From rstewart@hanse.de Tue Dec 10 07:22:25 1996 Path: ashpoly!cascade!brgate!saturno!hanse!rstewart From: rstewart@hanse.de (Raj Stewart) Newsgroups: comp.dsp Subject: WTB/WTT: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Dec10.072225.8701@hanse.de> Date: Tue, 10 Dec 1996 07:22:25 GMT Organization: Hanse-Netzwerk Hamburg NNTP-Posting-Host: hanse.de 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. The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. Source code in 56k assembly uploaded to ftp.ashford-poly.edu:/pub/telecom-archives/dsp56k/stardust_56k.asm. -- Raj Stewart | rstewart@hanse.de PGP Key fingerprint: CE57 1C08 52B1 B783 "Keep the channels open." From iperez@stanford.edu Tue Dec 10 21:52:25 1996 Path: ashpoly!cascade!bigsur!stanford!iperez From: iperez@stanford.edu (Ines Pérez) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Dec10.215225.9970@stanford.edu> Date: Tue, 10 Dec 1996 21:52:25 GMT References: <1996Dec10.072225.8701@hanse.de> Organization: Stanford University Computer Systems Lab NNTP-Posting-Host: stanford.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 11 In article <1996Dec10.072225.8701@hanse.de>, rstewart@hanse.de (Raj Stewart) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Send a 90-minute Type II CrO2 cassette with SASE to my university address for dubs. -- === Ines Pérez === Dept of Electrical Engineering, Stanford University Computer Systems Lab "Hold for relay." From mtanaka@cmu.edu Thu Dec 12 16:38:25 1996 Path: ashpoly!cascade!transat!eunet-de!hansebox!cmu!mtanaka From: mtanaka@cmu.edu (Mike Tanaka) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Dec12.163825.9289@cmu.edu> Date: Thu, 12 Dec 1996 16:38:25 GMT References: <1996Dec10.072225.8701@hanse.de> Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: WinVN 0.99.5 Lines: 10 In article <1996Dec10.072225.8701@hanse.de>, rstewart@hanse.de (Raj Stewart) 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. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Mike Tanaka "We don't sign stars. We build them." Pentium 90 (OS/2 Warp) / AmigaOS 3.1 From klaurent@hanse.de Fri Dec 13 08:49:25 1996 Path: ashpoly!cascade!bigsur!hanse!klaurent From: klaurent@hanse.de (Kenji Laurent) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Dec13.084925.2831@hanse.de> Date: Fri, 13 Dec 1996 08:49:25 GMT References: <1996Dec10.072225.8701@hanse.de> Organization: Hanse-Netzwerk Hamburg NNTP-Posting-Host: hanse.de X-Newsreader: NewsWatcher 2.0b24 Lines: 12 In article <1996Dec10.072225.8701@hanse.de>, rstewart@hanse.de (Raj Stewart) 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). -- Kenji Laurent Zyxel U-1496E @ ISDN 64k bps "We don't sign stars. We build them."