[Bf-committers] 5.1 surround sound in Blender?

[Knapp]

On Wed, Jun 2, 2010 at 5:33 PM, Roger Wickes <rogerwickes at yahoo.com> wrote:

> I started thinking about what this might look like in 2.7.
> Just as light raytraces and interference patterns and caustics are seen
> and emerge, so does sound.
>
> You would need a sound emitter panel for an object
> to load the sound file, with start/end frame and loop options
>
> You would need a sound reflection panel for an object
> to define the reflection frequency curve (what isnt reflected
> is absorbed). With presets this allows you to simulate whatever.
>
> 3D View would enable a loudspeaker icon to hover over the object to let you
> know that an object is broadcasting a sound
>
> Then a new type of object: microphone with appropriate icon of course
> with a panel to set that microphone to pick up sounds for a channel
> with a frequency response curve and level response curve, with presets
> of course to simulate condenser, lavalier, shotgun, and cardoid.
> It has a direction and cone, very similar to spot lamp settings.
> A mic could feed one or more channels at certain levels; e.g. a mic in
> front
> of the main actor would feed L and R channels evenly to overcome the
> mono -> stereo issue
>
> A standard library rig for 2 channel camera-mounted forward, 2-channel max
> separation,
> and 5.2 channel camera mounted, and 5.2 set mounted. You could for example
> parent the mic object to a character to simulate a wireless lav.
>
> To "render" all this, you would need a wave-broadcasting algorithim, which
> i guess
> is similar to a raytracer, to figure out what frequency and level is being
> sampled
> at that particular time, at 44.1 or 48 khz.
>
>  --Roger, not holding my breath.


I am with you on most of that but for the record side of it. Why do you need
the mics? The sound is only heard from the point of view of the camera or in
a game the player. The current camera should do fine as the point of
recording.

How strongly the actor is heard over the noise of the background could be
set by adjusting the volume on the transmission of the actor sound. All
source sounds only need to be mono but I guess they might be more or
less directional sort of like the lights.

A really basic implementation could  work simply by having the left/right
level adjusted and the volume based on simple distance away and the degree
to the right or left of the camera. The only difference about the scene
rendering is that you can hear things behind you that you can't see.

Maybe before we get into full absorption and reflection and echo effects we
might just have a node that sets room size and gives and ambiance to the
sound.

I too dream of 10.2 surround or perhaps binaural sound options! Perhaps each
sound could be adjusted and saved out on its own track for post production
work.

Anything would be better than what we have. Heck, as best as i can tell 2.5
does not even have pan yet.

Did you see the bit about Jack working? That will bring in some more tools,
if we can get jack going well on our system. Kubuntu 10.04 for me with the
lame stuff it has now causing popping on most editors that I have installed.


-- 
Douglas E Knapp

Creative Commons Film Group, Helping people make open source movies
with open source software!
http://bespin.org/~douglas/CommonsFilmGroup/phpBB3/index.php

Massagen Arbeit in Gelsenkirchen Buer
http://bespin.org/~douglas/tcm/ztab1.htm
Please link to me and trade links with me!

Open Source Sci-Fi mmoRPG Game project.
http://sf-journey-creations.wikispot.org/Front_Page
http://code.google.com/p/perspectiveproject/