We propose a neural network model to estimate word translation probabilities for Cross-Lingual Information Retrieval (CLIR). The model estimates better probabilities for word translations than automatic word alignments alone, and generalizes to unseen source-target word pairs. We further improve the lexical neural translation model (and subsequently CLIR), by incorporating source word context, and by encoding the character sequences of input source words to generate translations of out-of-vocabulary words. To be effective, neural network models typically need training on large amounts of data labeled directly on the final task, in this case relevance to queries. In contrast, our approach only requires parallel data to train the translation model, and uses an unsupervised model to compute CLIR relevance scores.

Individual ship detection from underwater audio is the task of deciding whether a specific ship is present, using sound captured by an underwater hydrophone. It is a task analogous to speaker identification (SID), in the sense that it is an open-class detection task; the ships present could be other irrelevant (“impostor”) ships, never encountered in the training data. We present two methodologies for tackling this problem, both motivated by our work in speech-related technologies: (i) one based on neural networks, which follows, to a large extent, the approach of [1], and (ii) one based on i-vectors and PLDA [2]. To the best of our knowledge, this is the first time that the topic of individual ship detection is approached as an open-class detection problem.

We present a three-pronged approach to improving Statistical Machine Translation (SMT), building on recent success in the application of neural networks to SMT. First, we propose new features based on neural networks to model various non-local translation phenomena. Second, we augment the architecture of the neural network with tensor layers that capture important higher-order interaction among the network units. Third, we apply multitask learning to estimate the neural network parameters jointly. Each of our proposed methods results in significant improvements that are complementary. The overall improvement is +2.7 and +1.8 BLEU points for Arabic-English and Chinese-English translation over a state-of-the-art system that already includes neural network features.

Jacob Devlin, Rabih Zbib, Zhongqiang Huang, Thomas Lamar, Richard Schwartz, John Makhoul. Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2014.

As shown in [1, 2], score normalization is of crucial importance for improving the Average Term-Weighted Value (ATWV) measure that is commonly used for evaluating keyword spotting systems. In this paper, we compare three different methods for score normalization within a keyword spotting system that employs phonetic search. We show that a new unsupervised linear fit method results in better-estimated posterior scores, that, when fed into the keyword-specific normalization of [1], result in ATWV gains of 3% on average. Furthermore, when these scores are used as features within a supervised machine learning framework, they result in additional gains of 3.8% on average over the five languages used in the first year of the IARPA-funded project Babel.

In this paper we describe the Vietnamese conversational telephone speech keyword spotting system under the IARPA Babel program for the 2013 evaluation conducted by NIST. The system contains several, recently developed, novel methods that significantly improve speech-to-text and keyword spotting performance such as stacked bottleneck neural network features, white listing, score normalization, and improvements on semi-supervised training methods. These methods resulted in the highest performance for the official IARPA Babel surprise language evaluation of 2013.

We present two techniques that are shown to yield improved Keyword Spotting (KWS) performance when using the ATWV/MTWV performance measures: (i) score normalization, where the scores of different keywords become commensurate with each other and they more closely correspond to the probability of being correct than raw posteriors; and (ii) system combination, where the detections of multiple systems are merged together, and their scores are interpolated with weights which are optimized using MTWV as the maximization criterion. Both score normalization and system combination approaches show that significant gains in ATWV/MTWV can be obtained, sometimes on the order of 8-10 points (absolute), in five different languages. A variant of these methods resulted in the highest performance for the official surprise language evaluation for the IARPA-funded Babel project in April 2013.

This paper investigates the use of several language model adaptation techniques applied to the task of machine translation from Arabic broadcast speech. Unsupervised and discriminative approaches slightly outperform the traditional perplexity-based optimization technique. Language model adaptation, when used for n-best rescoring, improves machine translation performance by 0.3-0.4 BLEU and reduces translation edit rate (TER) by 0.2-0.5% compared to an unadapted LM.

This paper presents a set of experiments that we conducted in order to optimize the performance of an Arabic/English machine translation system on broadcast news and conversational speech data. Proper integration of speech-to-text (STT) and machine translation (MT) requires special attention to issues such as sentence boundary detection, punctuation, STT accuracy, tokenization, conversion of spoken numbers and dates to written form, optimization of MT decoding weights, and scoring. We discuss these issues, and show that a carefully tuned STT/MT integration can lead to significant translation accuracy improvements compared to simply feeding the regular STT output to a text MT system.

