| Development of a Spoken Language System |
APR 92 |
242 pages |
| Authors:
S. Austin; D. Ayuso; C. Barry; M. Bates; R. Bobrow; BBN SYSTEMS AND TECHNOLOGIES CORP CAMBRIDGE MA
|
 | This report describes the activities performed and the progress made in the development of HARC (Hear And Respond to Continuous speech), BBN's spoken language system, from May 1, 1989 to February 29, 1992. Significant progress has been made both in terms of speed of understanding and accuracy of understanding. New search algorithms BBN developed during this period increased the speed of HARC by more than three orders of magnitude. The ... |
|
| Determination of Optimal Training Methodologies for Discrete/Dependent Speech Recognition (SR) Systems |
MAR 92 |
|
| Authors:
Mark C. Rhoads; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
 | A research experiment was conducted to determine whether various combinations of training methodologies and speaking voices would affect recognition accuracies amongst unique speaker dependent speech recognition (SR) systems. The experiment used a SR system (VOTAN VTR 6050II) which is based on VOTAN (proprietary) technology. Ten subjects trained five different voice patterns each and conducted four natural voice tests to compile statistics about the recognition accuracy for each pattern. Two patterns ... |
|
| Continuous Speech Recognition as an Input Method for Tactical Command Entry in the SH-60B Helicopter |
MAR 92 |
74 pages |
| Authors:
Richard A. Powers; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
| An experiment was conducted to determine whether a continuous speech recognition system would reduce the SH-60B Airborne Tactical Officer's taskload. The experiment made use of a Verbex Series 5000 speech recognizer. Ten subjects entered 45 commands frequently used by the Airborne Tactical Officer via two input methods: continuous voice and keying. The experiment was successful and demonstrated that continuous speech recognition is an effective means of reducing the Airborne Tactical ... |
|
| Evaluation of Lightweight and Low Profile Communications Devices for Respiratory Protective System 21 (RESPO 21) |
FEB 92 |
33 pages |
| Authors:
James E. Dvorsky; G. F. Renner; Kevin M. Taylor; William J. Williams; Kenneth J. Woodruff; BATTELLE COLUMBUS DIV OH
|
| The Chemical Research, Development, and Engineering Center (CRDEC) is entering development of the next generation of respiratory protection (RESPO 21) to replace the current M40 series of protective masks. One of the design goals of this development is to improve both face-to-face and electronic communications while wearing respiratory protection. Commercial voice amplification systems being evaluated for this application are too bulky for many field applications. An evaluation of miniaturized communication ... |
|
| BBN BYBLOS and HARC February 1992 ATIS Benchmark Results |
FEB 1992 |
7 pages |
| Authors:
Francis Kubala; Chris Barry; Madeleine Bates; Robert Bobrow; Pascale Fung; Robert Ingria; John Makhoul; Long Nguyen; Richard Schwartz; David Stallard; BBN SYSTEMS AND TECHNOLOGIES CORP CAMBRIDGE MA
|
| We present results from the February `92 evaluation on the ATIS travel planning domain for HARC, the BBN spoken language system (SLS). In addition, we discuss in detail the individual performance of BYBLOS, the speech recognition (SPREC) component. In the official Scoring, conducted by NIST, BBN's HARC system produced a weighted SLS score of 43.7 on all 687 evaluable utterances in the test set. This was the lowest error achieved ... |
|
| Feasibility Standards for Communicative Skills. Using Existing Instruments to Measure Noncommissioned Officers' Communicative Skills Proficiencies |
27 JAN 92 |
|
| Authors:
Carl R. Dolmetsch; Winn B. McDougal; Deborah N. Vause; MANDEX INC NEWPORT NEWS VA
|
| An investigation was conducted to determine the feasibility of using existing instruments to measure the communicative skills proficiencies (writing, speaking, and listening) of Noncommissioned Officers (NCOs) to the Army's standards. The literature was reviewed with an emphasis on issues associated with testing in general and proceeded to the specific areas of writing, speaking, and listening. Support was found for the interrelatedness of these skills (the 'whole language' approach). However, it ... |
