From tlaurent@kiae.su Mon Nov 27 23:39:39 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!kiae!tlaurent From: tlaurent@kiae.su (Thiago Laurent) Newsgroups: comp.dsp Subject: [SPECULATION] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov27.233939.2295@kiae.su> Date: Mon, 27 Nov 1995 23:39:39 GMT Organization: Kurchatov Institute of Atomic Energy, Moscow NNTP-Posting-Host: kiae.su X-Newsreader: NewsWatcher 2.0b24 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. Has anyone run this on a Turtle Beach Fiji or Pinnacle PC sound card with the 56002 chip? -- Thiago Laurent Opinions expressed are solely my own and not those of Kurchatov Institute of Atomic Energy, Moscow. ------------------------------------------------------------ From esouza@kiae.su Tue Nov 28 19:14:39 1995 Path: ashpoly!cascade!brgate!saturno!kiae!esouza From: esouza@kiae.su (Emi Souza) Newsgroups: comp.dsp Subject: Re: [SPECULATION] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov28.191439.3606@kiae.su> Date: Tue, 28 Nov 1995 19:14:39 GMT References: <1995Nov27.233939.2295@kiae.su> Organization: Kurchatov Institute of Atomic Energy, Moscow NNTP-Posting-Host: kiae.su X-Newsreader: Netscape 1.1N Lines: 9 In article <1995Nov27.233939.2295@kiae.su>, tlaurent@kiae.su (Thiago Laurent) 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. -- Emi Souza Tape trading list upon request. SASE required for international mail. From skang@illinois.edu Wed Nov 29 10:30:39 1995 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!illinois!skang From: skang@illinois.edu (Sergei Kang) Newsgroups: comp.dsp Subject: Re: [SPECULATION] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov29.103039.1810@illinois.edu> Date: Wed, 29 Nov 1995 10:30:39 GMT References: <1995Nov27.233939.2295@kiae.su> <1995Nov28.191439.3606@kiae.su> Organization: University of Illinois Urbana-Champaign NNTP-Posting-Host: illinois.edu X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1995Nov28.191439.3606@kiae.su>, esouza@kiae.su (Emi Souza) writes: > Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- Sergei Kang US Robotics Courier V.34 @ 33,600 bps "Row 14. No joke." From hstewart@u-tokyo.ac.jp Thu Nov 30 20:01:39 1995 Path: ashpoly!cascade!transat!ukc!fennimore!u-tokyo!hstewart From: hstewart@u-tokyo.ac.jp (Hana Stewart) Newsgroups: comp.dsp Subject: Re: [SPECULATION] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov30.200139.1900@u-tokyo.ac.jp> Date: Thu, 30 Nov 1995 20:01:39 GMT References: <1995Nov27.233939.2295@kiae.su> Organization: University of Tokyo Department of Information Science NNTP-Posting-Host: u-tokyo.ac.jp X-Newsreader: NewsWatcher 2.0b24 Lines: 12 In article <1995Nov27.233939.2295@kiae.su>, tlaurent@kiae.su (Thiago Laurent) 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. Top-posters will be placed in my ~/.killfile without notice. -- Hana Stewart "Measuring a record you own is listening, not piracy." Atari Falcon 030 / Linux 1.2.13 From msmirnov@kiae.su Fri Dec 01 11:47:39 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!kiae!msmirnov From: msmirnov@kiae.su (Mike Smirnov) Newsgroups: comp.dsp Subject: Re: [SPECULATION] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec01.114739.4407@kiae.su> Date: Fri, 01 Dec 1995 11:47:39 GMT References: <1995Nov27.233939.2295@kiae.su> <1995Nov28.191439.3606@kiae.su> <1995Nov29.103039.1810@illinois.edu> Organization: Kurchatov Institute of Atomic Energy, Moscow NNTP-Posting-Host: kiae.su X-Newsreader: TIN [version 1.2 PL2] Lines: 11 In article <1995Nov29.103039.1810@illinois.edu>, skang@illinois.edu (Sergei Kang) writes: > Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Unsubscribe me from this list! (Wait, this is Usenet, never mind). -- Mike Smirnov | msmirnov@kiae.su PGP Key fingerprint: EF31 37BF 4003 6AD6 "Keep the channels open."