Re: Linux binaural analysis (Catherine Emily Carr )


Subject: Re: Linux binaural analysis
From:    Catherine Emily Carr  <cc117(at)umail.umd.edu>
Date:    Fri, 5 Nov 1999 06:18:13 -0500

Eric Mousset <mousset(at)SEDAL.USYD.EDU.AU> wrote: >The department I am working for is considering purchasing some >equipment for research purposes in binaural hearing (HRTF-based sound >source localisation, amongst others). >The computer on which we are intending to run the (real-time) binaural >analyses is a PC running LINUX. > >1) Part of the question is general and applies to anyone interested > in real-time sound source localisation with a pair of mics as input: > There are apparently two main options for the acquisition of the > acoustic signals: a sound-card vs an A/D convertor. How do they compare? > >2) Linux-oriented question: Do most cards have drivers for Linux? > > >Many thanks in advance. > > >Eric. We at Mark Konishi's lab, Caltech, do exactly what you want to do, it seems. We use have computers running Linux 2.x and SunOS 4.1.x to do both behavioral studies and neurophysiology concerning sound localization in owls. We have done experiments with HRTF-based sound source localization, so what you seem to want to do can indeed be done. For neurophysiology, we use a custom written piece of software that runs under X-windows, and for the behavioral stuff we use Matlab with some extensions that we wrote to control the hardware. We use a data acquisition system from Tucker Davis Technologies, which consists of an ISA card connected to rack-mounted A/D and D/A conveters via fiber optic cables, in addition to filters, programmable attenuators, and some other equipment. This stuff works pretty well under Linux, with a few little glitches every few weeks, but that's it. I had thought as you did to use a sound card for my data acquisition device when I started four years ago. This was a royal pain in the butt. At that time, there were primitive drivers for a very few data acquisition cards. The Linux sound drivers in the 1.3.x kernels for the card I was using, the Gravis Ultrasound MAX, were not satisfactory, and after about six months of struggling with it, I admitted defeat, and we started looking for another system, eventually deciding on the Tucker Davis stuff. It would probably be easier to make the sound card work now, in the 2.2.x kernels. Many data acquisition cards have drivers (at varying levels of functionality) under Linux now: see the Comedi project <http://stm.lbl.gov/comedi/> If anyone has any other questions, they can e-mail me, since I'm not on this list. Chris -- cmalek(at)etho.caltech.edu "The price of secure connectivity is eternal vigilance!" -- cDc [PGP pubkey at ftp://ftp.etho.caltech.edu/pub/cmalek/cmalek.asc] -- Posted by: Catherine Carr Dept Biology Univ Maryland College Park MD 20742-4415 301-405-2085 (office) 301-405-6915 (lab) http://www.glue.umd.edu/~carr/


This message came from the mail archive
http://www.auditory.org/postings/1999/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University