|
| The Development of the Speaker Independent ARM Continuous Speech Recognition System |
JAN 92 |
26 pages |
| Authors:
M. J. Russell; ROYAL SIGNALS AND RADAR ESTABLISHMENT MALVERN (UNITED KINGDOM)
|
| This memorandum describes the development of a speaker independent continuous speech recognition system based on phoneme level hidden Markov models. The system is configured to recognise continuously spoken airborne reconnaissance reports, a task which involves a vocabulary of approximately 500 words. On a test set of speech from 80 male subjects, the final system achieves a word accuracy of 74.1% with no explicit syntactic constraints. |
|
| The Use of Linear Discriminant Analysis in the ARM Continuous Speech Recognition System |
JAN 92 |
|
| Authors:
S. M. Peeling; K. M. Ponting; ROYAL SIGNALS AND RADAR ESTABLISHMENT MALVERN (UNITED KINGDOM)
|
| Linear discriminant analysis is used to generate speech data transformations. This transformed data is then within the ARM continuous speech recognition system. Experiments are described using transformed data in conjunction with variable frame rate analysis and word transition penalties. Speaker independent results are reported which are as good as the best obtained previously using cosine transformations and variable frame rate analysis. The two sets of results are compared and commented ... |
|
| Interpretation as Abduction |
92 |
68 pages |
| Authors:
Jerry R. Hobbs; Mark Stickel; Douglas Appelt; Paul Martin; SRI INTERNATIONAL MENLO PARK CA ARTIFICIAL INTELLIGENCE CENTER
|
| Abduction is inference to the best explanation. In the TACITUS project at SRI we have developed an approach to abductive inference, called weighted abduction , that has resulted in a significant simplification of how the problem of interpreting texts is conceptualized. The interpretation of a text is the minimal explanation of why the text would be true. More precisely, to interpret a text, one must prove the logical form of ... |
|
| Communication and Attitude Revision |
92 |
14 pages |
| Authors:
Douglas E. Appelt; SRI INTERNATIONAL MENLO PARK CA
|
| Much recent research has been directed toward understanding those aspects of language use that fall into that somewhat ill-defined area between semantics and pragmatics. The linguistic phenomena that seem to fall into this area include presupposition, implicature, speech acts (especially performatives) , metonymy, and metaphor. These linguistic phenomena can be characterized by a failure of truth conditional semantics alone to provide a satisfactory account, which is manifested in an obvious ... |
|
| A Unified Abductive Treatment of the Intentional and Informational Aspects of Discourse Interpretation: A Preliminary Report |
92 |
16 pages |
| Authors:
Jerry R. Hobbs; SRI INTERNATIONAL MENLO PARK CA ARTIFICIAL INTELLIGENCE CENTER
|
| The problem of interpreting discourse has been subsumed under the general problem faced by intelligent agents of interpreting the situation they are in by explaining the observable facts. The possibility of interpreting an event as the saying by an intelligent agent of a meaningful stretch of discourse is given by an axiom-axiom (18). The ways in which a stretch of discourse can be analyzed into its parts are given by ... |
|
| Recognition Using Classification and Segmentation Scoring |
1992 |
6 pages |
| Authors:
Owen Kimball; Mari Ostendorf; Robin Rohlicek; BOSTON UNIV MA
|
| Traditional statistical speech recognition systems typically make strong assumptions about the independence of observation frames and generally do not make use of segmental information. In contrast, when the segmentation is known, existing classifiers can readily accommodate segmental information in the decision process. We describe an approach to connected word recognition that allows the use of segmental information through an explicit decomposition of the recognition criterion into classification and segmentation scoring. ... |
|
| Performance of SRI's Decipher Speech Recognition System on DARPA's CSR Task |
