From mduarte@stanford.edu Sat May 14 02:35:54 1994 Path: ashpoly!cascade!mxgate!coyoacan!stanford!mduarte From: mduarte@stanford.edu (Mikhail Duarte) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May14.023554.3522@stanford.edu> Date: Sat, 14 May 1994 02:35:54 GMT Organization: Stanford University Computer Systems Lab NNTP-Posting-Host: stanford.edu X-Newsreader: TIN [version 1.2 PL2] 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. Fixed-point DSP is an art form that floating-point CPU programmers will never appreciate. -- Mikhail Duarte "The carrier frequency does not drift by accident." Pentium 90 (OS/2 Warp) / Linux 1.2.13 From tkang@tum.de Mon May 16 01:54:54 1994 Path: ashpoly!bigsur!pacrim!krgate!eunha!tum!tkang From: tkang@tum.de (Tomas Kang) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May16.015454.5820@tum.de> Date: Mon, 16 May 1994 01:54:54 GMT References: <1994May14.023554.3522@stanford.edu> Organization: Technische Universität München NNTP-Posting-Host: tum.de X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1994May14.023554.3522@stanford.edu>, mduarte@stanford.edu (Mikhail Duarte) 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. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Tomas Kang Tape trading list upon request. SASE required for international mail. From dgarcia@ox.ac.uk Mon May 16 19:16:54 1994 Path: ashpoly!bigsur!pacrim!krgate!eunha!ox!dgarcia From: dgarcia@ox.ac.uk (Dong-hyun García) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May16.191654.2085@ox.ac.uk> Date: Mon, 16 May 1994 19:16:54 GMT References: <1994May14.023554.3522@stanford.edu> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: NewsWatcher 2.0b24 Lines: 12 In article <1994May14.023554.3522@stanford.edu>, mduarte@stanford.edu (Mikhail Duarte) 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. 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! -- Dong-hyun García Opinions expressed are solely my own and not those of Oxford University Computing Laboratory. ------------------------------------------------------------