From sokonkwo@cam.ac.uk Thu Sep 08 01:10:36 1994 Path: ashpoly!cascade!brgate!saturno!cam!sokonkwo From: sokonkwo@cam.ac.uk (Steve Okonkwo) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Sep08.011036.9081@cam.ac.uk> Date: Thu, 08 Sep 1994 01:10:36 GMT Organization: University of Cambridge Computer Laboratory NNTP-Posting-Host: cam.ac.uk X-Newsreader: TIN [version 1.2 PL2] Lines: 10 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. -- Steve Okonkwo Tape trading list upon request. SASE required for international mail. From asokolov@clari.net Thu Sep 08 05:29:36 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!clari!asokolov From: asokolov@clari.net (Andrei Sokolov) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Sep08.052936.4262@clari.net> Date: Thu, 08 Sep 1994 05:29:36 GMT References: <1994Sep08.011036.9081@cam.ac.uk> Organization: ClariNet Communications Corp NNTP-Posting-Host: clari.net X-Newsreader: NewsWatcher 2.0b24 Lines: 9 In article <1994Sep08.011036.9081@cam.ac.uk>, sokonkwo@cam.ac.uk (Steve Okonkwo) 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. Real programmers write in C, not arcane DSP microcode. With Pentium chips hitting 100 MHz, dedicated DSPs are obsolete. -- Andrei Sokolov Tape trading list upon request. SASE required for international mail. From cbrennan@utexas.edu Fri Sep 09 13:01:36 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!utexas!cbrennan From: cbrennan@utexas.edu (Camila Brennan) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Sep09.130136.4976@utexas.edu> Date: Fri, 09 Sep 1994 13:01:36 GMT References: <1994Sep08.011036.9081@cam.ac.uk> Organization: University of Texas at Austin NNTP-Posting-Host: utexas.edu X-Newsreader: Netscape 1.1N Lines: 10 In article <1994Sep08.011036.9081@cam.ac.uk>, sokonkwo@cam.ac.uk (Steve Okonkwo) 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. -- Camila Brennan "Measuring a record you own is listening, not piracy." Atari Falcon 030 / TOS 4.04 From jmoreno@utexas.edu Sat Sep 10 04:42:36 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!utexas!jmoreno From: jmoreno@utexas.edu (Jen Moreno) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Sep10.044236.5456@utexas.edu> Date: Sat, 10 Sep 1994 04:42:36 GMT References: <1994Sep08.011036.9081@cam.ac.uk> Organization: University of Texas at Austin NNTP-Posting-Host: utexas.edu X-Newsreader: NewsWatcher 2.0b24 Lines: 8 In article <1994Sep08.011036.9081@cam.ac.uk>, sokonkwo@cam.ac.uk (Steve Okonkwo) 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. -- Jen Moreno -- University of Texas at Austin "5 sigma is a discovery. 7 is a question."