1992 |
6 pages |
| Authors:
Hy Murveit; John Butzberger; Mitch Weintraub; SRI INTERNATIONAL MENLO PARK CA
|
| SRI has ported its DECIPHER(trademark) speech recognition system from DARPA's ATIS domain to DARPA's CSR domain (read and spontaneous Wall Street Journal speech). This paper describes what needed to be done to port DECIPHER(trademark), and reports experiments performed with the CSR task. The system was evaluated on the speaker-independent (SI) portion of DARPA's February 1992 "Dry-Run" WSJ0 test and achieved 17.1% word error without verbalized punctuation (NVP) and 16.6% error ... |
|
| Vocabulary and Environment Adaptation in Vocabulary-Independent Speech Recognition |
1992 |
7 pages |
| Authors:
Hsiao-Wuen Hon; Kai-Fu Lee; CARNEGIE-MELLON UNIV PITTSBURGH PA SCHOOL OF COMPUTER SCIENCE
|
| In this paper, we are looking into the adaptation issues of vocabulary-independent (VI) systems. Just as with speaker- adaptation in speaker-independent system, two vocabulary adaptation algorithms [5] are implemented in order to tailor the VI subword models to the target vocabulary. The first algorithm is to generate vocabulary-adapted clustering decision trees by focusing on relevant allophones during tree generation and reduces the VI error rate by 9%. The second algorithm, ... |
|
| Spontaneous Speech Collection for the CSR Corpus |
1992 |
7 pages |
| Authors:
Jared Bernstein; Denise Danielson; SRI INTERNATIONAL MENLO PARK CA
|
| As part of a pilot data collection for DARPA's Continuous Speech Recognition (CSR) speech corpus, SRI International experimented with the collection of spontaneous speech material. The bulk of the CSR pilot data was read versions of news articles from the Wall Street Journal (WSJ), and the spontaneous sentences were to be similar material, but spontaneously dictated. In the first pilot portion of the data collection, twelve subjects including nine journalists ... |
|
| Experimental Results for Baseline Speech Recognition Performance using Input Acquired from a Linear Microphone Array |
1992 |
7 pages |
| Authors:
Harvey F. Silverman; Stuart E. Kirtman; John E. Adock; Paul C. Meuse; BROWN UNIV PROVIDENCE RI DIV OF ENGINEERING
|
| In this paper, baseline speech recognition performance is determined both for a single remote microphone and for a signal derived from a delay-and-sum beamformer using an eight-microphone linear array. An HMM-based, connected-speech, 38-word vocabulary (alphabet, digits, 'space', 'period'), talker-independent speech recognition system is used for testing performance. Normal performance, with no language model, i.e., raw word-level performance, is currently about 81% for a set of talkers not in the training ... |
|
| Minimizing Speaker Variation Effects for Speaker-Independent Speech Recognition |
1992 |
7 pages |
| Authors:
Xuedong Huang; CARNEGIE-MELLON UNIV PITTSBURGH PA SCHOOL OF COMPUTER SCIENCE
|
| ABSTRACT For speaker-independent speech recognition, speaker variation is one of the major error sources. In this paper, a speaker-independentnor- malization network is constructed such that speaker variation effects can be minimized. To achieve this goal, multiple speaker clusters are constructed from the speaker-independent training database. A codeword-dependent neural network is associated with each speaker cluster. The cluster that contains the largest number of speakers is designated as the golden cluster. ... |
|
| Multiple Approaches to Robust Speech Recognition |
1992 |
7 pages |
| Authors:
Richard M. Stern; Fu-Hua Liu; Yoshiaki Ohshima; Thomas M. Sullivan; Alejandro Acero; CARNEGIE-MELLON UNIV PITTSBURGH PA SCHOOL OF COMPUTER SCIENCE
|
| This paper compares several different approaches to robust speech recognition. We review CMU's ongoing research in the use of acoustical pre-processing to achieve robust speech recognition, and we present the results of the first evaluation of pre- processing in the context of the DARPA standard ATIS domain for spoken language systems. We also describe and compare the effectiveness of three complementary methods of signal processing for robust speech recognition: acoustical ... |
|
