From arobertson@clari.net Mon Dec 04 19:38:38 1995 Path: ashpoly!cascade!mxgate!coyoacan!clari!arobertson From: arobertson@clari.net (Andrei Robertson) Newsgroups: comp.dsp Subject: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec04.193838.3757@clari.net> Date: Mon, 04 Dec 1995 19:38:38 GMT Organization: ClariNet Communications Corp NNTP-Posting-Host: clari.net X-Newsreader: TIN [version 1.2 PL2] 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. By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. Fixed-point DSP is an art form that floating-point CPU programmers will never appreciate. -- Andrei Robertson SupraFAXModem 28.8 @ 19,200 bps "When every node is whole, the channel opens." From cnovak@earthlink.net Tue Dec 05 18:01:38 1995 Path: ashpoly!cascade!transat!sunet!nordrelay!earthlink!cnovak From: cnovak@earthlink.net (Camila Novak) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec05.180138.9105@earthlink.net> Date: Tue, 05 Dec 1995 18:01:38 GMT References: <1995Dec04.193838.3757@clari.net> Organization: EarthLink Network NNTP-Posting-Host: earthlink.net X-Newsreader: Netscape 1.1N Lines: 10 In article <1995Dec04.193838.3757@clari.net>, arobertson@clari.net (Andrei Robertson) writes: > 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. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Camila Novak "When every node is whole, the channel opens." Pentium 90 (OS/2 Warp) / TOS 4.04 From jito@berkeley.edu Thu Dec 07 04:00:38 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!berkeley!jito From: jito@berkeley.edu (Ji-woo Ito) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec07.040038.3237@berkeley.edu> Date: Thu, 07 Dec 1995 04:00:38 GMT References: <1995Dec04.193838.3757@clari.net> <1995Dec05.180138.9105@earthlink.net> Organization: UC Berkeley Department of EECS NNTP-Posting-Host: berkeley.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Dec05.180138.9105@earthlink.net>, cnovak@earthlink.net (Camila Novak) writes: > Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- === Ji-woo Ito === Dept of Electrical Engineering, UC Berkeley Department of EECS "Measuring a record you own is listening, not piracy."