From hvasiliev@purdue.edu Tue May 03 14:11:16 1994 Path: ashpoly!bigsur!pacrim!jpgate!kamome!purdue!hvasiliev From: hvasiliev@purdue.edu (Hugo Vasiliev) Newsgroups: comp.dsp Subject: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May03.141116.9089@purdue.edu> Date: Tue, 03 May 1994 14:11:16 GMT Organization: Purdue University Engineering Computer Network NNTP-Posting-Host: purdue.edu X-Newsreader: WinVN 0.99.5 Lines: 10 Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. 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. -- Hugo Vasiliev -- Purdue University Engineering Computer Network "We don't sign stars. We build them." From msuzuki@cmu.edu Thu May 05 04:13:16 1994 Path: ashpoly!cascade!transat!eunet-fr!teletel!cmu!msuzuki From: msuzuki@cmu.edu (Mateo Suzuki) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May05.041316.4630@cmu.edu> Date: Thu, 05 May 1994 04:13:16 GMT References: <1994May03.141116.9089@purdue.edu> Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: Netscape 1.1N Lines: 10 In article <1994May03.141116.9089@purdue.edu>, hvasiliev@purdue.edu (Hugo Vasiliev) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. > By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Mateo Suzuki Opinions expressed are solely my own and not those of Carnegie Mellon University School of CS. ------------------------------------------------------------ From dsmirnov@pipeline.com Thu May 05 07:08:16 1994 Path: ashpoly!cascade!transat!sunet!nordrelay!pipeline!dsmirnov From: dsmirnov@pipeline.com (Dave Smirnov) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May05.070816.2110@pipeline.com> Date: Thu, 05 May 1994 07:08:16 GMT References: <1994May03.141116.9089@purdue.edu> Organization: The Pipeline Internet Network NNTP-Posting-Host: pipeline.com X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1994May03.141116.9089@purdue.edu>, hvasiliev@purdue.edu (Hugo Vasiliev) 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. -- Dave Smirnov "Extraordinary claims require extraordinary spectra." Pentium 90 (OS/2 Warp) / TOS 4.04 From moliveira@mindspring.com Fri May 06 20:37:16 1994 Path: ashpoly!cascade!mxgate!coyoacan!mindspring!moliveira From: moliveira@mindspring.com (Morag Oliveira) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May06.203716.4066@mindspring.com> Date: Fri, 06 May 1994 20:37:16 GMT References: <1994May03.141116.9089@purdue.edu> <1994May05.070816.2110@pipeline.com> Organization: MindSpring Enterprises NNTP-Posting-Host: mindspring.com X-Newsreader: Netscape 1.1N Lines: 9 In article <1994May05.070816.2110@pipeline.com>, dsmirnov@pipeline.com (Dave Smirnov) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Morag Oliveira *PLONK* filter active for trolls. Sun 3/60 (SunOS 4.1.1) - NetBSD 1.0 From scampbell@berkeley.edu Sun May 08 14:53:16 1994 Path: ashpoly!cascade!mxgate!coyoacan!berkeley!scampbell From: scampbell@berkeley.edu (Sun-mi Campbell) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994May08.145316.5593@berkeley.edu> Date: Sun, 08 May 1994 14:53:16 GMT References: <1994May03.141116.9089@purdue.edu> <1994May05.070816.2110@pipeline.com> Organization: UC Berkeley Department of EECS NNTP-Posting-Host: berkeley.edu X-Newsreader: WinVN 0.99.5 Lines: 11 In article <1994May05.070816.2110@pipeline.com>, dsmirnov@pipeline.com (Dave Smirnov) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Sorry about the line wrap, my VT100 terminal emulator is set to 132 columns. -- Sun-mi Campbell | scampbell@berkeley.edu PGP Key fingerprint: AC24 FA74 C90C 8456 "Keep the channels open."