From vdubois@tum.de Wed Jun 14 16:59:27 1995 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!tum!vdubois From: vdubois@tum.de (Valeria Dubois) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Jun14.165927.9384@tum.de> Date: Wed, 14 Jun 1995 16:59:27 GMT Organization: Technische Universität München NNTP-Posting-Host: tum.de X-Newsreader: TIN [version 1.2 PL2] Lines: 11 Porting stardust.c to the NeXTcube and Atari Falcon DSP subsystems. The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. However, when downsampling from 44.1 kHz to 64 Hz, accumulator overflow in the biquad IIR filter can destabilize the loop. Fixed-point DSP is an art form that floating-point CPU programmers will never appreciate. -- === Valeria Dubois === Dept of Electrical Engineering, Technische Universität München "The hatch is open. Is anyone listening?" From hdubois@cern.ch Thu Jun 15 11:40:27 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!cern!hdubois From: hdubois@cern.ch (Heather Dubois) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Jun15.114027.4193@cern.ch> Date: Thu, 15 Jun 1995 11:40:27 GMT References: <1995Jun14.165927.9384@tum.de> Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: Netscape 1.1N Lines: 9 In article <1995Jun14.165927.9384@tum.de>, vdubois@tum.de (Valeria Dubois) writes: > Porting stardust.c to the NeXTcube and Atari Falcon DSP subsystems. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Heather Dubois "Do not worship the copy." Macintosh IIfx / FreeBSD 2.1