In this paper, we present a novel approach for morphological de-composition in large vocabulary Arabic speech recognition. It achieved low out-of-vocabulary (OOV) rate as well as high recognition accuracy in a state-of-the-art Arabic broadcast news transcription system. In this approach, the compound words are decomposed into stems and affixes in both language training and acoustic training data. The decomposed words in the recognition output are re-joined before scoring. Four algorithms are experimented and compared in this work. The best system achieved 1.9% absolute reduction (9.8% relative) in word error rate (WER) when compared to the 64K-word baseline. The recognition performance of this system is also comparable to a 300K-word recognition system trained on the normal words. In the meantime, the decomposed system is much faster in terms of speed and also needs less memory than the systems with larger than 64K vocabularies.

Summary form only given. The state of the art in automatic speech recognition has reached the point that searching for and extracting information from large speech repositories or streaming audio has become a growing reality. This paper summarizes the technologies that have been instrumental in making audio as searchable as text, including speech recognition, speaker clustering, segmentation, and identification; topic classification; and story segmentation. Once speech is turned into text, information extraction methods can then be applied, such as named entity extraction, finding relationships between named entities, and resolution of anaphoric references. Examples of deployed systems for information extraction from speech, which incorporate some of the aforementioned technologies, will be given.

This paper aims to quantify the main error types the 2004 BBN speech recognition system made in the broadcast news (BN) and conversational telephone speech (CTS) DARPA EARS evaluations. We show that many of the remaining errors occur in clusters rather than isolated, have specific causes, and differ to some extent between the BN and CTS domains. The correctly recognized words are also clustered and are highly correlated with regions where the system produces a single hypothesized choice per word. A statistical analysis of some well-known error causes (out-of-vocabulary words, word fragments, hesitations, and unlikely language constructs) was performed in order to assess their contribution to the overall word error rate (WER). We conclude with a discussion of the lower bound on the WER introduced by the human annotator disagreement

This paper describes the progress made in the transcription of broadcast news (BN) and conversational telephone speech (CTS) within the combined BBN/LIMSI system from May 2002 to September 2004. During that period, BBN and LIMSI collaborated in an effort to produce significant reductions in the word error rate (WER), as directed by the aggressive goals of the Effective, Affordable, Reusable, Speech-to-text [Defense Advanced Research Projects Agency (DARPA) EARS] program. The paper focuses on general modeling techniques that led to recognition accuracy improvements, as well as engineering approaches that enabled efficient use of large amounts of training data and fast decoding architectures. Special attention is given on efforts to integrate components of the BBN and LIMSI systems, discussing the tradeoff between speed and accuracy for various system combination strategies. Results on the EARS progress test sets show that the combined BBN/LIMSI system achieved relative reductions of 47% and 51% on the BN and CTS domains, respectively.

This paper describes the BBN English Broadcast News transcription system developed for the EARS Rich Transcription 2004 (RT04) evaluation. In comparison to the BBN RT03 system, we achieved around 22% relative reduction in word error rate for all EARS BN development test sets. The use of additional acoustic training data acquired through Light Supervision based on thousands of hours of found data made the biggest contribution to the improvement. Better audio segmentation, through the use of an online speaker clustering algorithm and chopping speaker turns into moderately long utterances, also contributed substantially to the improvement. Other contributions, even of modest size but adding up nicely, include using discriminative training for all acoustic models, using word duration as an additional knowledge source during N-best rescoring, and using updated lexicon and language models.

We report on the results of the first evaluations for the BBN/LIMSI system under the new DARPA EARS program. The evaluations were carried out for conversational telephone speech (CTS) and broadcast news (BN) for three languages: English, Mandarin, and Arabic. In addition to providing system descriptions and evaluation results, the paper highlights methods that worked well across the two domains and those few that worked well on one domain but not the other. For the BN evaluations, which had to be run under 10 times real-time, we demonstrated that a joint BBN/LIMSI system with a time constraint achieved better results than either system alone.

This paper describes the development of the BBN Audio Indexing System for broadcast news in Arabic. Key issues addressed in this work revolve around the three major components of the audio indexing system: automatic speech recognition, speaker identification, and named entity identification. The system deals with several challenges introduced by the Arabic language, including the absence of short vowels in written text and the presence of compound words that are formed by the concatenation of certain conjunctions, prepositions, articles, and pronouns, as prefixes and suffixes to the word stem. The lack of short vowels in the transcripts prompted a novel solution that further demonstrated the power of hidden Markov models to deal with ambiguity. Another challenge was the acquisition of appropriate language modeling data, given the absence of broadcast news data for that purpose. We present performance results for all three components of the Audio Indexing System, which we believe represent the state of the art for Arabic broadcast news.

