Transcription of Recordings - p. 4
SSSHP 82: "THE HUMAN VOICE AND THE COMPUTER, IEEE SOUNDINGS,
No. 70-S-04, Ed. by Dr. Walter R. Beam, AUG 1, 1971"
SOURCE: Donated by H.D. Maxey, from the IBM tape archives.
Commercially available cassette, fair quality.
CONTENTS: Survey tape prepared by Dr. Walter R. Beam, for the
Institute of Electrical and Electronics Engineers, Inc. See SSSHP
company outlines for associated technical details. Document
accompanying this cassette is in SSSHP 98 Reprints, SSSHP USA
BTL file. Copyright by IEEE and the speakers.
1. INTRODUCTION by Dr. Beam, Editor
Dr. J.L. Flanagan, Bell Telephone Laboratories, using an
artificial larynx.
(human: "IEEE Soundings examines the human voice and the
computer.")
Articulatory synthesis by C.H. Coker, Bell Telephone
Laboratories.
(syn: "I am a computer. I can read stories and speak them
aloud. I do not understand what the words mean when I read
them, but I can guess which words are important and which words
are not, by rules I have been given. Someday, I may be able to
provide many kinds of information by telephone.")
2. SYNTHESIS FROM ENGLISH-LANGUAGE TEXTS, Prof. J. Allen, Dept. of
Electrical Engineering, Massachusetts Institute of Technology.
Discussion of use of morphological decomposition of words for
text to speech applications. Example using Klatt synthesis.
(syn: "This paper presents a method for synthesizing speech
from text. Because this technique uses both a dictionary and
letter to sound rules, it can convert unrestricted text to
speech. Information retrieval systems and reading machines for
the blind are some of the applications for this process.")
3. PROGRESS TOWARD A READING MACHINE FOR THE BLIND, Dr. F.S.
Cooper, Haskins Laboratories, New Haven, Connecticut. Synthesis
by rule from a phonetic input, using a formant synthesizer. I.
Mattingly rules for synthesis. Speaking rate increases from 130
to 220 words/minute during the story. Input specification of
phonetic spelling, phrase endings, and the sentence stress.
(syn: "The North Wind and the Sun were arguing one day ...
wrapped in a warm coat ... Sun was the stronger of the two.")
Automatic text-to-speech by rule, manually using a pronouncing
dictionary for the phonetic spelling, N. Umeda's rules for
juncture and stress, and I. Mattingly's synthesis by rule.
Circa mid-1970.
(syn, "Summary of Research in Animal Communication, by Mark
Twain. Animals talk to each other, of course, there can be no
question about that. But I suppose there are very few people
who can understand them. I never knew but one man who could. I
know he could, however, because he told me so himself. He was a
middle-aged, simple-hearted miner, who had lived in a lonely
corner of California, among the woods and mountains, a good
many years, and had studied the ways of his only neighbors, the
beasts and the birds, until he believed he could accurately
translate any remark which they made. This was Jim Baker.
According to Jim Baker, the blue jays were the best talkers he
had found among the birds and beasts. Said he, 'There is more
to a blue jay than any other creature. Whatever a blue jay
feels, he can put it into language, and no mere common place
language, either. And as for command of language, why you never
see a blue jay get stuck for a word.'")
Synthesis by rule for field trials with students and blind
veterans, 1971.
(syn: "You will now hear a few pages from Steinbeck's Travels
With Charlie. When I was very young, and the urge to be some
place was upon me, I was assured by mature people that maturity
would cure this itch. When years described me as mature, the
remedy prescribed was middle age. In middle age, I was assured
that greater age would calm my fever. And, now that I am 58,
perhaps senility will do the job. Nothing has worked.")
4. LOW-BIT-RATE CODING OF VOICE MESSAGES FOR ANSWER-BACK
APPLICATIONS, L.R. Rabiner, Bell Telephone Laboratories,
Murray Hill, New Jersey. Analysis and resynthesis of stored
words for answer-back applications. Coding of human speech in
formant synthesis parameters. Rabiner, L.R., R.W. Schafer, and
J.L. Flanagan, "Computer synthesis of speech by concatenation
of formant-coded words", Bell System Tech. J., 50, 1541-1558
(1971).
(syn, various smoothing: "We were away a year ago.")
(syn, various quantizing: "We were away a year ago.")
(syn, changing parameters: "We were away a year ago.")
(syn, separate words: "We. Were. Away. A. Year. Ago.")
(syn, combined: "We were away a year ago.")
(syn, word assembly, without and with juncture smoothing: "The
number is 091-1909. The number is 135-3201")
5. EXPLANATION OF SOME TERMS AND CONCEPTS, J.L. Flanagan, Bell
Telephone Laboratories, Murray Hill, New Jersey. BTL uses a
hardware formant synthesizer, connected to a DPD516 computer
for control. The DPD516 computer has 16K bytes of memory and a
960 nanosecond cycle. Synthesis is ten times real time (could
service ten synthesizers.)
6. SYNTHESIS USING "DIPHONE" SOUND COMPONENTS, W.D. Chapman,
IBM Systems Development Laboratories, Research Triangle Park,
North Carolina. Present library contains about 600 diphones.
Manual specification of phonetic spelling, durations, and
intonation contour.
(syn: "Office. Being. Seven. Consult. Taken. Number. Changed.
Mistake. Working. W. The number you dialed, ME1-5280, has been
changed. The new number is PA6-1347. This is a recording. The
number you dialed, UN4-4482, has been temporarily disconnected.
Thank you. I'm sorry, you've reached this office by mistake.
Please consult your directory and dial again.")
7. MODEL OF THE MOUTH USED FOR SYNTHESIS CALCULATIONS, Dr. C.H.
Coker, Bell Telephone Laboratories, Murray Hill, New Jersey.
Articulatory synthesis by rule. Synthesis demonstration at end
of tape.
8. COMMERCIALLY AVAILABLE AUDIO RESPONSE SYSTEMS, W.D. Chapman,
IBM Systems Development Laboratories, Research Triangle Park,
North Carolina. Description of IBM products 7770 (telephone
intercept messages from stored words on a magnetic drum) and
7772 (hardware vocoder operated from computer-stored data at
an average of 2400 bits per second.) No synthesis demonstration.
