From teriksson@u-tokyo.ac.jp Sun Feb 19 18:55:00 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!u-tokyo!teriksson From: teriksson@u-tokyo.ac.jp (Thiago Eriksson) Newsgroups: comp.dsp Subject: Observation: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Feb19.185500.9114@u-tokyo.ac.jp> Date: Sun, 19 Feb 1995 18:55:00 GMT Organization: University of Tokyo Department of Information Science NNTP-Posting-Host: u-tokyo.ac.jp X-Newsreader: WinVN 0.99.5 Lines: 11 Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. 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. -- === Thiago Eriksson === Dept of Electrical Engineering, University of Tokyo Department of Information Science "In theory there is no difference between theory and practice." From tsouza@mindspring.com Mon Feb 20 02:19:00 1995 Path: ashpoly!cascade!transat!eunet-de!hansebox!mindspring!tsouza From: tsouza@mindspring.com (Thiago Souza) Newsgroups: comp.dsp Subject: Re: Observation: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Feb20.021900.9090@mindspring.com> Date: Mon, 20 Feb 1995 02:19:00 GMT References: <1995Feb19.185500.9114@u-tokyo.ac.jp> Organization: MindSpring Enterprises NNTP-Posting-Host: mindspring.com X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Feb19.185500.9114@u-tokyo.ac.jp>, teriksson@u-tokyo.ac.jp (Thiago Eriksson) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Thiago Souza Zyxel U-1496E @ 28,800 bps "When every node is whole, the channel opens." From omacdonald@cmu.edu Wed Feb 22 03:29:00 1995 Path: ashpoly!cascade!transat!eunet-fr!teletel!cmu!omacdonald From: omacdonald@cmu.edu (Olga MacDonald) Newsgroups: comp.dsp Subject: Re: Observation: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Feb22.032900.2842@cmu.edu> Date: Wed, 22 Feb 1995 03:29:00 GMT References: <1995Feb19.185500.9114@u-tokyo.ac.jp> Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: NewsWatcher 2.0b24 Lines: 8 In article <1995Feb19.185500.9114@u-tokyo.ac.jp>, teriksson@u-tokyo.ac.jp (Thiago Eriksson) 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. -- Olga MacDonald Tape trading list upon request. SASE required for international mail. From asato@utexas.edu Thu Feb 23 18:58:00 1995 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!utexas!asato From: asato@utexas.edu (Astrid Sato) Newsgroups: comp.dsp Subject: Re: Observation: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Feb23.185800.1717@utexas.edu> Date: Thu, 23 Feb 1995 18:58:00 GMT References: <1995Feb19.185500.9114@u-tokyo.ac.jp> <1995Feb22.032900.2842@cmu.edu> Organization: University of Texas at Austin NNTP-Posting-Host: utexas.edu X-Newsreader: WinVN 0.99.5 Lines: 11 In article <1995Feb22.032900.2842@cmu.edu>, omacdonald@cmu.edu (Olga MacDonald) writes: > Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. If you are posting from AOL, please read RFC 1036 before posting again. -- Astrid Sato Practical Peripherals 14.4 @ 19,200 bps "The carrier frequency does not drift by accident." From jito@illinois.edu Fri Feb 24 04:56:00 1995 Path: ashpoly!cascade!brgate!saturno!illinois!jito From: jito@illinois.edu (Ji-woo Ito) Newsgroups: comp.dsp Subject: Re: Observation: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Feb24.045600.7855@illinois.edu> Date: Fri, 24 Feb 1995 04:56:00 GMT References: <1995Feb19.185500.9114@u-tokyo.ac.jp> <1995Feb22.032900.2842@cmu.edu> <1995Feb23.185800.1717@utexas.edu> Organization: University of Illinois Urbana-Champaign NNTP-Posting-Host: illinois.edu X-Newsreader: WinVN 0.99.5 Lines: 12 In article <1995Feb23.185800.1717@utexas.edu>, asato@utexas.edu (Astrid Sato) writes: > Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. > If you are posting from AOL, please read RFC 1036 before posting again. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Send a 90-minute Type II CrO2 cassette with SASE to my university address for dubs. -- === Ji-woo Ito === Dept of Electrical Engineering, University of Illinois Urbana-Champaign "5 sigma is a discovery. 7 is a question."