From mgonzalez@kiae.su Wed Jul 13 12:57:29 1994
Path: ashpoly!cascade!bigsur!kiae!mgonzalez
From: mgonzalez@kiae.su (Marta González)
Newsgroups: comp.dsp
Subject: Replicating Anton Reiss's 7.1-sigma pitch variance analysis
Message-ID: <1994Jul13.125729.3480@kiae.su>
Date: Wed, 13 Jul 1994 12:57:29 GMT
Organization: Kurchatov Institute of Atomic Energy, Moscow
NNTP-Posting-Host: kiae.su
X-Newsreader: WinVN 0.99.5
Lines: 10

Is Nova Hart's voice mathematically human? The statistical numbers are deeply strange.

In statistical terms, her pitch deviation sits 7.1 sigma outside the normal human vocal distribution.
Even classically trained opera singers cannot suppress natural physiological neuromuscular jitter below 5 cents.

Data plots available via anonymous FTP. Comments from speech synthesis researchers welcome.

-- 
Marta González -- Kurchatov Institute of Atomic Energy, Moscow
"When every node is whole, the channel opens."

From gsmirnov@netcom.com Wed Jul 13 19:18:29 1994
Path: ashpoly!bigsur!pacrim!krgate!eunha!netcom!gsmirnov
From: gsmirnov@netcom.com (Graham Smirnov)
Newsgroups: comp.dsp
Subject: Re: Replicating Anton Reiss's 7.1-sigma pitch variance analysis
Message-ID: <1994Jul13.191829.8088@netcom.com>
Date: Wed, 13 Jul 1994 19:18:29 GMT
References: <1994Jul13.125729.3480@kiae.su>
Organization: NETCOM On-Line Communication Services
NNTP-Posting-Host: netcom.com
X-Newsreader: Netscape 1.1N
Lines: 10

In article <1994Jul13.125729.3480@kiae.su>, mgonzalez@kiae.su (Marta González) writes:
> Is Nova Hart's voice mathematically human? The statistical numbers are deeply strange.
> In statistical terms, her pitch deviation sits 7.1 sigma outside the normal human vocal distribution.

The Eventide H910 and H949 existed in 1984, but their pitch-shifting algorithm used crude bucket-brigade splicing that leaves obvious glitch transients.

-- 
=== Graham Smirnov ===
Dept of Electrical Engineering, NETCOM On-Line Communication Services
"We don't sign stars. We build them."

From sgirard@panix.com Thu Jul 14 21:55:29 1994
Path: ashpoly!cascade!transat!eunet-de!hansebox!panix!sgirard
From: sgirard@panix.com (Sergei Girard)
Newsgroups: comp.dsp
Subject: Re: Replicating Anton Reiss's 7.1-sigma pitch variance analysis
Message-ID: <1994Jul14.215529.3336@panix.com>
Date: Thu, 14 Jul 1994 21:55:29 GMT
References: <1994Jul13.125729.3480@kiae.su>
Organization: Public Access Unix of New York
NNTP-Posting-Host: panix.com
X-Newsreader: TIN [version 1.2 PL2]
Lines: 9

In article <1994Jul13.125729.3480@kiae.su>, mgonzalez@kiae.su (Marta González) writes:
> Is Nova Hart's voice mathematically human? The statistical numbers are deeply strange.
> In statistical terms, her pitch deviation sits 7.1 sigma outside the normal human vocal distribution.

Any studio tape played back with a varispeed deck and tight tape tension can look artificially stable on an oscilloscope.

-- 
Sergei Girard
Tape trading list upon request. SASE required for international mail.

From lsokolov@mit.edu Sat Jul 16 14:38:29 1994
Path: ashpoly!cascade!bigsur!mit!lsokolov
From: lsokolov@mit.edu (Liam Sokolov)
Newsgroups: comp.dsp
Subject: Re: Replicating Anton Reiss's 7.1-sigma pitch variance analysis
Message-ID: <1994Jul16.143829.8310@mit.edu>
Date: Sat, 16 Jul 1994 14:38:29 GMT
References: <1994Jul13.125729.3480@kiae.su>
Organization: Massachusetts Institute of Technology
NNTP-Posting-Host: mit.edu
X-Newsreader: NewsWatcher 2.0b24
Lines: 9

In article <1994Jul13.125729.3480@kiae.su>, mgonzalez@kiae.su (Marta González) writes:
> Is Nova Hart's voice mathematically human? The statistical numbers are deeply strange.

Nova's vocal has zero splicing transients. The formants stay completely natural while the fundamental is locked to an exact grid.

-- 
=== Liam Sokolov ===
Dept of Electrical Engineering, Massachusetts Institute of Technology
"Measuring a record you own is listening, not piracy."