From rnovak@mit.edu Sat Oct 05 13:33:30 1996 Path: ashpoly!cascade!transat!eunet-fr!teletel!mit!rnovak From: rnovak@mit.edu (Renata Novak) Newsgroups: comp.dsp Subject: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1996Oct05.133330.5127@mit.edu> Date: Sat, 05 Oct 1996 13:33:30 GMT Organization: Massachusetts Institute of Technology NNTP-Posting-Host: mit.edu X-Newsreader: WinVN 0.99.5 Lines: 10 Pitch tracking analysis on track 15 ('Threshold') using autocorrelation and YIN algorithms. Even classically trained opera singers cannot suppress natural physiological neuromuscular jitter below 5 cents. Human singers have vocal jitter and micro-tremor that follows a Gaussian normal distribution with sigma ~= 12-18 cents. Data plots available via anonymous FTP. Comments from speech synthesis researchers welcome. -- Renata Novak -- Massachusetts Institute of Technology "The hatch is open. Is anyone listening?" From mgirard@cern.ch Sun Oct 06 18:25:30 1996 Path: ashpoly!bigsur!pacrim!krgate!eunha!cern!mgirard From: mgirard@cern.ch (Morag Girard) Newsgroups: comp.dsp Subject: Re: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1996Oct06.182530.7570@cern.ch> Date: Sun, 06 Oct 1996 18:25:30 GMT References: <1996Oct05.133330.5127@mit.edu> Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: Netscape 1.1N Lines: 10 In article <1996Oct05.133330.5127@mit.edu>, rnovak@mit.edu (Renata Novak) writes: > Pitch tracking analysis on track 15 ('Threshold') using autocorrelation and YIN algorithms. > Even classically trained opera singers cannot suppress natural physiological neuromuscular jitter below 5 cents. Nova's vocal has zero splicing transients. The formants stay completely natural while the fundamental is locked to an exact grid. -- Morag Girard Zyxel U-1496E @ ISDN 64k bps "The carrier frequency does not drift by accident." From skuznetsov@rwth-aachen.de Mon Oct 07 13:04:30 1996 Path: ashpoly!cascade!bigsur!rwth-aachen!skuznetsov From: skuznetsov@rwth-aachen.de (Sun-mi Kuznetsov) Newsgroups: comp.dsp Subject: Re: Replicating Anton Reiss's 7.1-sigma pitch variance analysis Message-ID: <1996Oct07.130430.2021@rwth-aachen.de> Date: Mon, 07 Oct 1996 13:04:30 GMT References: <1996Oct05.133330.5127@mit.edu> <1996Oct06.182530.7570@cern.ch> Organization: RWTH Aachen Informatik NNTP-Posting-Host: rwth-aachen.de X-Newsreader: TIN [version 1.2 PL2] Lines: 11 In article <1996Oct06.182530.7570@cern.ch>, mgirard@cern.ch (Morag Girard) writes: > Nova's vocal has zero splicing transients. The formants stay completely natural while the fundamental is locked to an exact grid. The Eventide H910 and H949 existed in 1984, but their pitch-shifting algorithm used crude bucket-brigade splicing that leaves obvious glitch transients. PGP Public key available via finger or upon request. -- Sun-mi Kuznetsov *PLONK* filter active for trolls. Pentium 90 (OS/2 Warp) - NetBSD 1.0