| Language Understanding Research at Paramax |
1992 |
3 pages |
| Authors:
Deborah A. Dahl; Carl Weir; Suzanne L. Taylor; Lewis M. Norton; Marcia C. Linebarger; Mark Lipshutz; PARAMAX SYSTEMS CORP PAOLI PA
|
| Language understanding work at Paramax focuses on applying general-purpose language understanding technology to spoken language understanding, text understanding, and document processing, integrating language understanding with speech recognition, knowledge-based information retrieval and image understanding. |
|
| Improving State-of-the-Art Continuous Speech Recognition System Using the N-Best Paradigm with Neural Networks |
1992 |
6 pages |
| Authors:
S. Austin; G. Zavaliagkost; J. Makhoul; R. Schwartz; BBN SYSTEMS AND TECHNOLOGIES CORP CAMBRIDGE MA
|
| In an effort to advance the state of the art in continuous speech recognition employing hidden Markov models (HMM), Segmental Neural Nets (SNN) were introduced recently to ameliorate the well- known limitations of HMMs, namely, the conditional-independence limitation and the relative difficulty with which HMMs can handle segmental features. We describe a hybrid SNN/I-IMM system that combines the speed and performance of our HMM system with the segmental modeling capabilities ... |
|
| ANDVT Rate Conversion Algorithm (from 2400 b/s to 1200 b/s) |
27 DEC 91 |
28 pages |
| Authors:
George S. Kang; Larry J. Fransen; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
| The 2400-b/s linear predictive coder (LPC) is currently being widely deployed to support tactical voice communication over narrowband channels. However, lower-data-rate voice encoders are needed for special applications: improved performance in high-bit-error conditions and implementation of integrated voice/data systems. In this report, we generated a rate-conversion algorithm to compress voice data rate from 2400 b/s to 1200 b/s. Rate reduction is effected outside the Advanced Narrowband Digital Voice Terminal (ANDVT), ... |
|
| Preliminary Results on the Use of Linear Discriminant Analysis in the ARM Continuous Speech Recognition System |
16 DEC 91 |
|
| Authors:
S. M. Peeling; K. M. Ponting; ROYAL SIGNALS AND RADAR ESTABLISHMENT MALVERN (UNITED KINGDOM)
|
| Liner discriminant analysis is used to generate speech data transformations. This transformed data is then used within the ARM continuous speech recognition system. Preliminary results are presented from experiments using transformed data alone and also in conjunction with one, or both, of word transition penalties and variable frame rate analysis. Speaker dependent results are reported which are significantly better than the best obtained previously. |
|
| Function Prediction Using Recurrent Neural Networks |
DEC 91 |
111 pages |
| Authors:
Randall L. Lindsey; AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF ENGINEERING
|
| A fully recurrent neural network was applied to the function prediction problem. The real-time recurrent learning (RTRL) algorithm was modified and tested for use as a viable function predictor. The modification gave the algorithm a variable learning rate and a linear/sigmoidal output selection. Verifying the networks ability to temporally learn both the classic exclusive-OR (XOR) problem and the internal state problem, the network was then used to simulate the frequency ... |
|
| Technical Feasibility of Digital Three-Dimensional Cellular Communications for Air Traffic Control Applications. Volume 1. Introduction and Summary |
DEC 91 |
|
| Authors:
B. E. White; M. Leiter; R. I. Millar; J. L. Ramsey; R. D. Sakamoto; MITRE CORP BEDFORD MA
|
| MITRE's Center for Advanced Aviation System Development (CAASD) has proposed a set of concepts for improving VHF communications for air traffic control applications. One idea, called CTAG for Cellular Trunked Air Ground (CTAG) communications is to extend land-mobile cellular-trunked digital communications technology to air-ground communication between pilots and controllers. This study was aimed at addressing the technical feasibility of this approach. Detailed results show that significant benefits can indeed be ... |
|