We present a novel architecture for providing automated telephone Directory Assistance (DA). The architecture couples a large-vocabulary, statistical n-gram, speech recognition engine with a statistical retrieval system. The use of a statistical n-gram allows for the recognition of unconstrained spoken queries while the statistical retrieval engine allows for an inexact match between a particular spoken query and the training data. Allowing for unconstrained recognition and an inexact match provides the framework for high levels of automation. Once the retrieval engine returns a ranked set of frequently requested telephone numbers (FRN), the rejection module uses a classifier to compute a confidence-like score that is used to make the automation decision. With actual customer calls into an operational, automated DA call center and an FRN set size of 25000 numbers, the new architecture is capable of delivering more than 17% correct automation at a false accept rate of 0.76%.

We present a method for performing optical character recognition (OCR) of text in video images. Recognition of videotext is a challenging problem due to various factors such as the presence of rich, dynamic backgrounds, low resolution, color, etc. Our strategy is to process the video images to produce high-resolution binarized text images that resemble printed text. We describe a novel clustering and relaxation procedure that combines stroke and color information to separate the text from the background. The binarized text image is then recognized with our Byblos OCR engine (Natarajan et al., 2001; Schwartz et al., 1996) using hidden Markov models trained on similar data. We present experimental results on a video-data corpus collected from broadcast news programs. Currently the system delivers a character error rate of 8.3% on independent multi-font test data from this corpus.

This paper presents a script-independent methodology for optical character recognition (OCR) based on the use of hidden Markov models (HMM). The feature extraction, training and recognition components of the system are all designed to be script independent. The training and recognition components were taken without modification from a continuous speech recognition system; the only component that is specific to OCR is the feature extraction component. To port the system to a new language, all that is needed is text image training data from the new language, along with ground truth which gives the identity of the sequences of characters along each line of each text image, without specifying the location of the characters on the image. The parameters of the character HMMs are estimated automatically from the training data, without the need for laborious handwritten rules. The system does not require presegmentation of the data, neither at the word level nor at the character level. Thus, the system is able to handle languages with connected characters in a straightforward manner. The script independence of the system is demonstrated in three languages with different types of script: Arabic, English, and Chinese. The robustness of the system is further demonstrated by testing the system on fax data. An unsupervised adaptation method is then described to improve performance under degraded conditions.

With the advent of essentially unlimited data storage capabilities and with the proliferation of the use of the Internet, it becomes reasonable to imagine a world in which it would be possible to access any of the stored information at will with a few keystrokes or voice commands. Since much of this data will be in the form of speech from various sources, it becomes important to develop the technologies necessary for indexing and browsing such audio data. This paper describes some of the requisite speech and language technologies that would be required and introduces an effort aimed at integrating these technologies into a system, called Rough 'n' Ready, which indexes speech data, creates a structural summarization, and provides tools for browsing the stored data. The technologies highlighted in the paper include speaker-independent continuous speech recognition, speaker segmentation and identification, name spotting, topic classification, story segmentation, and information retrieval. The system automatically segments the continuous audio input stream by speaker, clusters audio segments from the same speaker, identifies speakers known to the system, and transcribes the spoken words. It also segments the input stream into stories, based on their topic content, and locates the names of persons, places, and organizations. These structural features are stored in a database and are used to construct highly selective search queries for retrieving specific content from large audio archives.

We present some recent advances in the BBN BYBLOS OCR system. This OCR system can be used to recognize Arabic, Chinese, and English with high accuracy. A major change in the system is the use of continuous-density HMMs, which allow us to take advantage of a large amount of training data and to use unsupervised adaptation methods to improve accuracy in many cases, e.g., on degraded data. Another advance is the substantial increase in recognition speed. With this increased speed, the system is fast enough for practical use on Arabic and English data. The extension of the system to Chinese further demonstrated the language independence of this system and showed that this system can be used on languages with large character sets and complicated character structures. The Chinese OCR system yielded high accuracy on newspaper data.

This paper is concerned with techniques for performing robust OCR of degraded documents, such us faxed text, using a hidden Markov model (HMM) based OCR system. We present two strategies for dealing with degraded documents. The first strategy is to train the system on degraded documents that have been subjected to the same, or similar, degradation process as the documents to be recognized. The second, more sophisticated, strategy is to use adaptation to adjust the parameters of the trained model in order to improve recognition accuracy on a specific document. This adjustment of model parameters is typically posed as a constrained optimization problem wherein a certain prespecified objective function is to be optimized. We present a comparative study of two objective functions. The likelihood function and the posterior probability. A variation of the basic posterior probability method is also discussed. Using adaptation with a model trained on fax-degraded data we have reduced, by a factor of three, the character error rate on fax-degraded text images generated from the University of Washington English Image Database I.

In this paper, we describe our recent work in fast automatic transcription of broadcast news programming from radio and television. Given our state-of-the-art BBN BYBLOS primary system [1] running at 230 times real time (230xRT) we show that eliminating and approximating many computationally expensive components speeds up the system by a factor of more than 20 without significant loss in recognition accuracy. This is accomplished without retraining or changing the base-line system structure. The components of the primary system which are refined include segmentation, adaptation, decoding, cross-word rescoring with adaptation, and system combination.