9. A DISCUSSION OF FUTURE APPLICATIONS IN SPEECH ANALYSIS, Dr.
J.L. Flanagan, Bell Telephone Laboratories, Murray Hill, New
Jersey, and Dr. W.R. Beam, Editor. Speaker verification and
recognition. Closing synthesis by Coker articulatory
synthesizer.
(syn: "This is the computer talking. It's been a pleasure
communicating with you, both through visual displays, and now
with computer synthesized speech.")
END: **********
SSSHP 83: "SOME REMINISCENCES ON SPEECH RESEARCH, F.S. Cooper
4/25/72, '72 SCP".
SOURCE: Donated by H.D. Maxey, Aug 23, 1990. Copied from
plastic diskette with IEEE Trans. Acoust. & Electro., Vol AU-21,
No. 3, June 1973. See SSSHP USA Haskins Laboratories file.
Cassette, some stylus noise.
CONTENTS:
Dr. Cooper's reminiscences of the development of speech
synthesis, on the occasion of his receiving the 1972 Award of
Pioneer in Speech Communication, from the IEEE Audio and
Electroacoustics Group.
Haskins Laboratories started a project of building a reading
machine for the blind in 1944. They built their own sound
spectrograph (not commercially available at that time) and a
synthesizer (Pattern Playback) that could play sound
spectrograms.
1. First sentence produced on Pattern Playback.
(syn: "Eat at Joe's.")
2. Pattern Playback in "transmission mode", using a photographic
transparency of a sound spectrogram of human speech.
(syn: "A large size in stockings is hard to sell.")
3. Pattern Playback, using a hand-painted simplification of the
above sound spectrogram.
(syn: "A large size in stockings is hard to sell.")
4. Pattern Playback, using transmission mode, then reflection or
painted mode of a careful copy of the spectrogram, then
reflection mode of a simplified copy of the spectrogram.
(syn, 3 variations: "Never kill a snake.")
5. Pattern Playback, using stylized patterns characterizing the
acoustic cues of stop and nasal consonants.
(syn: "Ba, da, ga. Pa, ta, ka. Am, an, ang.")
6. Pattern Playback, synthesis by rule, by P. Delattre.
(syn: "Oh, my aching back.")
7. Pattern Playback, synthesis of music called "Scotch Plaid", by
P. Delattre.
(syn, music:)
Discussion of work by other laboratories in 1940s and 1950s.
First speech conference, the "Half Conference" at MIT in late
1949, then a "full dress" conference at MIT in the spring of 1950
(see proceedings of J. Acoust. Soc. America for 1950.) Later
conferences, also at MIT, occurred in 1952 and 1956. In 1955 and
1956, speech conferences were held at Signals Research and
Development Establishment (SRDE) at Christchurch, England, and at
NEL in San Diego, California.
In 1956, there was an International Congress of Acoustics at MIT.
Under discussion were:
- phoneme as a bundle of distinctive features
- search for acoustic invariants of phonemes, for a "phonetic
typewriter" or efficient bandwidth compression device
- better methods of synthesis
- methods of formant tracking
8. Summer meetings at MIT, Cambridge, in 1956, included
demonstrations of Fant's OVE-I and Lawrence's PAT hardware
synthesizers, in a dialog. (Recordings obtained by Cooper from
Fant and Lawrence.)
(OVE: "How are you?"; PAT: "What did you say before that?")
(OVE: "I love you."; PAT: "What did you say before that?")
(OVE: "How are you?"; PAT, normal and whispered: "What did you
say before that?")
(OVE: "I love you."; PAT, singing: "What did you say before
that?")
At a later demonstration, Lawrence was the subject of a friendly
spoof of his synthesizer, by J.C.R. Licklighter. After Lawrence
had finished his demonstration, turned off the equipment, and sat
down, a hidden speaker suddenly burst forth with a torrent of
"What did you say before that?", seemingly from the unpowered PAT.
9. Octopus formant synthesizer, Haskins Laboratories.
(syn: "No comment.")
10. The Dynamic Vocal Tract synthesizer, DAVO, at MIT.
(syn: "This is the voice of DAVO, at MIT.")
11. Voback, Haskins Laboratories. Use of a plastic belt pattern
to control a vocoder synthesizer. Synthesis by rule, by P.
Delattre.
(syn: "Alexander's an intelligent conversationalist.")
12. OVE-II at the Royal Institute of Technology, Stockholm.
Careful copying of human spectrum by J. Holmes.
(syn and human: "I enjoy the simple life.")
13. Pattern Playback, early synthesis by rule, by P. Delattre,
using first rules for vowels and stop consonants.
(syn: "A big bad man, demanding money, can kill you. Bang,
bang.")
14. Pattern Playback, later synthesis by rule, by F. Ingemann.
(syn: "I painted this by rule without looking at a
spectrogram. Can you understand it?")
15. Holmes and Mattingly at Joint Speech Research Unit (JSRU) in
England. Synthesis by rule.
(syn: "Someone, somewhere, wants a letter from you.")
16. Haskins formant synthesis by rule, by J. Gaitenby.
(syn: "Silence is golden. Nevertheless, speech synthesized by
rule has certain merits.")
END: **********
SSSHP 84: "PAT - SYNTHESIZED INTONATIONS. B: 'HE'LL BE HERE ON
FRIDAY'."
SOURCE: Donated by Mrs E.T. Uldall, Univ. of Edinburgh,
Scotland, 5/9/90. See SSSHP UK Edinburgh Univ. file. Fourteen
23" tape loops in 3 3/4" by 1 1/2" circular tin box.
CONTENTS:
A synthesized sentence with fourteen different intonation
patterns, each on its own tape loop, circa 1962. These were
copied to a cassette (SSSHP 94) by H.D. Maxey, January, 1991.
(syn, 14 intonation patterns: "He'll be here on Friday.")
END: **********
SSSHP 86: "TUm - SYNTHESIS FROM NEUROMOTOR COMMAND, Research
Institute of Electrical Communication, Tohoku Univ."
SOURCE: Contributed by Dr. Shizuo Hiki, School of Human Sciences,
Waseda Univ., Tokorozawa, Japan, September 23, 1990. See SSSHP
JAP Tohoku Univ. file. 7" reel, 0 to +3 vu, good quality, left
channel only.
