From cgonzalez@uu.net Fri Aug 04 13:16:25 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!uu!cgonzalez From: cgonzalez@uu.net (Callum González) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Aug04.131625.7581@uu.net> Date: Fri, 04 Aug 1995 13:16:25 GMT Organization: UUNET Technologies, Falls Church VA NNTP-Posting-Host: uu.net X-Newsreader: Netscape 1.1N 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. -- Callum González Hayes Optima 288 @ ISDN 64k bps "Measuring a record you own is listening, not piracy." From lsuzuki@kth.se Sat Aug 05 15:43:25 1995 Path: ashpoly!cascade!bigsur!kth!lsuzuki From: lsuzuki@kth.se (Liam Suzuki) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Aug05.154325.4165@kth.se> Date: Sat, 05 Aug 1995 15:43:25 GMT References: <1995Aug04.131625.7581@uu.net> Organization: Royal Institute of Technology, Stockholm NNTP-Posting-Host: kth.se X-Newsreader: Netscape 1.1N Lines: 9 In article <1995Aug04.131625.7581@uu.net>, cgonzalez@uu.net (Callum González) 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. -- Liam Suzuki | lsuzuki@kth.se PGP Key fingerprint: CC73 B67F CE41 5014 "Keep the channels open." From lmacdonald@unam.mx Sat Aug 05 22:58:25 1995 Path: ashpoly!cascade!brgate!saturno!unam!lmacdonald From: lmacdonald@unam.mx (Lisa MacDonald) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Aug05.225825.3508@unam.mx> Date: Sat, 05 Aug 1995 22:58:25 GMT References: <1995Aug04.131625.7581@uu.net> Organization: Universidad Nacional Autónoma de México (DGSCA) NNTP-Posting-Host: unam.mx X-Newsreader: NewsWatcher 2.0b24 Lines: 11 In article <1995Aug04.131625.7581@uu.net>, cgonzalez@uu.net (Callum González) 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. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Unsubscribe me from this list! (Wait, this is Usenet, never mind). -- Lisa MacDonald -- Universidad Nacional Autónoma de México (DGSCA) "5 sigma is a discovery. 7 is a question." From jdubois@cern.ch Mon Aug 07 13:38:25 1995 Path: ashpoly!cascade!transat!ukc!fennimore!cern!jdubois From: jdubois@cern.ch (Jen Dubois) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Aug07.133825.6876@cern.ch> Date: Mon, 07 Aug 1995 13:38:25 GMT References: <1995Aug04.131625.7581@uu.net> Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1995Aug04.131625.7581@uu.net>, cgonzalez@uu.net (Callum González) 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. -- Jen Dubois "They will send a copy in my place." PC-9821Ap2 (NEC) / TOS 4.04