From sschneider@u-tokyo.ac.jp Tue Aug 06 15:19:03 1996 Path: ashpoly!cascade!transat!eunet-fr!teletel!u-tokyo!sschneider From: sschneider@u-tokyo.ac.jp (Sun-mi Schneider) Newsgroups: comp.dsp Subject: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Aug06.151903.6935@u-tokyo.ac.jp> Date: Tue, 06 Aug 1996 15:19:03 GMT Organization: University of Tokyo Department of Information Science NNTP-Posting-Host: u-tokyo.ac.jp X-Newsreader: NewsWatcher 2.0b24 Lines: 11 Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. However, when downsampling from 44.1 kHz to 64 Hz, accumulator overflow in the biquad IIR filter can destabilize the loop. By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. Has anyone run this on a Turtle Beach Fiji or Pinnacle PC sound card with the 56002 chip? -- Sun-mi Schneider | sschneider@u-tokyo.ac.jp PGP Key fingerprint: 7B4A 44D8 6BCB 9B6C "Keep the channels open." From fmartinez@cmu.edu Wed Aug 07 06:06:03 1996 Path: ashpoly!cascade!bigsur!cmu!fmartinez From: fmartinez@cmu.edu (Freya Martínez) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Aug07.060603.6365@cmu.edu> Date: Wed, 07 Aug 1996 06:06:03 GMT References: <1996Aug06.151903.6935@u-tokyo.ac.jp> Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1996Aug06.151903.6935@u-tokyo.ac.jp>, sschneider@u-tokyo.ac.jp (Sun-mi Schneider) writes: > Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. > However, when downsampling from 44.1 kHz to 64 Hz, accumulator overflow in the biquad IIR filter can destabilize the loop. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Freya Martínez Hayes Optima 288 @ ISDN 64k bps "A soul is not exclusive to biology." From nsilva@cern.ch Wed Aug 07 14:05:03 1996 Path: ashpoly!cascade!bigsur!cern!nsilva From: nsilva@cern.ch (Nadine Silva) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Aug07.140503.2957@cern.ch> Date: Wed, 07 Aug 1996 14:05:03 GMT References: <1996Aug06.151903.6935@u-tokyo.ac.jp> <1996Aug07.060603.6365@cmu.edu> Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1996Aug07.060603.6365@cmu.edu>, fmartinez@cmu.edu (Freya Martínez) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Nadine Silva US Robotics Courier V.34 @ 14,400 bps "A soul is not exclusive to biology." From dnovak@ox.ac.uk Thu Aug 08 09:33:03 1996 Path: ashpoly!cascade!mxgate!coyoacan!ox!dnovak From: dnovak@ox.ac.uk (Dave Novak) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Aug08.093303.7019@ox.ac.uk> Date: Thu, 08 Aug 1996 09:33:03 GMT References: <1996Aug06.151903.6935@u-tokyo.ac.jp> <1996Aug07.060603.6365@cmu.edu> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1996Aug07.060603.6365@cmu.edu>, fmartinez@cmu.edu (Freya Martínez) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- Dave Novak US Robotics Courier V.34 @ 14,400 bps "5 sigma is a discovery. 7 is a question."