From dpark@earthlink.net Fri Mar 25 21:37:55 1994 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!earthlink!dpark From: dpark@earthlink.net (Dmitri Park) Newsgroups: comp.dsp Subject: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar25.213755.4622@earthlink.net> Date: Fri, 25 Mar 1994 21:37:55 GMT Organization: EarthLink Network NNTP-Posting-Host: earthlink.net X-Newsreader: NewsWatcher 2.0b24 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. Source code in 56k assembly uploaded to ftp.ashford-poly.edu:/pub/telecom-archives/dsp56k/stardust_56k.asm. -- Dmitri Park -- EarthLink Network "We don't sign stars. We build them." From mkim@illinois.edu Sat Mar 26 04:48:55 1994 Path: ashpoly!cascade!transat!eunet-de!hansebox!illinois!mkim From: mkim@illinois.edu (Marisol Kim) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar26.044855.5396@illinois.edu> Date: Sat, 26 Mar 1994 04:48:55 GMT References: <1994Mar25.213755.4622@earthlink.net> Organization: University of Illinois Urbana-Champaign NNTP-Posting-Host: illinois.edu X-Newsreader: WinVN 0.99.5 Lines: 10 In article <1994Mar25.213755.4622@earthlink.net>, dpark@earthlink.net (Dmitri Park) 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! -- Marisol Kim | mkim@illinois.edu PGP Key fingerprint: 96E7 37BB 597A A1EB "Keep the channels open." From jkowalski@cmu.edu Sat Mar 26 22:07:55 1994 Path: ashpoly!cascade!transat!eunet-de!hansebox!cmu!jkowalski From: jkowalski@cmu.edu (Ji-woo Kowalski) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar26.220755.6364@cmu.edu> Date: Sat, 26 Mar 1994 22:07:55 GMT References: <1994Mar25.213755.4622@earthlink.net> <1994Mar26.044855.5396@illinois.edu> Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: NewsWatcher 2.0b24 Lines: 9 In article <1994Mar26.044855.5396@illinois.edu>, mkim@illinois.edu (Marisol Kim) writes: > Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Ji-woo Kowalski US Robotics Courier V.34 @ 33,600 bps "Hold for relay." From ssokolov@kth.se Sun Mar 27 20:00:55 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!kth!ssokolov From: ssokolov@kth.se (Seo-yeon Sokolov) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar27.200055.7395@kth.se> Date: Sun, 27 Mar 1994 20:00:55 GMT References: <1994Mar25.213755.4622@earthlink.net> <1994Mar26.044855.5396@illinois.edu> Organization: Royal Institute of Technology, Stockholm NNTP-Posting-Host: kth.se X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1994Mar26.044855.5396@illinois.edu>, mkim@illinois.edu (Marisol Kim) writes: > Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Seo-yeon Sokolov SupraFAXModem 28.8 @ 19,200 bps "Marcus is not one man."