.html

That map reminds me a lot the technology tree of Civilization :D