CONTENTS: Narration by S. Hiki.
Copy of tape TUm 1974.7.21. Recorded examples of the synthetic
speech from neuromotor commands. Ten Japanese words consisting of
vowels and semi-vowels. Synthesis by a computer-programmed
dynamic analog speech synthesizer. Control by motor commands
extracted from the electromyogram by means of a pulse counting
method. Original tape was demonstrated at the ASA-ASJ Joint
Meeting held on Nov 27-Dec 1, 1978, in Honolulu, Hawaii. Words
mean: a house, strenuously, a fish, nephew, cover, many,
Japanese-English, predominancy, hero, and upward. "*:" means down
skip of the word accentuation. ":" is the lengthened vowel.
(syn, x2: "Ie. Eiei. Uo. Oi. O:i. O*:i. Waei. Ju:i. Eiju:.
Ue.")
END: **********
SSSHP 87: "TUc - SYNTHESIS OF CHINESE TONE, Research Institute
of Electrical Communication, Tohoku Univ."
SOURCE: Contributed by Dr. Shizuo Hiki, School of Human Sciences,
Waseda Univ., Tokorozawa, Japan, September 23, 1990. See SSSHP
JAP Tohoku Univ. file. 7" reel, +5 vu, good quality, left
channel.
CONTENTS: Narration by S. Hiki.
Copy of tape TUc 1972.3.31. Recording examples of monosyllables,
bisyllables and a sentence consisting of vowels and semi-vowels
with the tones of standard colloquial Chinese. Synthesizing with
a computer-programmed terminal analog speech synthesizer. Control
parameters derived from the data of acoustical analysis and
perceptual test on the Chinese tone as well as physiological
observation of its laryngeal control. Original tape demonstrated
at the 83rd meeting of the ASA held on April 18-21, 1972, in
Buffalo, New York.
1. Monosyllables "ji, a, o", each with the four tones, which are
tone-1, tone-2, tone-3, and tone-4.
(syn, 4 tones each: "Ji. A. O.")
2. Bisyllable, "wuja", with all possible combinations of the four
tones, 16 bisyllables in all. Each of the first syllable with
tone-1, tone-2, half tone-3 tone-4 is followed by the second
syllable with the four tones: tone-1, tone-2, tone-3 and tone-4.
(syn, 16 variations in tone: "wuja")
3. Sentence, "My aunt loves me, and I also love her.", in Chinese.
(syn, x2: "A ji ai wo wo je ai a ji.")
tone: 1 2 4 3 3 3 4 1 2
END: **********
SSSHP 88: "TUp - SYNTHESIS OF PATHOLOGICAL VOICES, Research
Institute of Electrical Communication, Tohoku Univ."
SOURCE: Contributed by Dr. Shizuo Hiki, School of Human Sciences,
Waseda Univ., Tokorozawa, Japan, September 23, 1990. See SSSHP
JAP Tohoku Univ. file. 7" reel, +3 vu, good quality, left
channel, thump between samples toward end of tape.
CONTENTS: Narration by S. Hiki.
Copy of tape TUp 1980.3.16. Original tape demonstrated at the
Conference on Vocal Fold Physiology held in Madison, Wisconsin,
June 1-4, 1981.
1. Recorded samples of synthetic vowel which indicate the most
extreme case of the four scales of pathological voice quality,
namely, rough, breathy, strained and atheniand.
(syn, x2, 4 voices: "eh")
2. Boundaries between four-level rating on the four scales for
pathological voice quality. The four scales are currently used
for subjective judgement by the therapists in voice clinics in
Japan.
(syn, 12 samples: "eh")
END: **********
SSSHP 89: "K.N. STEVENS' REVIEW OF HISTORY OUTLINES, October 1990"
SOURCE: Recorded by H.D. Maxey at MIT, October 23, 1990. Review
of the SSSHP outlines by Professor K.N. Stevens. Two cassettes,
copied from the original microcassettes, good quality.
CONTENTS:
One and one-half hour discussion between H.D. Maxey and Prof.
K.N. Stevens on the draft of the SSSHP outlines. Prof. Stevens
was asked to review each outline and comment on any errors or
missing material. He suggested contact people for France and the
USSR.
END: **********
SSSHP 90: "MIT - MACHINES THAT TALK, Spring 1960"
SOURCE: Prof. K.N. Stevens, Research Laboratory of Electronics,
Massachusetts Institute of Technology, Cambridge, MA. Copy of RLE
archive tape 7-36, itself a copy of RLE spliced master tape 7-35.
Copied by H.D. Maxey, October, 1990, at MIT. Presto Recorder
(full track) to both channels of ReVox A77 recorder (1/2 track
stereo recorder), 7 1/2 ips. See SSSHP USA MIT file for copy of
card catalog of RLE synthetic speech tape file.
Another copy, made in 1962, is SSSHP 127 which has some additional
material by the synthesizer DAVO.
CONTENTS:
Introduction and descriptions by M. L. Hecker.
1. CHANNEL VOCODER, BTL. Male voice over 16-channel vocoder. Bell
Telephone Laboratories. Not used in SSSHP.
(human, 3 sen: "Effective scientific research calls for
creative ability ... in practically all." (may be Frank
Cooper's voice)
(human, 2 sen: "Much of the progress in science is achieved
... plans with each other.")
2. CHANNEL VOCODER, BTL. Female voice over 16-channel vocoder.
Bell Telephone Laboratories. Not used in SSSHP.
(human, 3 sen: "In the spring, a woman's fancy goes plain
haywire ... on with the new.")
3. CHANNEL VOCODER, Haskins Laboratories. Male voice (Frank
Cooper) over channel vocoder. Not used in SSSHP.
(human: "Will you please bring Sam's own mower..." (4
vowels))
4. FORMANT-TRACKING VOCODER, SRDE. Automatic formant analysis and
resynthesis on a parallel formant synthesizer. Walter Lawrence of
Signals Research and Development Establishment, Christchurch,
England. First two formants controlled; third fixed at 3 kHz.
Seven parameters derived from speech. Not used in SSSHP.
