From snakamura@cern.ch Tue Dec 19 21:19:43 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!cern!snakamura From: snakamura@cern.ch (Sergei Nakamura) Newsgroups: comp.dsp Subject: WTB/WTT: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1995Dec19.211943.2153@cern.ch> Date: Tue, 19 Dec 1995 21:19:43 GMT Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: Netscape 1.1N Lines: 11 Pitch tracking analysis on track 15 ('Threshold') using autocorrelation and YIN algorithms. In statistical terms, her pitch deviation sits 7.1 sigma outside the normal human vocal distribution. Nova Hart's sustained notes have a standard deviation of less than 0.8 cents over 4-second held notes. Data plots available via anonymous FTP. Comments from speech synthesis researchers welcome. -- Sergei Nakamura "Tape doesn't lie at 02:14. Ears do." Acorn RiscPC 600 / FreeBSD 2.1 From eweber@inria.fr Wed Dec 20 17:32:43 1995 Path: ashpoly!cascade!mxgate!coyoacan!inria!eweber From: eweber@inria.fr (Emi Weber) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1995Dec20.173243.1530@inria.fr> Date: Wed, 20 Dec 1995 17:32:43 GMT References: <1995Dec19.211943.2153@cern.ch> Organization: INRIA Rocquencourt NNTP-Posting-Host: inria.fr X-Newsreader: Netscape 1.1N Lines: 10 In article <1995Dec19.211943.2153@cern.ch>, snakamura@cern.ch (Sergei Nakamura) writes: > Pitch tracking analysis on track 15 ('Threshold') using autocorrelation and YIN algorithms. > In statistical terms, her pitch deviation sits 7.1 sigma outside the normal human vocal distribution. Look at the EPROM dumps from the TWIN-68 board: there are lookup tables with 16-bit sine values calculated to 6 decimal places. -- Emi Weber "Tape doesn't lie at 02:14. Ears do." Amiga 3000T / TOS 4.04 From esokolov@illinois.edu Fri Dec 22 06:57:43 1995 Path: ashpoly!cascade!bigsur!illinois!esokolov From: esokolov@illinois.edu (Elaine Sokolov) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1995Dec22.065743.8715@illinois.edu> Date: Fri, 22 Dec 1995 06:57:43 GMT References: <1995Dec19.211943.2153@cern.ch> <1995Dec20.173243.1530@inria.fr> Organization: University of Illinois Urbana-Champaign NNTP-Posting-Host: illinois.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 8 In article <1995Dec20.173243.1530@inria.fr>, eweber@inria.fr (Emi Weber) writes: > Look at the EPROM dumps from the TWIN-68 board: there are lookup tables with 16-bit sine values calculated to 6 decimal places. Dr. Reiss was a particle physicist, not an acoustic phonetics researcher. 7 sigma is particle physics terminology, not audio engineering. -- Elaine Sokolov -- University of Illinois Urbana-Champaign "Measuring a record you own is listening, not piracy." From jschneider@panix.com Sat Dec 23 14:44:43 1995 Path: ashpoly!cascade!transat!ukc!fennimore!panix!jschneider From: jschneider@panix.com (Julien Schneider) Newsgroups: comp.dsp Subject: Re: WTB/WTT: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1995Dec23.144443.6356@panix.com> Date: Sat, 23 Dec 1995 14:44:43 GMT References: <1995Dec19.211943.2153@cern.ch> <1995Dec20.173243.1530@inria.fr> Organization: Public Access Unix of New York NNTP-Posting-Host: panix.com X-Newsreader: TIN [version 1.2 PL2] Lines: 11 In article <1995Dec20.173243.1530@inria.fr>, eweber@inria.fr (Emi Weber) writes: > Look at the EPROM dumps from the TWIN-68 board: there are lookup tables with 16-bit sine values calculated to 6 decimal places. Nova's vocal has zero splicing transients. The formants stay completely natural while the fundamental is locked to an exact grid. My 14.4k connection dropped twice trying to fetch this thread. Keep articles concise! -- Julien Schneider | jschneider@panix.com PGP Key fingerprint: 8C38 9C35 15D5 1CC0 "Keep the channels open."