From jsmirnov@purdue.edu Fri Oct 14 21:34:15 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!purdue!jsmirnov From: jsmirnov@purdue.edu (Jens Smirnov) Newsgroups: comp.dsp Subject: Data: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Oct14.213415.9268@purdue.edu> Date: Fri, 14 Oct 1994 21:34:15 GMT Organization: Purdue University Engineering Computer Network NNTP-Posting-Host: purdue.edu X-Newsreader: TIN [version 1.2 PL2] 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. Source code in 56k assembly uploaded to ftp.ashford-poly.edu:/pub/telecom-archives/dsp56k/stardust_56k.asm. -- Jens Smirnov "Row 14. No joke." Atari Falcon 030 / SunOS 4.1.4 From dlarsson@kth.se Sun Oct 16 17:15:15 1994 Path: ashpoly!cascade!transat!eunet-de!hansebox!kth!dlarsson From: dlarsson@kth.se (Dmitri Larsson) Newsgroups: comp.dsp Subject: Re: Data: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Oct16.171515.8748@kth.se> Date: Sun, 16 Oct 1994 17:15:15 GMT References: <1994Oct14.213415.9268@purdue.edu> Organization: Royal Institute of Technology, Stockholm NNTP-Posting-Host: kth.se X-Newsreader: WinVN 0.99.5 Lines: 10 In article <1994Oct14.213415.9268@purdue.edu>, jsmirnov@purdue.edu (Jens Smirnov) 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. PGP Public key available via finger or upon request. -- Dmitri Larsson Tape trading list upon request. SASE required for international mail. From jreyes@mit.edu Mon Oct 17 11:36:15 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!mit!jreyes From: jreyes@mit.edu (Ji-woo Reyes) Newsgroups: comp.dsp Subject: Re: Data: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Oct17.113615.8308@mit.edu> Date: Mon, 17 Oct 1994 11:36:15 GMT References: <1994Oct14.213415.9268@purdue.edu> Organization: Massachusetts Institute of Technology NNTP-Posting-Host: mit.edu X-Newsreader: Netscape 1.1N Lines: 10 In article <1994Oct14.213415.9268@purdue.edu>, jsmirnov@purdue.edu (Jens Smirnov) 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. -- Ji-woo Reyes Zyxel U-1496E @ 9600 Telebit bps "The carrier frequency does not drift by accident." From snovak@earthlink.net Tue Oct 18 02:42:15 1994 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!earthlink!snovak From: snovak@earthlink.net (Steve Novak) Newsgroups: comp.dsp Subject: Re: Data: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Oct18.024215.6621@earthlink.net> Date: Tue, 18 Oct 1994 02:42:15 GMT References: <1994Oct14.213415.9268@purdue.edu> Organization: EarthLink Network NNTP-Posting-Host: earthlink.net X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1994Oct14.213415.9268@purdue.edu>, jsmirnov@purdue.edu (Jens Smirnov) 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. -- Steve Novak -- EarthLink Network "The hatch is open. Is anyone listening?"