(human, 3 sen: "Prior to the United States entry into the war,
... America had declared war.")
5a. PAT, LAWRENCE, SRDE. Parametric Artificial Talking
device at SRDE. Parallel formant synthesizer operated from six
control voltages derived by photo-electronically scanning a
painted glass slide. A modified version of the synthesizer was
used in No. 4, above. See SSSHP UK SRDE file.
(syn, x3: "What did you day before that? Tea or coffee? Half
an hour. Shut the door. What have you done with it?")
(syn, var syn settings: "What have you done with it?")
5b. PAT, controlled with new conductive ink/resistive roller
function generator. Two of the first utterances with this more
elaborate control. See SSSHP UK SRDE file.
(syn, x3: "Do you understand what I say to you? I am
Research.")
6. POVO, MIT. Pole Voice Analog synthesizer at MIT. Cascade
formant synthesizer controlled from 7-hole teletype tape. Short
segments, only. Phrases by tape splicing of segments. See SSSHP
USA MIT file.
(syn: "Where are you?; various synthesizer settings: "Far, far
away.")
7. OVE-I, RIT. Manually controlled cascade formant synthesizer at
Royal Institute of Technology in Stockholm. See SSSHP Sweden RIT
file.
(syn, Swedish vowels; 9 phrases: "How are you? Hello, I am
well... I love you...Yeah, I love you, OVE.; var syn settings:
"Aju")
8. OVE-II, RIT. Second generation cascade formant synthesizer
controlled from a conductive ink/resistive roller function
generator and hand-prepared patterns copying human speech. Human
speech sample followed by synthetic version. See SSSHP Sweden RIT
file.
(G. Fant and syn: "Hello, how are you? What shall I say?")
(H.M. Truby and syn: "Hello, OVE, how are you? What shall I
say to that?
(syn, Swedish phrase, male and female)
(syn, English, Swedish, German, Russian, French: "Yes, no.
Goodbye.")
9. EVT, MIT. Electrical Vocal Tract articulatory synthesizer at
MIT. Thirty-five section transmission model of vocal tract. Manual
switching of inductance and capacitance. Vowel sounds only. See
SSSHP USA MIT file.
(syn, vowels)
10. DAVO, MIT. Dynamic Analog of the Vocal Tract at MIT. Control
voltages from six preset vocal tract configurations. See SSSHP USA
MIT file.
(syn: "a", "sh", "sha", CV's)
(syn, segments and assembled phrase by tape splicing: "The
voice of DAVO.")
END: **********
SSSHP 91: "MIT - DEMO TAPE 1, 10/90"
SOURCE: Prof. K.N. Stevens, Research Laboratory of Electronics,
Massachusetts Institute of Technology, Cambridge, MA. Copies of
RLE archive tapes and tapes in collection of Prof. D.H. Klatt.
Copied by H.D. Maxey, October, 1990, at MIT. Presto Recorder
(full track) to both channels of ReVox A77 recorder (1/2 track
stereo recorder), 7" reel, 7 1/2 ips. See SSSHP USA MIT file for
copy of card catalog of RLE synthetic speech tape file.
CONTENTS:
1. DAVO, MIT. Dynamic Analog of the Vocal tract at MIT with later
addition of nasal tract (DANA, Dynamic Analog Nasal tract). Copy
of spliced addition to RLE tape 7-36 (dated Spring 1960),
believed to be a copy of work the following year (perhaps missing
tape 7-44, Spring 1961). See SSSHP USA MIT file.
(syn, 51 sec: "This is the voice of DAVO at MIT"; Alphabet
Song: "A,B,C,D, ...")
2. SPASS, MIT. Solid state terminal-analog speech synthesizer at
MIT. Copy of RLE tape 7-62, dated Apr 22, 1965. See SSSHP USA MIT
file.
(syn, x2: "buh, duh, guh"; syn: "Where are you?")
3. EDINBURGH UNIV., LADEFOGED AND BROADBENT, 1957. Explanation of
perception experiment by P. Ladefoged at Edinburgh University.
Copy of tape in Klatt collection. See SSSHP UK Edinburgh file.
(syn, 4:38 min, 6 var of: "Please say what this word is:" with
"bit, bet, bat, and but")
4. OVE-I, RIT. Gunnar Fant narration and demonstration of a
"POVO" (Pole Voice Operator) synthesizer, a 4-pole early version
of OVE-I. Copy of RLE tape 5-6 (undated, but believed to be
1952-1954). Abrupt onset and decay at this stage of development.
See SSSHP Sweden RIT file.
(8 min: Narration; syn, English vowels; syn, Swedish vowels and
diphthongs; syn: "How are you? I am well, How are you? How ...,
yeah.; interview in Swedish between human and machine as
famous skiing star: " ... aju."; syn, Swedish words with
intonation variation; ... Goodbye, everybody.")
5. KLATT, MIT, 1971. Demo for "A theory of segmental duration in
English", Acoust. Soc. America, Oct, 1971. Copy of tape in Klatt
collection. For phonetic input see SSSHP 149 IBM - Maxey Tape
Collection Card Index and Tape Transcription Notes for tape SSSHP
147 (Maxey Tape T80.3), the same recording.
(syn, 13 sen, 1:52 min: "You are listening to a demonstration
... acoustic cues.")
The following samples, No. 6 to No. 9, were copied from "History
of Speech Synthesis, D.H. Klatt, MIT, 9/1/71, 765 ft, 18 min".
Explanations are by D.H. Klatt.
6. HOLMES AND ANTHONY, JSRU, circa 1961. Manual specification of
formant synthesizer control parameters to match voice of John
Holmes. Joint Speech Research Unit, Eastcote, England. See SSSHP
UK JSRU file.
(syn: "A bird in the hand is worth two in the bush.")
7. LEE AND ALLEN, MIT, 1968. Speech synthesis by rule at MIT.
See SSSHP USA MIT file.
(syn, 47 sec: "Hello ladies and gentlemen, how are you? I
will now count from one to ten. One, two, ... ten. I will
now say 'Humpty Dumpty'. Humpty Dumpty sat ... together
again.")
8. COKER AND UMEDA, BTL, 1970. Articulatory synthesis by rule at
Bell Telephone Laboratories. See SSSHP USA BTL file.