| Technical Feasibility of Digital Three-Dimensional Cellular Communications for Air Traffic Control Applications. Volume 2. Example System Design Details |
DEC 91 |
|
| Authors:
M. Leiter; R. I. Millar; J. L. Ramsey; B. E. White; W. J. Wilson; MITRE CORP BEDFORD MA
|
| MITRE's Center for Advanced Aviation System Development (CAASD) has proposed a set of concepts for improving VHF communications for air traffic control applications. One idea, called CTAG for Cellular Trunked Air Ground (CTAG) communications is to extend land-mobile cellular-trunked digital communications technology to air-ground communication between pilots and controllers. This study was aimed at addressing the technical feasibility of this approach. Detailed results show that significant benefits can indeed be ... |
|
| Technical Feasibility of Digital Three-Dimensional Cellular Communications for Air Traffic Control Applications. Volume 3. Ground Network Architecture |
DEC 91 |
|
| Authors:
M. Leiter; R. I. Millar; J. L. Ramsey; R. D. Sakamotor; B. E. White; MITRE CORP BEDFORD MA
|
| MITRE's Center for Advanced Aviation System Development (CAASD) has proposed a set of concepts for improving VHF communications for air traffic control applications. One idea, called CTAG for Cellular Trunked Air Ground (CTAG) communications is to extend land-mobile cellular-trunked digital communications technology to air-ground communication between pilots and controllers. This study was aimed at addressing the technical feasibility of this approach. Detailed results show that significant benefits can indeed be ... |
|
| Natural Language Based Inference Procedures Applied to Schubert's Steamroller |
DEC 91 |
14 pages |
| Authors:
Robert Givan; David McAllester; Sameer Shalaby; MASSACHUSETTS INST OF TECH CAMBRIDGE ARTIFICIAL INTELLIGENCE LAB
|
| We have previously argued that the syntactic structure of natural language can be exploited to construct powerful polynomial time inference procedures. This paper supports the earlier arguments by demonstrating that a natural language based polynomial time procedure can solve Schubert's steamroller in a single step.... Natural language, Automated theorem proving, Tractable inference, Schubert's steamroller, Automated inference, Socratic proof systems. |
|
| A Study of Channel-Access Schemes for Integrated Voice/Data Radio Networks |
29 NOV 91 |
45 pages |
| Authors:
Jeffrey E. Wieselthier; Anthony Ephremides; NAVAL RESEARCH LAB WASHINGTON DC
|
| In this report we address the major issues associated with channel access in integrated wireless networks, and we propose and analyze the Pure- Reservation Voice/Data Non-Interleaved-Frame Fixed-Length (PR-VD-NIFFL) protocol. This scheme is well suited to either satellite or to terrestrial networks. A two-dimensional first-order Markov chain model for this scheme is presented, and techniques that exploit the structural properties of this chain to simplify the evaluation of the equilibrium state, ... |
|
| Feasibility of Using Radioactive Tracers for Studies of Permeation of Chemicals through Protective Clothing Materials |
OCT 91 |
72 pages |
| Authors:
Iver Drabaek; Christian Ursin; DANISH ISOTOPE CENTRE BRONDBY
|
| This report describes the examination of the feasibility of using radioactive tracers, i.e., carbon 14, in studies of permeation of chemicals through protective clothing. The primary method in the feasibility study has been microscale autoradiography. The technique, well known from biological science, has been modified for work with highly diffusible chemicals and polymers. As a result, the concentration profile inside the material from a given carbon 14 labeled permeant is ... |
|
| JPRS Report Science & Technology Europe Twenty-Fourth Isata International Symposium on Automotive Technology and Automation |
05 SEP 1991 |
|
| Authors:
JOINT PUBLICATIONS RESEARCH SERVICE ARLINGTON VA
|
| Selections from the Twenty-Fourth International Symposium on Automotive Technology and Automation held in Florence, 20-24 May 1991. |
|
| An Analysis of Aircrew Communication Patterns and Content |
03 SEP 91 |
|
| Authors:
Randall L. Oser; Carolyn Prince; Ben B. Morgan Jr.; Steven S. Simpson; NAVAL TRAINING SYSTEMS CENTER ORLANDO FL
|
