13 Jan
2011
13 Jan
'11
8:41 p.m.
On Thursday 13 January 2011 19:48:38 Tom Szilagyi wrote:
On 13 January 2011 19:44, Brett McCoy
<idragosani(a)gmail.com> wrote:
Hm, that would then represent a stereo source in a virtual room recorded to a
stereo sink. Nice. Not.
What kind of purpose has this??? When I want reverb, I want to place the
source in a certain position (via pan for example) and get the reverb for the
room with the source "at that position", not the reverb of a stereo-
reproduction of that source at another position, put into the virtual room and
recorded in stereo again...
The only places where convolution-reverbs make sense imho:
1) Place all sources in short distance to each other compared to the virtual
room. Then you have a mono-to-stereo (when you mix to stereo). This will
actually give good results.
2) Use an impuls-response-package with several inputs for several positions
with each the stereo-output of the listening position. And the route your
signals to the different source-inputs depending on the position in the room
you want to give them. Gives better results then #1 but requires better
reverbs.
3) Fold the ambient of your stereo-song played on a stereo system in the
virtual room and record it at another position. Sounds like fun, but works
reliable only when you use head-phones for listening. And is not worth the
computation time for music that is to be played on any system from high-end
stereos to mp3-players and mobile phones.
Have fun,
Arnold
Can you explain what the difference is here with
'stereo' and 'True
Stereo', as far as convolution reverb goes?
Quoting the IR homepage:
[...] IR supports so-called 'True Stereo' impulses. These are four
channel impulses that describe a full convolution matrix (the four
channels contain, in order, the convolution paths L->L, L->R, R->L,
R->R).<end quote>