Article View: comp.dsp
Article #2605Re: dsp audio algorithms
From: Adam McKeag
Date: Sun, 12 May 1996 00:00
Date: Sun, 12 May 1996 00:00
45 lines
2415 bytes
2415 bytes
The technique of recreating room responses in the manner described is called "Auralisation". The basic idea is to measure or calculate the impulse response(s) of the room and filter "dry" (anechoically recorded) material with these impulse responses. There are several software modelling programs available for calculating accurate room impulse responses (given room geometry, reflection of walls etc). The one I have experience with is CATT Acoustic from Computer Aided Theatre Technique, Gothenburg, Sweden. Such programs are used by the architectural acoustics people to try out hall designs before they are built. Special hardware takes care of the real-time filtering. If you want to do anything more than the first order reflections (from walls and other surfaces) you are going to need some very long FIR filters, as room impulse responses can last several seconds. To get the proper effect you really do need to implement the reverberant late part of the response. To do this in real time means you will need an FFT based filter implementation, which will give you more programming work to do. There is a technique that reduces the computation required at the expense of quality. You just pick out the major peaks in your long impulse response and only implement those. This is essentially what you get when you press the "hall" or "stadium" button on your CD player or digital amplifier. It doesn't sound really good. Good luck, Regards, Adam McKeag DSP Design Engineer Lake DSP Pty. Ltd. Sydney Australia PS. You can find out (a bit) more about this by looking at http://www.lakedsp.com. Sorry it's not more comprehensive.
Technical Details
Message-ID:
<4n5tst$6jj@metro.usyd.edu.au>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailReferences:
<4lq9js$grl@yama.mcc.ac.uk> <31860E11.41C6@research.att.com> <318DC55F.237D@bham.ac.uk> <Pine.OSF.3.91.960511103646.18880B-100000@internet>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: f8fcc,dbd46f298cf7c1de X-Google-Attributes: gidf8fcc,public X-Google-Thread: 1133f6,295b4c51f862d35e X-Google-Attributes: gid1133f6,public From: Adam McKeag <adam@lake.com.au> Subject: Re: dsp audio algorithms Date: 1996/05/12 Message-ID: <4n5tst$6jj@metro.usyd.edu.au> X-Deja-AN: 154494337 distribution: inet references: <4lq9js$grl@yama.mcc.ac.uk> <31860E11.41C6@research.att.com> <318DC55F.237D@bham.ac.uk> <Pine.OSF.3.91.960511103646.18880B-100000@internet> content-type: text/plain; charset=us-ascii organization: Lake DSP mime-version: 1.0 newsgroups: comp.dsp,fj.comp.dsp x-mailer: Mozilla 1.22 (Windows; I; 32bit)