| The findings reported here represent a detailed analysis of tactical rotary-wing aircrew communication patterns and content. This research is part of an extensive effort to investigate the nature of tactical aircrew coordination and to develop effective mission-oriented aircrew coordination training. The primary objectives of this research were to answer the following questions: (1) What specific communication patterns and content are demonstrated by different helicopter crewmembers (i.e., Helicopter Aircraft Commander - ... |
|
| The Use of Distinctive Features for Automatic Speech Recognition |
SEP 91 |
93 pages |
| Authors:
Helen M. Meng; MASSACHUSETTS INST OF TECH CAMBRIDGE LAB FOR COMPUTER SCIENCE
|
| One of the most critical and yet unsolved problems in phonetic recognition is the transformation of the continuous speech signal to a discrete representation for accessing words in the lexicon. In order to find an efficient description of speech for recognition tasks, our research investigates the use of distinctive features. Distinctive features are a small set of linguistic units which have the potential advantage of enabling us to describe contextual ... |
|
| Report of Potential Applications of Voice Technology to Armor Training |
SEP 91 |
33 pages |
| Authors:
Albert L. Ingram; SCIENTIFIC SYSTEMS INC CAMBRIDGE MA
|
| This report presents potential applications of current low-risk voice technology to military instruction. Three programs of instruction were analyzed to select specified training tasks that might benefit from the use of this technology. In general, two classes of tasks were identified as potential candidates. The first class consists of those tasks that have a high oral content and are good candidates for computer-based instruction and interactive video instruction. The second ... |
|
| Satellite Data Link Validation Test Plan |
SEP 91 |
|
| Authors:
Hilda M. DiMeo; FEDERAL AVIATION ADMINISTRATION TECH- NICAL CENTER ATLANTIC CITY NJ
|
| This document describes the validation process of a satellite data link which will be conducted by the Federal Aviation Administration (FAA) technical Center. Aircraft equipped with satellite communication avionics will relay progress reports through a satellite to a ground earth station (GES). These will be compared directly to voice reports made using high-frequency (HF) radio. The results of the comparisons between the different links will be used to determine the ... |
|
| The Effects of Speech Intelligibility Level on Concurrent Visual Task Performance |
SEP 91 |
33 pages |
| Authors:
David G. Payne; Leslie J. Peters; Deborah P. Birkmire; Georges R. Garinther; HUMAN ENGINEERING LAB ABERDEEN PROVING GROUND MD
|
| Two experiments were performed to determine if changes in speech intelligibility level can impact performance levels in concurrent visual tasks. The auditory task used in both experiments was the auditory memory search task in which subjects memorized a set of word and then decided whether auditorally presented probe items were members of the memorized set. Experiment (1) used an unstable tracking task as the visual task, and experiment (2) used ... |
|
| Speech Recognition Application in C.I.C |
SEP 91 |
58 pages |
| Authors:
Constantinos P. Leventis; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
| The use of a continuous voice recognition system for data input to tactical tables in the combat information center would improve the man-machine interface and decrease the reaction time of operators who run the tables. The results of this study show that the delay times of trained personnel using manual typing input methods were far greater than when they used continuous speech input to run two tactical tables. Using a ... |
|
| Diagnostic Rhyme Test Statistical Analysis Programs |
AUG 91 |
|
| Authors:
A. Sim; R. Bain; A. J. Belyavin; R. L. Pratt; ROYAL SIGNALS AND RADAR ESTABLISHMENT MALVERN (UNITED KINGDOM)
|
|
| Research Laboratory of Electronics Progress Report Number 133 |
AUG 91 |
|
| Authors:
Jonathan Allen; Daniel Kleppner; MASSACHUSETTS INST OF TECH CAMBRIDGE RESEARCH LAB OF ELECTRONICS
|