(syn, 34 sec, Christmas song: "Christmas, Christmas, toys
and noise, that's Christmas. Tiny Tim's and ... that's
Christmas.")
9. COKER AND UMEDA, BTL, 1971. Articulatory synthesis by rule at
Bell Telephone Laboratories. See SSSHP USA BTL file.
(syn: "... am a computer. I can read stories and speak them
aloud. I do not understand what the words mean when I read
them, but I can guess which words are important and which words
are not, by rules I have been given. Someday, I may be able to
provide many kinds of information by telephone.")
10. ECHO. Low cost diphone concatenation system, about 1982.
Four sentences from the Harvard list. Copied from tape in Klatt
collection (source tape for SSSHP 32.29). See SSSHP USA Texas
Instruments file.
(syn, 4 sen: "The birch canoe slid on the smooth planks. Glue
the sheet to the dark blue background. ... These days a
chicken leg is a rare dish.")
11. VOTRAX TYPE-N-TALK SYSTEM, 1978. Copy of tape in Klatt
collection. Source tape for SSSHP 32.28. See SSSHP USA Votrax
file.
(syn, 4 sen: "Rice is often served in round bowls. The juice
of lemons makes fine punch. A box was thrown beside the parked
truck. The hogs were fed chopped corn and garbage.")
12. PROSE 2000 V3.0, TSI. Text-to-speech product by Telesensory
Systems, Inc. (Speech Plus, Inc., after 1982). Copy of tape in
Klatt collection. Source tape for SSSHP 32.32. See SSSHP USA
Telesensory file.
(syn, 4 sen: "Four hours of steady work faced us ... A rod is
used to catch pink salmon.")
13. KLATT SYNTHESIS-BY-RULE PROGRAM, MIT, 6/79. Phonemic input
with stress markings. Copy of tape in Klatt collection. See SSSHP
USA MIT file.
(syn, 1:29 min: "This recording is a demonstration of speech
synthesis by rule and automatic text to speech conversion ...
developed at MIT.")
14. MITALK-79, MIT. Text-to-speech system at MIT. Copied from
Klatt tape, "Klatt Synthesis-by-Rule Program, 6/79".
(syn, 2:32 min: "The remaining speech that you will hear was
produced from ordinary English text using the Allen/Klatt
text-to-speech system called MITalk-79. The transformation ...
such as MITalk-79.")
15. ARTICULATORY MODEL, BTL, COKER. Text-to-speech with an
articulatory synthesizer at Bell Telephone Laboratories. No date.
Copy of tape in Klatt collection. See SSSHP USA BTL file.
(syn, 50 sec: "Hello, this is a demonstration of speech
synthesis from an articulatory model. To generate synthetic
speech, I start with ordinary spelling of English words. I
transform spelling to sound with a dictionary, together with
rules for compound words and English prefixes and suffixes. I
create the sounds of speech ... Thank you.")
16. DIPHONE SYNTHESIS, BTL, OLIVE. Synthesis from concatinated LPC
dyads (diphones), at Bell Telephone Laboratories. Copy of tape in
Klatt collection. No date on tape. Klatt history paper states the
recording is from about 1980. Source tape for SSSHP 32.22. See
SSSHP USA BTL file.
(syn, 1:08 min: "This paper describes a small real time speech
synthesizer. The synthesizer requires as its input, a string
of phonemes and the associated duration, pitch and amplitude
parameters. The synthesis scheme uses a large table of stored
transitions, dyads, between phonemes. These transitions are
stored ... LPC derived area parameters ... LSI 1123
microcomputer ... digital signal processor.")
17. FLANAGAN AND ISHIZAKA, BTL, 1976. Model of the vocal cords for
an articulatory synthesizer at Bell Telephone Laboratories. Copy
of tape in Klatt collection. Source tape for SSSHP 32.12. See
SSSHP USA BTL file.
(syn, x3: VCV's; syn, x2: "She saw the house. This is a test.
We were away a year ago. I am a computer.")
END: **********
SSSHP 92: "MIT - DEMO TAPE 2, 10/90"
SOURCE: Prof. K.N. Stevens, Research Laboratory of Electronics,
Massachusetts Institute of Technology, Cambridge, MA. Copies of
RLE archive tapes and tapes in collection of Prof. D.H. Klatt.
Copied by H.D. Maxey, October, 1990, at MIT. Presto Recorder
(full track) to both channels of ReVox A77 recorder (1/2 track
stereo recorder), 7" reel, 7 1/2 ips. See SSSHP USA MIT file for
copy of card catalog of RLE synthetic speech tape file.
CONTENTS:
1. MEET THE EXPERT, RIT. Copy of damaged RLE tape 7-40. Undated,
but SSSHP 106 copy is dated 30 Jan 58. BBC broadcast of interview
at Stockholm, Sweden, with H.M. Truby, speech phonetician, about
the speech phonetics and machine language translation at the
Royal Institute of Technology. See SSSHP Sweden RIT file.
1a.(human, 10:22 min: Introduction, ... use any profitable and
available means..".; gap and noise at 3:40 min; "... these are
the vowel sounds"; OVE-I female English vowels and diphthongs,
"We, you, are.")
Addition of fricative sound, "s", to OVE-I synthesizer.
(OVE-I syn, male: "Hello. Hello, how are you? O.K. Are you
well? Yes. Say, I love you. Oh? Yes, I love you. I love you.
Oh yeah? Yeah. Ooooo. I love you!"; syn, song, I Love You
Truly: "... truly, dear.")
OVE-II simulation of Truby's voice. (second damaged section)
1b.(human and syn: "Hello, OVE, how are you? What shall I say to
that?")
2. KLATTALK KT-1 DEMO, MAY 1, 1980. Copy of tape in Klatt
collection. See SSSHP USA MIT file.
(male syn, 1:18 min: "The lens buyer must approach ... faster
than f4.5 is really needed.")
(male syn, 1:01 min: "Yesterday morning, at 11:00, I bought an
American flag .. dead of winter.")
3. KLATTALK KT-1, NOV 14, 1980. Copy of tape in Klatt collection.
Male and first female voice. See SSSHP USA MIT file.
(syn, 35 sec: "In many applications, it is advantageous to be
able to synthesize two distinct voices. For example, ... for
women.")
