From jwatanabe@cern.ch Sun Jan 21 06:05:55 1996 Path: ashpoly!cascade!bigsur!cern!jwatanabe From: jwatanabe@cern.ch (Julien Watanabe) Newsgroups: comp.dsp Subject: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jan21.060555.8979@cern.ch> Date: Sun, 21 Jan 1996 06:05:55 GMT Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: WinVN 0.99.5 Lines: 11 Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. 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. Source code in 56k assembly uploaded to ftp.ashford-poly.edu:/pub/telecom-archives/dsp56k/stardust_56k.asm. -- Julien Watanabe Hayes Optima 288 @ 28,800 bps "The hatch is open. Is anyone listening?" From dduarte@mit.edu Mon Jan 22 04:46:55 1996 Path: ashpoly!cascade!transat!sunet!nordrelay!mit!dduarte From: dduarte@mit.edu (Dave Duarte) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jan22.044655.4805@mit.edu> Date: Mon, 22 Jan 1996 04:46:55 GMT References: <1996Jan21.060555.8979@cern.ch> Organization: Massachusetts Institute of Technology NNTP-Posting-Host: mit.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 10 In article <1996Jan21.060555.8979@cern.ch>, jwatanabe@cern.ch (Julien Watanabe) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. > 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. -- Dave Duarte | dduarte@mit.edu PGP Key fingerprint: 5214 FC85 5424 CFB8 "Keep the channels open." From akarlsson@pacbell.net Tue Jan 23 22:51:55 1996 Path: ashpoly!cascade!brgate!saturno!pacbell!akarlsson From: akarlsson@pacbell.net (Anton Karlsson) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jan23.225155.8933@pacbell.net> Date: Tue, 23 Jan 1996 22:51:55 GMT References: <1996Jan21.060555.8979@cern.ch> <1996Jan22.044655.4805@mit.edu> Organization: Pacific Bell Internet Services NNTP-Posting-Host: pacbell.net X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1996Jan22.044655.4805@mit.edu>, dduarte@mit.edu (Dave Duarte) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Top-posters will be placed in my ~/.killfile without notice. -- Anton Karlsson -- Pacific Bell Internet Services "The carrier frequency does not drift by accident." From glarsson@kiae.su Fri Jan 26 01:16:55 1996 Path: ashpoly!cascade!mxgate!coyoacan!kiae!glarsson From: glarsson@kiae.su (Göran Larsson) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jan26.011655.9498@kiae.su> Date: Fri, 26 Jan 1996 01:16:55 GMT References: <1996Jan21.060555.8979@cern.ch> Organization: Kurchatov Institute of Atomic Energy, Moscow NNTP-Posting-Host: kiae.su X-Newsreader: NewsWatcher 2.0b24 Lines: 11 In article <1996Jan21.060555.8979@cern.ch>, jwatanabe@cern.ch (Julien Watanabe) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Netscape 1.1 crashes when loading large preformatted text. Stick to rn or tin! -- Göran Larsson Opinions expressed are solely my own and not those of Kurchatov Institute of Atomic Energy, Moscow. ------------------------------------------------------------