Advertisement

Spatial simulation part2 - need advice

Started by August 05, 2003 04:20 PM
3 comments, last by Claymore 21 years, 3 months ago
here''s the problem.. If we assume that the sound is coming from the left and the listener''s year drums are 12cm apart, the left year should recieve unmodified dry signal and the right year signal filtered by virtual head (lowpass at 1.5khz) and delayed by, for example, 0.036ms. BUT once the signal is reproduced by headphones/speakers, the left year will also recieve 2 times lowpassed (second lowpass is done by the real head) signal delayed by 0.036+0.36ms (0.036ms = interaural sound travel time in case of headphones) and amplified by "right speaker - right year distance"/"right speaker - left year deistance" meanwhile the right year will recieve one more filtered signal, this time with 0.36ms delay and amplified by "left speaker - left year"/"left speaker - right year distance". This leads to low frequency accentuation. Now if we generate these unwanted signals, invert and add them to the mix, the problem will be corrected but only for a certain pair of years and a certain pair of speakers of a certain position. Is there any other way one could get rid of this problem without making the spatial image confused? (preequalizing the signal seems to affect the image because of pinna-related spatial freqency response, for example)
-Claymore-
You''re trying to emulate the job of the head when the head is going to do it anyway. Make life easier by letting the head do it''s thang?!

Just play the original sound into the left ear only; end of problem. That way the right ear receives the naturally filtered sound once.
Advertisement
If I did as you said, only two source positions could be simulated properly - full left and full right - which wouldn''t be such a problem if it didn''t sound bad.
-Claymore-
OK, so give up.

The only way what your attempting would work is if you had 2 direct feeds into the brain, thus bypassing the ears altogether whilst simultaneously recreating the effects of the head.
IMO spatial sound never works unless it''s trained for specifically for each user... Basic panning will do for left/right but anything else just slows down the program yet sound no better... at least that is my experience with any so-called spatial sound I''ve heard.

Mark
Bytten Independent Games Magazine
http://www.bytten.com
Sound Effects For Game Developers
http://www.indiesfx.co.uk

This topic is closed to new replies.

Advertisement