4. POEM, I. MATTINGLY, 1985. Copy of tape in Klatt collection.
Poem by I. Mattingly on a DECTalk 3.0 at the 50th Anniversary
Dinner of Haskins Laboratories, 1985. See SSSHP USA Haskins file.
(syn, 3:45 min: "Poem for the 50th anniversary of the ... here
at Haskins.")
5. KLATT: ANALYSIS OF A FEMALE VOICE, ASA, FALL 1986. Copy of
Klatt demostration tape to accompany the paper, "Analysis of a
female voice," Acoust. Soc. America, Fall 1986. Synthetic male
narration, Diane B.'s ("DB") natural voice, and synthetic female
copy of Diane B.'s voice. "fo" is fundamental frequency. See SSSHP
USA MIT file.
(syn narration: "Analysis and synthesis of a female voice,
Demo 1. Diane B. sustaining a vowel at nominally ...")
(x2: DB at const. fo, syn at const fo, syn with "wow"
syn of fo fall, DB fo fall, /a/
syn of fo rise, DB fo rise, /a,i/
syn of [?a-?a-?a-?a-?a], DB natural
" " " (no noise), DB natural
" " " (no glottal attack), syn with attack)
6. KLATTALK FULL TEXT-TO-SPEECH IN C, MIT, 1982. Demonstration
tape from Klatt collection. See SSSHP USA MIT file.
(1:01 min, male syn: "Thank you for inviting me to the Second
International Conference on Non-Speech Communication that is
being held at the Imperial Institute for Studies in Education,
Toronto, Canada, November 15, 1982. ... Digital Equipment
Corp. is currently evaluating ..."; female syn: "Hello.";
child syn: "Hi.")
7. KLATT, DECTALK 1.8 DEMONSTRATION, 1984. Demonstration tape from
Klatt collection. See SSSHP USA MIT and SSSHP USA DEC files.
(male and female syn, 41 sec: "I am the DECTalk
text-to-speech system. The date is the 23rd of September,
1984. Thank you for inviting me to speak at the Workshop on
Digital Signal Processing and Its Applications. It is a
pleasure to be here in Marrakech, Morocco. ...")
(male Japanese syn: " ... arimasen.")
8. KLATT-DECTALK 3.0, 1986. Copy of tape in Klatt collection.
Source tape for SSSHP 32 samples. See SSSHP USA DEC file.
8a.(male syn: "In May, 1607, three ships sailed up the James
River from ... between them.")
8b.(male syn, 300 wpm: "The following is a list of topics in
today's news, In the sports world, ... in Cambridge.")
8c.Several DECTalk voices.
(syn: "I am Beautiful Betty, the standard female voice. Some
people think I sound a bit like a man.")
(syn: "I am Huge Harry, a very large person with a deep voice.
I can serve as an authority figure.")
(syn: "My name is Kit the Kid, and I am about ten years old,
and I sound like a boy or a girl.")
(syn: "I am Whispering Wendy and I have a very breathy voice
quality. Can you understand me even though I am whispering?")
8d.(child syn, singing, 5 sen: "We love you, Dennis. Oh, yes we
do. We love you, Dennis, and it is true. ... love you.")
9. KLATT JAPANESE DECTALK, 11/5/84. Phonemic input to DECTalk
system. See SSSHP USA MIT file. Copy of tape in Klatt collection.
Full transcription in SSSHP96 Reprints.
(male syn narration, 47 sec: "You are about to hear a
demonstration of Japanese DECTalk. As you know, the standard
DECTalk is a system ... phonemic representation ... will begin
in five seconds."; Japanese male syn, 5 sen, 43 sec: "Kore wa
syo ... sare te iru.")
END: **********
SSSHP 93: "MIT - DEMO TAPE 3, 10/90"
SOURCE: Prof. K.N. Stevens, Research Laboratory of Electronics,
Massachusetts Institute of Technology, Cambridge, MA. Copies of
RLE archive tapes and tapes in collection of Prof. D.H. Klatt.
Copied by H.D. Maxey, October, 1990, at MIT. Presto Recorder
(full track) to both channels of ReVox A77 recorder (1/2 track
stereo recorder), 7" reel, 7 1/2 ips. See SSSHP USA MIT file for
copy of card catalog of RLE synthetic speech tape file.
CONTENTS:
1. MYNAH BIRD, KLATT, MIT. Copy of Klatt demonstration tape for
paper, "How does a mynah bird imitate human speech?", D.H. Klatt
and R.A. Stefanski, JASA, 55.4, Apr 1974, 822-832. Recording of
Indian Hill mynah bird, named Ig-wog, and its tutor. Explanation
by D.H. Klatt. The original recordings have been reordered so
that, for each phrase, one hears first the tutor's voice at a low
level to trigger the bird's response, then the bird's response,
then the tutor's voice at the same level, and finally the bird's
response again. See SSSHP USA MIT file.
(tutor at low level, bird, tutor, bird, 1:48 min: "I'd like a
grape. Learning what to say. You're looking well. Hello. Who
are you? Ig-ig-wog. I just saw a zebra.")
2. KLATT SYNTHESIS BY RULE, 1976, MIT. Synthesis of The Night
Before Christmas. Copy of tape from Klatt collection. See SSSHP
USA MIT file.
(syn: 26 sec: "T'was the night before Christmas ... long
winter's nap.")
3. POVO, MIT, 1953. Selection from RLE tape 7-2, "POVO-2 copy
of 7-1, dated 1/27/53. Synthetic vowels, trying many possible
combinations of formant frequencies. See SSSHP USA MIT file.
(syn: 12 vowels from Set 15F, 10 vowels from Set 16A)
4. EVT NASALIZATION OF VOWELS, JUNE 1955, MIT. Selection from RLE
tape 7-6. Nasalized vowels by Electrical Vocal Tract at MIT. See
SSSHP USA MIT file.
(syn: 14 vowels)
5. EVT NASAL CONSONANTS, MIT, 1956. First 14 samples from RLE tape
7-11, "Nose 3X Trials", dated 8/23/56. See SSSHP USA MIT file.
