Lets talk audio. - TTLG Forums

Chade on 21/11/2013 at 00:15

Eh, this bit of the article hints pretty strongly that it's not in anymore:

Quote Posted by TFA

Weir: ... What I was trying to do on Thief was kind of tackle this slightly ... What my ambition was, which of course we never realized, was that you could start a level with one piece of music and end it with a totally different piece of music, and you'd never know when the music had been changed.

DET: Has that never been done before?

Weir: I don't think it has. Music's been changed but it's always different cues. Our system solved the technical problem but we didn't quite solve the human problem.

DET: The direction the game's taken?

Weir: Yeah.

[Before you ask, Weir does elaborate upon this but only off the record.]

Impossible to say exactly what happened. Could have been one influential employee who didn't like it. Could have been they wanted to keep the same tone throughout the mission. Could have been anything really.

jay pettitt on 21/11/2013 at 00:19

Oh good spot. I somehow missed that.

I also missed the original (http://www.ttlg.com/forums/showthread.php?t=142473) TTLG T4 audio thread which was probably a perfectly good thread. If a mod passes by can they shift this there?

nickie on 21/11/2013 at 08:44

Your wish is my command.

jay pettitt on 21/11/2013 at 09:59

Oooh! I'll have Tea then please.

nickie on 21/11/2013 at 10:00

With or without milk and sugar?

henke on 21/11/2013 at 10:16

Quote Posted by Chade

Eh, this bit of the article hints pretty strongly that it's not in anymore

To me it reads more like the generative audio is still in, but the system that changed it completely over the course of a mission is out.

Anyway, I'd have to hear this in-game for myself in order to judge if it's any good or not. I do love the idea of generative music simply because from a technical standpoint it sounds cool, but whether or not it'd actually make for a more pleasant soundscape than just regular background music I can't say.

jay pettitt on 21/11/2013 at 10:29

There's some mumblings of something that sounds like generated audio in the background of the gameplay demos, but it's pretty lacklustre compared to what we heard in Weir's Stealing Sound presentation. Maybe that's because the demos were kinda rushed through and the swells and grinds don't get a chance to seep into the brain properly... Maybe it's because there's a guy talking over the top... But mostly it sounds like it's been toned down and Weir's audio maybe isn't going to be a big feature of the game. Shame that.

cyrosis on 21/11/2013 at 16:19

Quote Posted by jay pettitt

You're much better off using geometry to calculate reverb - like TrueAudio does ~ but the last thing you need is to lock that stuff down in propriety solutions which is what has happened in the past and what AMD are doing now. You need to get that stuff in OpenAL so when you make a game it's a core feature for all your customers, not just some.

As far as I know, TrueAudio just uses convolution reverb, which is even more static than what's possible with EAX or other traditional digital reverbs. Using in game geometry to calculate reverb is ray tracing, which is what A3D offered in Thief 1 and 2, but I haven't heard it mentioned from any of AMD's press.

Convolution reverb is great for audio production where you are attempting to imitate an immobile instrument, and if the IR is properly captured, it offers the most realistic possible static reverb right now. For a game however, the audio sources are constantly moving around the environment, so having a perfect imitation of a static reverb isn't all that fantastic. On top of this, unless the IR's are captured exceedingly well, their noise floors will be audible (especially with many modern games using dynamic compressors), and attempting to aggressively cut these noise floors can result in truncated reverb tails.

Best case scenario leaves us with a reverb that offers the same features as EAX 5, but with slightly more realistic environments. Worst case, is they still don't offer the proper occlusion, transitions or environment combining that EAX 5 did; as these will require additional filters to post process the convolutions static output. Since EAX and most software reverb is algorithm bases, they can be changed directly from the source without adding to the signal chain, making some of those effects easier and less processor intensive to implement. To me, ray traced reverb is still the perfect solution for games, but it doesn't look like anyone wants to pick that up again.

Lastly, if you watch AMD's "GPU '14 Tech Day" video, the Thief audio lead talks briefly about how they will be using the new convolution reverb in true audio.

jay pettitt on 21/11/2013 at 16:56

Oh right - I just assumed from the feature set and that it was done in the gfx card that it'd be linked to rendering geometry that it was dynamic. Ho hum.

Darkness_Falls on 21/11/2013 at 21:53

Quote Posted by jay pettitt

I've tried listening as hard as I can with my special headphones on during the gameplay demo vids, and I can't hear much audio at all. The reverb sounds quite well done when NuGarrett is indoors, but that's as far as I get. I'm slightly worried (is worried the right word?) that nice reverb will be an AMD exclusive.

These two videos are probably the two best videos I've seen for hearing game audio and ambient music:
- (http://www.dailymotion.com/video/x15c9i6_thief-eurogamer-expo-2013-demo-gameplay_videogames) Eurogamer Expo 2013 Gameplay Demo - Some Commentary
- (http://www.gametrailers.com/videos/7h5jx0/thief-e3-2013--stealth-gameplay-demo) E3 Gameplay Demo - No Commentary (6:43-7:30 is a nice moment for ambient music. I hate all the gamey Focus, "Completed Step / New Step," and pickpocket *whooshing* sounds going on during this time, though.)

For the E3 one, I can already tell hearing the mob yelling in the background (probably a looping track) will get a bit distracting for me. I hope it isn't heard on multiple missions throughout the game.