Sound Engine work-flow

[soundgeek2496]

I've found some interesting work-flow, tutorials on the net and youtube with using Wwise, but have found little examples of using Fmod. From my understanding Fmod is a fairly popular sound engine who's licensing is much cheaper vs. others? I have experience with everything about sound up to the point of its implementation into a specific game platform, and I'm curious as to the ideology of how these sound engines are used from start to finish.

[george]

Hey Soundgeek,

I have used Fmod (though a much older version of it) recently. It basically takes the onus off the game engine in regards to handling sound implementation and execution. I have not yet used Wwise, so I can't speak for the differences between the two. Fmod provides a pretty simple gui with which to create sound packages that are referenced within the game engine (it's compatible with Unreal Engine obviously). Creating packages that are non-streaming and streaming is pretty simple too... although atm I have to run full audio builds every couple of days which takes about 6 + hours.. But, as I said this implementation and version of Fmod that I'm using is fairly old and probably should have been re-thought before starting production (not my decision unfortunately ).

That all being said I have also heard really good things about it as well. We usually use our own proprietary engine that we have developed which includes our own sound engine that was written from the ground up. I think that's really the best way to go if you have the dev/programming time to make it happen. Although, these days, it's getting harder and harder to ignore 3rd party solutions like Wwise and Fmod for their relative ease and simplicity of integration, with full feature sets out of the box.

Hope this helps a bit?

George

[soundgeek2496]

That does help a bit, but what I'm looking for really is (and I apologize if this is crossing the disclosure line), what is the assembly line process of capturing your sound to finalizing it into the game with regards of using game engines (3rd party or in-house software). What delivery methods are required/expected from programmers from the audio designers/editors?

[george]

Well, this could be a lengthy post

I am actually planning on writing an article for the site that goes through this process. It really varies from studio to studio as to what the delivery process is, but in my experience, I am responsible for implementing the sound into the engine. What does this mean? Basically I create all the audio then hook up each sound into the engine (either onto animations, sound markers for ambiences etc or sound triggers). Obviously I am in control of volume, pitch, randomization, whether the sound is looping or not, etc.

As far as the programmers go, they are there to support audio by fixing bugs or possibly implementing new code features (like DSP effects, etc).

That's the basic rundown..

George

[soundgeek2496]

Thanks, I'll be looking forward to the article.

[Xan]

George's answer covers it all really (and is much more succinct than mine, below).

For my part I'll say that we don't really outsource, except for speech. Basically I 'deliver' stuff I make in the way I know we'll use it.

But if I sit back and examine the process, I can perhaps elaborate about the initial phases. The following might be useful information to give a person making the files on my behalf:

We render/bounce/mix/bake audio files out at a fairly high rate, 44.1K minimum. It's later that the programs we use to build the game resample the file (down-grade the sample rate to recoup memory). We prefer Wave files, usually PCM but other flavors are fine; BWAV, etc. I've worked with another company's tools before and they exclusively used AIFFs. There's no real difference, just somebody's personal preference.

Certain things about the way the sound effect is composed are important.

Space taken up in memory is always an issue so we'd prefer nice, tight audio. Long tail ends of sounds that are almost inaudible, 'white space' as seen in a wave editor, can be trimmed off. If I'm working on large sets of files (like speech or collections of physics sounds) I like to use an auto-trim/crop plugin to do this.

This next bit might be harder for a freelancer to factor as it requires a bit of foresight into how a sound might be used and it's perceived importance in a soundtrack. Basically, because we might end up downgrading the sample rate of a file for playback in game, I try to compose the sound so that it will sound (almost) identical when played back at this lower rate. Cloth movement foley (especially for Non-player characters) probably won't be important so it will go in at a lower sample rate. Therefore it'd be daft to use lots of high frequency content - or even make that content the important part of the sound - because it will get naturally 'filtered out' during the resampling process.

...yeah like George said. lots of stuff could be written here. But that's the end of my lunch break. Sorry if some of that seems basic, patronizing, I'm not sure where you're coming from exactly.

Also, do you get what I mean about resampling? That was all a bit rushed so let me know and I can explain more about how and why we do it, and in what instances we do it (for example that's sort of old hat, we don't need to resample nearly so often on the next-gen platforms, only on the Wii).

[Xan]

Yeah, ok. So... I saw you did cinematics for Hellgate.

I'm guessing about 2% of that was useful

[soundgeek2496]

I believe I see what you're saying about re-sampling certain sounds that can maintain their composure even at lower rates, and or possibly sounds that are not as dominate or in focus in the mix or event to preserve memory usage yet still play a part of what is being heard. As for Hellgate, I personally didn't work on it, but I worked with the cinematic sound designer for a year (with other projects, like Army of Two). He was a great guy and often had me come into his office to check out his latest mixes on those cinematics.