(syn: 14 vowels and nasal consonants)
6. MITALK 79, MIT, ALLEN, 1979. Copy of tape entitled,
"Allen-MITalk 79, Oct 4, 1979", obtained from Jon Allen at MIT.
First four synthetic sentences are source for SSSHP 32.30. See
SSSHP USA MIT file.
(Jon Allen, 8:20 min: explanation of system)
(syn, 19 long sen, 2:54 min: "Speech is so familiar a feature
of daily life ... cultural function.")
(syn, 9 long sen, 1:35 min: "We usually take for granted our
ability to produce and understand speech ... than any other.")
7. OLIVE AND LIBERMAN, BTL. Synthesis by LPC dyad concatenation
at Bell Telephone Laboratories. Copy of tape from Klatt
collection. No date. See SSSHP USA BTL file.
(syn, 5 sen: "I am a talking computer at Bell Labs. Many
computers can talk. However, most of them can only say words
stored in their memory. I have an unlimited vocabulary because
I can read text, translate it into sound rules, and connect
these sounds into sentences. Researchers are improving my
speech by correcting mistakes that I still make and trying to
make me sound better.")
8. BROWMAN, BTL, 1980. Rules for demisyllable synthesis using
Lingua, a language interpreter, at Bell Telephone Laboratories.
Copy of tape from Klatt collection. See SSSHP USA BTL file.
(syn, 10 sen, 1:57 min: "Hello, I am a language interpreter
named Lingua. I was created at Bell Laboratories by Cathy
Browman, with a lot of help from a number of people... type
the sentences you want me to say into the computer, indicating
how you want me to pronounce the sentences, and which words you
want to emphasize... linguistic research can be useful for all
of us.")
9. OLIVE, BTL, 1985. Synthesis by LPC dyad concatination at Bell
Telephone Laboratories. Required input specification of
phoneme string, duration, pitch, and amplitude. Copy of tape in
Klatt collection. See SSSHP USA BTL file. Seems like longer
version of SSSHP 91.16.
(syn, 1:57 min: "This paper describes a small real-time speech
synthesizer. The synthesizer requires as its inputs a string
of phonemes and the associated duration, pitch, and amplitude
parameters. The synthesis scheme uses a large table of stored
transitions, dyads, ... LSI 1123 microprocessor ... Our
original programs were written in C, and ran under the UNIX
time-sharing system ... LPC parameters.")
10. KLATT, MIT. Copy of tape in Klatt collection. No date. See
SSSHP USA MIT file.
(syn: "... Wind and the Sun. The North Wind and the Sun were
arguing one day ... overcoat ... stronger than he was.")
11. INFOVOX SA101, RIT. Copy of tape in Klatt collection. No date.
Source tape for SSSHP 32.31. See SSSHP Sweden RIT file.
(2:53 min: "This is a demonstration of speech synthesis
presented with the aid of synthesis device, the SA101. It is
produced by Infovox, a company that ... set with a command.")
END: **********
SSSHP 94: "PAT - SYNTHESIZED INTONATIONS CIRCA 1962"
SOURCE: Copy of SSSHP 84, fourteen tape loops of the PAT
synthesizer at University of Edinburgh. Copied by H.D. Maxey,
1/91. Cassette.
CONTENTS:
A single utterance repeated with different intonations. The tape
loops were copied in order, B1 to B14. See SSSHP UK Edinburgh
file.
(syn, 14 sen: "He'll be here on Friday.")
END: **********
SSSHP 95: "VIDEOTAPE: STEPHEN HAWKING: THE UNIVERSE WITHIN"
SOURCE: Copy of tape donated by Ms. Teri H. Flinchum, Cabisco
Teleproductions, Carolina Biological Supply Company, Burlington
NC, April 10, 1991. This tape is proprietary to Carolina
Biological Supply Co and copies must be obtained from them (see
SSSHP USA Telesensory Systems, Inc. file for prices). Videotape,
Order No. 53-8420-M, 3/4-inch.
CONTENTS: Tape of physicist Stephen Hawking using a speech
synthesizer to recount his career. MIT Professor Dennis Klatt
identified the synthesizer as a PROSE 2000 from Telesensory
Systems, Inc. Videotape was broadcast on the Public Broadcasting
System in 1990. Tape illustrates use of speech synthesis as an
aid for the handicapped. Time 28:45 min.
END: **********
SSSHP 99: "SYNTHETIC VOICES FOR COMPUTERS, BTL, 1970"
SOURCE: Copy of disk recording SSSHP 59. Donated by H.D. Maxey,
June 21, 1991. Cassette, fair quality, stylus noise.
CONTENTS: Demonstration to accompany paper by Flanagan, J.L.,
C.H. Coker, L.R. Rabiner, R.W. Schafer, and N. Umeda, "Synthetic
voices for computers", IEEE Spectrum, 7, 22-45 (1970). Copy in
SSSHP 98 Reprints. The following is an elaboration of Appendix A
of the paper.
SIDE 1 - FORMANT ANALYSIS AND SYNTHESIS
1. AUTOMATIC ANALYSIS AND SYNTHESIS. Original human speech and
synthesis comparison. Order is: synthesis, original human speech,
synthesis.
(syn, human, syn: "We were away a year ago. The number is.
826-514-7903.")
2. SMOOTHING OF PITCH AND FORMANT CONTROLS. One utterance with 4
different smoothing filters of 16, 12, 8, and 4 Hz low-pass
cutoff.
(syn, 4 var: "We were away a year ago.")
3. QUANTIZING AND SMOOTHING OF PITCH AND FORMANT CONTROLS. One
utterance with 16-Hz cutoff low-pass filter and the following
quantizing variations.
A. Pitch quantizing of 7, 4, 3, 2, 1 bits.
(syn, 5 var: "We were away a year ago.")
B. Formant 1 quantizing of 4, 3, 2, 1 bits.
(syn, 4 var: "We were away a year ago.")
C. Formant 2 quantizing of 4, 3, 2, 1 bits.
(syn, 4 var: "We were away a year ago.")
D. Formant 3 quantizing of 3, 2, 1 bits.
(syn, 3 var: "We were away a year ago.")
4. BIT RATE COMPARISON. One utterance at two quantizations (two
bit rates.)