| This Progress Report contains both a statement of research objectives and a summary of research efforts for research projects listed. Partial contents include: (1) Submicron structures technology and research; (2) Microstructural evolution in thin films of electronic materials; (3) Focused ion beam fabrication; (4) Chemical reaction dynamics at surfaces; (5) Measurement of electron-phonon interactions through large-amplitude phonon excitation; (6) Chemical beam epitaxy of compound semiconductors; (7) High-frequency INAlAs/ InGaAs metal-insulator-doped ... |
|
| Cochannel Talker Interference Suppression |
26 JUL 91 |
|
| Authors:
M. A. Zissman; MASSACHUSETTS INST OF TECH LEXINGTON LINCOLN LAB
|
| Cochannel talker interference suppression is defined as the processing of a waveform containing two simultaneous speech signals, referred to as the target and the jammer, to produce a signal containing an estimate of the target speech signal alone. The first part of this report describes the evaluation of a simulated suppression system that attenuates the jammer component of a cochannel signal, given the voicing states (voiced, unvoiced, silent) of the ... |
|
| Research and Development for Digital Voice Processing (DVP) |
JUL 91 |
141 pages |
| Authors:
J. D. Tardelli; P. A. LaFollette; C. M. Walter; J. LeBlanc; P. D. Gatewood; ARCON CORP WALTHAM MA
|
| This technical report covers a variety of topics and research areas. An update of the Canonical Coordinate (CC) Transformation process for digital speech compression based on non-Euclidean error minimization criteria is given. A method of transforming empirical filter sets to unitary error metrics is introduced and a sample error metric is developed. Research into frequency shift invarient transformation is presented and its utility for developing CC error metrics is discussed. ... |
|
| Detection of Prosodics by Using a Speech Recognition System |
JUL 91 |
26 pages |
| Authors:
N. A. Hupp; NAVAL OCEAN SYSTEMS CENTER SAN DIEGO CA
|
| The problem was to determine the ability of a speech recognizer to extract prosodic speech features, such as pitch and stress, and to examine these features for application to future voice recognition systems. The Speech Systems Incorporated (SSI) speech recognizer demonstrated that it could detect prosodic features and that these features do indicate the word and/or syllable that is stressed by the speaker. The research examined the effect of prosodics, ... |
|
| UBA EX-14 MOD 1 Communications/Noise Study |
JUN 91 |
|
| Authors:
Jerry D. Pelton; NAVY EXPERIMENTAL DIVING UNIT PANAMA CITY FL
|
| During Deep Dive 91, several new communications devices for improving communications with divers using the UBA EX-14 MOD 1 diving system were tested. In the past, it has been very difficult to communicate with divers wearing the UBA EX-14 MOD 1 diving helmet due to the noise level in the helmet. This problem is compounded with two divers in the water. |
|
| Modeling the Speech Communication Effect on Performance: Message Complexity |
JUN 91 |
|
| Authors:
Andrew M. Rose; Donna J. Mayo; Janice C. Redish; HUMAN ENGINEERING LAB ABERDEEN PROVING GROUND MD
|
| This document describes a research plan which will lead to the development of a model that relates speech communication to performance. The experiments to be conducted answer questions about the effect of spoken message set complexity on operational performance. A message set has been defined as the set of all possible response-triggering messages within the boundaries imposed by the situation or setting. First-year and long-term objectives are presented. An experimental ... |
|
| The Effects of User's Training on the Performance of an Automatic Speech Recognizer for a Self-Paced Task |
APR 91 |
|
| Authors:
Christopher C. Smyth; HUMAN ENGINEERING LAB ABERDEEN PROVING GROUND MD
|
| The results of a recent experiment concerning the effects of training on the performance of subjects using the automatic speech recognizer are reported. Over a 5-day period, 20 military enlisted grade male subjects were trained and tested in using a connected speech (speaker-dependent) machine automatic speech recognizer in a self-paced task controlling a generic tactical display by voice command. Experimental results show that a majority of the subjects had little ... |
