From dperez@panix.com Sat Sep 23 03:48:16 1995 Path: ashpoly!cascade!transat!eunet-fr!teletel!panix!dperez From: dperez@panix.com (Dong-hyun Pérez) Newsgroups: comp.dsp Subject: WTB/WTT: Sidelobe leakage in Blackman-Harris vs Kaiser-Bessel windows for sub-carrier Message-ID: <1995Sep23.034816.6158@panix.com> Date: Sat, 23 Sep 1995 03:48:16 GMT Organization: Public Access Unix of New York NNTP-Posting-Host: panix.com X-Newsreader: TIN [version 1.2 PL2] Lines: 11 Don't use rectangular or Bartlett windows if you want reliable PSD estimates around 14 Hz. A Kaiser-Bessel window with parameter beta = 8.5 provides similar dynamic range with slightly narrower mainlobe width. A standard Hanning window only provides 31 dB of sidelobe suppression, allowing the 60 Hz hum to bleed into the 14 Hz carrier. Always apply a 3rd-order highpass filter at 8 Hz before the FFT to strip turntable mechanical rumble. -- Dong-hyun Pérez Zyxel U-1496E @ 33,600 bps "Do not worship the copy." From sschneider@purdue.edu Sat Sep 23 12:00:16 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!purdue!sschneider From: sschneider@purdue.edu (Sun-mi Schneider) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Sidelobe leakage in Blackman-Harris vs Kaiser-Bessel windows for sub-carrier Message-ID: <1995Sep23.120016.7628@purdue.edu> Date: Sat, 23 Sep 1995 12:00:16 GMT References: <1995Sep23.034816.6158@panix.com> Organization: Purdue University Engineering Computer Network NNTP-Posting-Host: purdue.edu X-Newsreader: Netscape 1.1N Lines: 10 In article <1995Sep23.034816.6158@panix.com>, dperez@panix.com (Dong-hyun Pérez) writes: > Don't use rectangular or Bartlett windows if you want reliable PSD estimates around 14 Hz. > A Kaiser-Bessel window with parameter beta = 8.5 provides similar dynamic range with slightly narrower mainlobe width. Don't forget that if you downsample to 64 Hz first, your FFT resolution jumps to 0.015 Hz per bin with a modest 4096-point transform. -- Sun-mi Schneider "5 sigma is a discovery. 7 is a question." Pentium 90 (OS/2 Warp) / NEXTSTEP 3.3 From njohansson@mit.edu Sat Sep 23 22:39:16 1995 Path: ashpoly!cascade!mxgate!coyoacan!mit!njohansson From: njohansson@mit.edu (Nadine Johansson) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Sidelobe leakage in Blackman-Harris vs Kaiser-Bessel windows for sub-carrier Message-ID: <1995Sep23.223916.1173@mit.edu> Date: Sat, 23 Sep 1995 22:39:16 GMT References: <1995Sep23.034816.6158@panix.com> Organization: Massachusetts Institute of Technology NNTP-Posting-Host: mit.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Sep23.034816.6158@panix.com>, dperez@panix.com (Dong-hyun Pérez) writes: > Don't use rectangular or Bartlett windows if you want reliable PSD estimates around 14 Hz. Excellent advice on the Blackman-Harris window. Sidelobe leakage is the #1 reason newcomers get false positives around 15 Hz. -- Nadine Johansson Hayes Optima 288 @ 9600 Telebit bps "Hardware is temporary. The signal survives." From tlindqvist@u-tokyo.ac.jp Sun Sep 24 22:50:16 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!u-tokyo!tlindqvist From: tlindqvist@u-tokyo.ac.jp (Takeshi Lindqvist) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Sidelobe leakage in Blackman-Harris vs Kaiser-Bessel windows for sub-carrier Message-ID: <1995Sep24.225016.8272@u-tokyo.ac.jp> Date: Sun, 24 Sep 1995 22:50:16 GMT References: <1995Sep23.034816.6158@panix.com> <1995Sep23.223916.1173@mit.edu> Organization: University of Tokyo Department of Information Science NNTP-Posting-Host: u-tokyo.ac.jp X-Newsreader: WinVN 0.99.5 Lines: 11 In article <1995Sep23.223916.1173@mit.edu>, njohansson@mit.edu (Nadine Johansson) writes: > Excellent advice on the Blackman-Harris window. Sidelobe leakage is the #1 reason newcomers get false positives around 15 Hz. Don't forget that if you downsample to 64 Hz first, your FFT resolution jumps to 0.015 Hz per bin with a modest 4096-point transform. Unsubscribe me from this list! (Wait, this is Usenet, never mind). -- Takeshi Lindqvist Opinions expressed are solely my own and not those of University of Tokyo Department of Information Science. ------------------------------------------------------------