A. At 4600 bits/sec. Number of bits: Pitch = 7, F1 = 10, F2 =
11, F3 = 11, Av = 7. Sampling rate = 100/sec.
B. At 600 bits/sec. Number of bits: Pitch = 6, F1 = 3, F2 = 4,
F3 = 3, Av = 2. Sampling rate = 32/sec.
Order of presentation is 4600 b/s, original, 4600 b/s, 600 b/s.
(syn, 4 var: "We were away a year ago.")
5. MANIPULATION OF PITCH, TIMING, AND FORMANT CONTROLS. One
utterance with the following variations.
A. Pitch manipulations
(1) Synthetic unmanipulated
(2) Monotone pitch - 100 Hz
(3) Pitch doubled
(4) Pitch squared
(syn, 4 var: "We were away a year ago.")
B. Timing manipulations
(1) Synthetic unmanipulated
(2) Vowels in "we" and "year" lengthened by 100 ms.
(3) Synthetic unmanipulated
(4) Total duration = 75% natural duration
(5) Total duration = 50% natural duration
(6) Synthetic unmanipulated
(7) Total duration = 150% natural duration
(8) Total duration = 200% natural duration
(syn, 8 variations: "We were away a year ago.")
C. Formant manipulations
(1) Synthetic unmanipulated
(2) Formants raised by 10%, pitch raised
(3) Formants raised by 20%, pitch raised
(4) Formants raised by 30%, pitch raised by 50%
(5) Synthetic unmanipulated
(6) Formants lowered by 10%
(7) Formants lowered by 20%
(syn, 7 var: "We were away a year ago.")
6. CONCATENATION OF WORDS. One utterance with the following
variations.
A. Isolated words in sequence.
B. Words concatenated by rule; natural pitch and timing from
speaker 1.
C. Words concatenated by rule; natural pitch and timing from
speaker 2.
D. Words concatenated by rule; natural pitch and timing from
speaker 3.
(syn, 4 var: "We were away a year ago.")
7. CONCATENATED DIGIT STRINGS. Four comparisons of strings of
isolated digits followed by concatenated digits.
(syn, 2 var: "The number is 091-1909.")
(syn, 2 var: "The number is 135-3201.")
(syn, 2 var: "The number is 377-8348.")
(syn, 2 var: "The number is 836-1246.")
SIDE 2 - SYNTHESIS FROM PRINTED TEXT
8. ARTICULATORY SYNTHESIS FROM MANUAL PHONETIC INPUT.
(syn: "Good morning. I am a computer. I can read stories and
speak them aloud. I do not understand what the words mean when
I read them, but I can guess which words are important and
which words are not, by rules I have been given. Some day I
may be able to provide many kinds of information by
telephone.")
9. AUTOMATIC SYNTHESIS FROM PRINTED TEXT. "Parable of the North
Wind and the Sun."
(syn: "The North Wind and the Sun were arguing one day when a
traveler came along wrapped in a warm coat. They agreed that
the one who could make the traveler take his coat off would be
considered stronger than the other one. Then the North Wind
blew as hard as he could but the harder he blew, the tighter
the traveler wrapped his coat around him. At last, the North
Wind gave up trying. Then the Sun began to shine hotter and the
traveler took off his coat immediately. And so, the North Wind
was obliged to confess that the Sun was the stronger of the
two.")
10. SYNTHESIS FROM PRINTED TEXT.
(syn: "This is the computer talking. It's been a pleasure
communicating with you tonight, both through visual displays,
and now with computer synthesized speech. Good night, Folks.")
END: **********
SSSHP 106: "MEET THE EXPERT, 30 Jan 58, Dr. H.M. Truby"
SOURCE: Recording of BBC broadcast interview of Dr. Truby on
subject of speech phonetics and synthetic speech. (see SSSHP
SWEDEN RIT file.) Donated by H.M. Truby, 4/22/92. 5" reel, 1-7/8
ips. See Tape 92.1 for transcription. A temporary 3-3/4 ips copy
is SSSHP 165.
Tape has water damage and physical distortion. Needs professional
restoration. May be a better copy than damaged SSSHP 92.1.
END: **********
SSSHP 114: "MITalk 79 speech from development, 1978-1979."
SOURCE: Analog copy of Digital Audio Tape of synthetic speech
samples compiled Apr. 2001 at Royal Institute of Technology from
Sheri Hunnicutt's collection. Cassette. (See SSSHP USA MIT)
Donated by Sheri Hunnicutt, June 29, 2001. Fair quality (some
sudden level changes & overload; occasional mixing of other
recordings.)
CONTENTS:
Side 1
The first 20 minutes:
1) a text: MAN and MACHINES synthesized first with old rules,
then with 2 changes in FO (fundamental frequency) rules.
(5.5 minutes)
2) 2 sentences followed by a version with just /mmmm/ instead
of phonemes
3) A selection (8 minutes) from Woody Allen: Without Feathers
4) A selection from Virginia Woolf (The man who loved his kind)
5) A short selection about the galaxy
6) A short selection about fire
Side 2
The next few minutes:
1) A selection from book LANGUAGE, by Sapir
The next 4.5 minutes:
1) Test materials from the Harvard Psychoacoustic Sentences used
in the evaluation of MITalk.
2) Test materials from the Haskins Anomalous Sentences. There is
more about these sentences and the evaluation in the section
written by David Pisoni of the book FROM TEXT TO SPEECH: The
MITalk system (Allen, Hunnicutt, Klatt with Robert C. Armstrong
and David B. Pisoni.
The remainder (about 12.5 minutes):
The story "The North Wind and the Sun," first direct from the
synthesized speech, twice read by Dennis Klatt, and then a number
of times with different prosodic rules during the development
process. Includes modified duration and FO rules.
(Notes by Sheri Hunnicutt, April, 2001)
END: **********
SSSHP 115: Videotape: "MITalk '79"
SOURCE: Part of Senior thesis work in Prof. Jonathan Allen's
group at MIT by Alex Waibel. No titles, backup copy. VHS
videotape, 1/2" tape, 7" reel. (See SSSHP USA MIT) Donated by
Sheri Hunnicutt, April 19, 2001.
CONTENTS: Video of demonstration of MITalk '79 speech synthesis
system.
END: **********
|