|
| Analysis of the Phoneme Recognition Performance of the Arm Continuous Speech Recognition System |
21 MAR 91 |
36 pages |
| Authors:
S. R. Browning; ROYAL SIGNALS AND RADAR ESTABLISHMENT MALVERN (UNITED KINGDOM)
|
| This memorandum presents the results of a phonetically motivated analysis of the speech recognition system developed as part of the ARM (Airborne Reconnaissance Mission) project. The aim of the work described here is to investigate to what extent errors can be explained by phonetic effects; those which cannot may indicate where models may be improved. The background to the investigation, and the problems of evaluating phoneme recognition performance are described, ... |
|
| Using Genetic Algorithms to Select and Create Features for Pattern Classification |
11 MAR 91 |
|
| Authors:
E. I. Chang; R. P. Lippmann; MASSACHUSETTS INST OF TECH LEXINGTON LINCOLN LAB
|
| Genetic algorithms were used to select and create features and to select reference exemplar patterns for machine vision and speech pattern classification tasks. On a 15-feature machine-vision inspection task, it was found that genetic algorithms performed no better than conventional approaches to feature selection but required much more computation. For a speech recognition task, genetic algorithms required no more computation time than traditional approaches but reduced the number of features ... |
|
| Acoustic Noise Recording and DRT/Dam Source Tape Generation |
MAR 91 |
|
| Authors:
Charles Teacher; John D. Tardelli; Paul D. Gatewood; ARCON CORP WALTHAM MA
|
| This study was aimed at the determination of procedures and techniques by which binaural sound fields created by Air Force vehicles could be accurately recorded in insitu and reproduced in a Rome Laboratory/RL sound chamber in order to enable accurate simulations for the generation of intelligibility and quality source material for the testing of narrowband digital voice communication systems. The sound fields were digitally recorded through a set of spectrally ... |
|
| Speech Recognition and the Telecommunications Emergency Decision Support System |
MAR 91 |
|
| Authors:
Nancy C. Browne; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
| The purpose of this thesis is to provide a feasibility study for incorporating speech recognition into the Telecommunications Emergency Decision Support System (TEDSS) developed by the National Communications System (NCS) and contained on a Compaq 386. The three types of speech recognition systems that were used are: the DragonDictate, a software driven system, the Verbex Series 5000, a system contained in a peripheral device, and the KeyTronic Speech Recognition System, ... |
|
| A Physiologically Motivated Front End for Speech Recognition |
26 FEB 91 |
|
| Authors:
Thao K. Nguyen; Richard P. Lippmann; Bernard Gold; Douglas B. Paul; MASSACHUSETTS INST OF TECH LEXINGTON LINCOLN LAB
|
| A physiological front-end preprocessor for speech recognition was evaluated using a large isolated word database in noisy and quiet environments. The front end was based on the ensemble interval histogram (EIH) model developed by Oded Ghitza, which provides phase or synchrony information similar to that available on the auditory nerve. A modified EIH front end was implemented and tested using a Lincoln hidden Markov model (HMM) isolated word recognizer with ... |
|
| High-Quality 800-b/s Voice Processing Algorithm |
25 FEB 91 |
|
| Authors:
G. S. Kang; L. J. Fransen; NAVAL RESEARCH LAB WASHINGTON DC
|
| The 2400-b/s linear predictive coder (LPC) is currently being widely deployed to support tactical voice communication over narrowband channels. However, there is a need for lower-data-rate voice encoders for special applications: improved performance in high-bit-error conditions, low- probability-of-intercept (LPI) voice communications, and integrated voice/data systems. As a result of continued research, the intelligibility of very-low- data-rate (600 to 800 b/s) voice processors has steadily improved, from the upper 70s about ... |
|
Voice Communications
