next »

[Sleepy Volcano]

Anyone do any work with binaural audio for game sound design? I've been experimenting with binaural recording since the early days the portable mini disc was relevant, and lately have found myself doing a lot of it again. Seems like a no brainer for immersive audio. Especially for cut-scenes and player sounds. A lot of programmers I've worked with shy away from it because they are so used to positioning mono source sounds in a 3D space to create faux 3D immersion, but binaural is so amazing. Anyone done anything?

[Daniel Pellicer]

The problem with binaural is that as soon as you use speakers you dont appreciate it properly but otherwise it would be very cool to have more of it in our daily videogame dose ;g

[dawid w. mika]

Maybe this is a good way for indie. Binaural recording is much easier and cheaper way to achieve dimensional sound (for me it's also more convincing). I'm thinking on investing in something like that: http://3diosound.com/index.php?main_page=product_info&products_id=45 - does anyone tried to used it for sound design?

As to speakers, I remember that a few weeks ago I read something about works on binaural speakers and it has something to do with transaural technique.

[Sleepy Volcano]

Wow, that 3DIO looks interesting. I'm using the Soundman OKM mics paired with their dummy head straight into my zoom H4N. It translates incredibly well to speakers. Of course it doesnt have the 360 effect without speakers, but the stereo separation is fantastic. I know you really can't do binaural for ALL sound design. But just for a few "WTF" splashes here and there and for ambience / player sounds.... like footsteps, inventory sounds etc.... it really makes a huge difference for the headphones experience. With the future of VR looking pretty bright, seems like now is the time to discuss raising the bar of audio immersion.

[MoritzPGKatz]

Wow, that 3DIO looks interesting. I'm using the Soundman OKM mics paired with their dummy head straight into my zoom H4N. It translates incredibly well to speakers. Of course it doesnt have the 360 effect without speakers, but the stereo separation is fantastic. I know you really can't do binaural for ALL sound design. But just for a few "WTF" splashes here and there and for ambience / player sounds.... like footsteps, inventory sounds etc.... it really makes a huge difference for the headphones experience. With the future of VR looking pretty bright, seems like now is the time to discuss raising the bar of audio immersion.

Well put!

The binaural technique, like 3D Virtual Reality goggles, is actually pretty old. Back in the 70's, Krautrock acts like Can produced whole albums using dummy head stereophony.

And, like 3D VR goggles, purely binaural audio work somehow never really caught on or found its way into mainstream production - yet.

Good chance for Indie devs? Maybe.
I've tinkered with binaural recordings using the same setup (OKM+Zoom), and while the results are interesting, I feel that - assuming you're not going for ultra-realism - you often still need that "artificially direct" smack with direct control over the stereo width and placement. Needless to say, creating the perception of a three-dimensional space using well-mic'd and well-placed mono signals and reverberation is a skill that has more uses in sound design as well as in music.

Anyway, the trend in 3D gaming is towards audio engines that simulate acoustic properties like reflection, absorption or even precise timing more and more realistically. Look at what graphic engines do with light nowadays - and in many ways, sound distribution is a lot less complex...

Interesting topic at any rate - we'll see what the future brings!

Cheers,
Moritz

[Schrompf]

Someone at QuoVadis told me that the Wise Audio library already does that: calculating impulse responses from geometry and applying those one a per-playback base to sounds. I wonder if you could also calculate or derive an abstract impulse response for positions alone. To reproduce binaural sound for arbitrary sounds. But I guess if it's possible, someone already did it.

[chris_b]

Someone at QuoVadis told me that the Wise Audio library already does that: calculating impulse responses from geometry and applying those one a per-playback base to sounds. I wonder if you could also calculate or derive an abstract impulse response for positions alone. To reproduce binaural sound for arbitrary sounds. But I guess if it's possible, someone already did it.

There's a whole set of recorded impulses for differeent angles and elevations here:

http://sound.media.mit.edu/resources/KEMAR.html

[Daniel Pellicer]

Someone at QuoVadis told me that the Wise Audio library already does that: calculating impulse responses from geometry and applying those one a per-playback base to sounds. I wonder if you could also calculate or derive an abstract impulse response for positions alone. To reproduce binaural sound for arbitrary sounds. But I guess if it's possible, someone already did it.

There's a whole set of recorded impulses for differeent angles and elevations here:

http://sound.media.mit.edu/resources/KEMAR.html

Schrompf you just need to do the convolution between thos and your arbitrary sound and you would have it alredy. I did something like this as a university little project.

[MoritzPGKatz]

Someone at QuoVadis told me that the Wise Audio library already does that: calculating impulse responses from geometry and applying those one a per-playback base to sounds. I wonder if you could also calculate or derive an abstract impulse response for positions alone. To reproduce binaural sound for arbitrary sounds. But I guess if it's possible, someone already did it.

There's a whole set of recorded impulses for differeent angles and elevations here:

http://sound.media.mit.edu/resources/KEMAR.html

Schrompf you just need to do the convolution between thos and your arbitrary sound and you would have it alredy. I did something like this as a university little project.

Could you please elaborate? Didn't quite get that...   

[Daniel Pellicer]

Someone at QuoVadis told me that the Wise Audio library already does that: calculating impulse responses from geometry and applying those one a per-playback base to sounds. I wonder if you could also calculate or derive an abstract impulse response for positions alone. To reproduce binaural sound for arbitrary sounds. But I guess if it's possible, someone already did it.

There's a whole set of recorded impulses for differeent angles and elevations here:

http://sound.media.mit.edu/resources/KEMAR.html

Schrompf you just need to do the convolution between thos and your arbitrary sound and you would have it alredy. I did something like this as a university little project.

Could you please elaborate? Didn't quite get that...   

Take A as your arbitrary sound or Head related transfer function sound then B as the opposite.

Take matlab and take a look to the conv function. It multiplies the two sounds in a way that your arbitrary sound will be filtered to sound as its in a point in space.



Then once you understand how it works you can implement yourself in your program or look for a simmilar function that already makes that.

The different head related transfer function have normally a azimuth and a left or right position angle (I forgot how is that called) if you want to place the different sounds automatically you will have to take different HRTF depending on those two parameters.

PD: Sorry I know Im not clarifying it too good, but Im a little bit sleepy . If any doubts ask again

[MoritzPGKatz]

Pretty wild! Thanks for clarifying!

Cheers,
Moritz