We present an omnifont, unlimited-vocabulary OCR system for English and Arabic. The system is based on hidden Markov models (HMM), an approach that has proven to be very successful in the area of automatic speech recognition. We focus on two aspects of the OCR system. First, we address the issue of how to perform OCR on omnifont and multi-style data, such as plain and italic, without the need to have a separate model for each style. The amount of training data from each style, which is used to train a single model, becomes an important issue in the face of the conditional independence assumption inherent in the use of HMMs. We demonstrate mathematically and empirically how to allocate training data among the different styles to alleviate this problem. Second, we show how to use a word-based HMM system to perform character recognition with unlimited vocabulary. The method includes the use of a trigram language model on character sequences. Using all these techniques, we have achieved character error rates of 1.1 percent on data from the University of Washington English Document Image Database and 3.3 percent on data from the DARPA Arabic OCR Corpus.

We address the problem of determining what size test set guarantees statistically significant results in a character recognition task, as a function of the expected error rate. We provide a statistical analysis showing that if, for example, the expected character error rate is around 1 percent, then, with a test set of at least 10,000 statistically independent handwritten characters (which could be obtained by taking 100 characters from each of 100 different writers), we guarantee, with 95 percent confidence, that: (1) the expected value of the character error rate is not worse than 1.25 E, where E is the empirical character error rate of the best recognizer, calculated on the test set; and (2) a difference of 0.3 E between the error rates of two recognizers is significant. We developed this framework with character recognition applications in mind, but it applies as well to speech recognition and to other pattern recognition problems.

This paper describes the speaker adaptive training (SAT) approach for speaker independent (SI) speech recognizers as a method for joint speaker normalization and estimation of the parameters of the SI acoustic models. In SAT, speaker characteristics are modeled explicitly as linear transformations of the SI acoustic parameters. The effect of inter-speaker variability in the training data is reduced, leading to parsimonious acoustic models that represent more accurately the phonetically relevant information of the speech signal. The proposed training method is applied to the Wall Street Journal (WSJ) corpus that consists of multiple training speakers. Experimental results in the context of batch supervised adaptation demonstrate the effectiveness of the proposed method in large vocabulary speech recognition tasks and show that significant reductions in word error rate can be achieved over the common pooled speaker-independent paradigm.

We describe the first sentence understanding system that is completely based on learned methods both for understanding individual sentences, and determining their meaning in the context of preceding sentences. We divide the problem into three stages: semantic parsing, semantic classification, and discourse modeling. Each of these stages requires a different model. When we ran this system on the last test (December, 1994) of the ARPA Air Travel Information System (ATIS) task, we achieved a 13.7% error rate. The error rate for those sentences that are context-independent (class A) was 9.7%.

The authors present a set of techniques for omnifont, unlimited-vocabulary OCR, within the context of a system based on hidden Markov models (HMM). First, they address the issue of how to perform OCR on omnifont and multi-style data, such as plain and italic, without the need to have a separate model for each style. The amount of training data from each style, which is used to train a single model, becomes an important issue in the face of the conditional independence assumption inherent in the use of HMMs. They demonstrate mathematically and empirically how to allocate training data among the different styles to alleviate this problem. Second, they show how to use a word-based HMM system to perform character recognition with unlimited vocabulary. The method includes the use of a trigram language model on character sequences. Using all these techniques, they have achieved character error rates of 1.1% on data from the University of Washington English Document Image Database and 3.3% on data from the DARPA Arabic OCR Corpus.

We describe a new algorithm for topic classification that allows discrimination among thousands of topics. A mixture of topics explicitly models the fact that each story has multiple topics, that different words are related to different topics, and that most of the words are not related to any topic. The resulting model, trained by EM, has sharper distributions of words that result in more accurate topic classification. We tested the algorithm on transcribed broadcast news texts. When trained on one year of stories containing over 5,000 different topics and tested on new (later) stories the first choice topic was among the manually annotated choices 76% of the time.

We use a hidden Markov model (HMM) based continuous speech recognition system to perform off-line character recognition (OCR) of Arabic printed text. The HMM trainer and recognizer are used without change, however we modify the feature extraction stage to compute features relevant to OCR. Although we begin by segmenting the page into a collection of lines, no further segmentation is necessary for either recognition or training. Experiments on the ARPA Arabic data corpus yield a range of character error rates from under one percent for a single computer font to 2.8% for multiple-font recognition of a wide range of material from books, magazines and newspapers.

In this paper we show how continuous speech recognition methods can be used for character recognition resulting in a technology that is language independent and does not require presegmentation of the data at the character and word levels. In multifont experiments on the ARPA Arabic OCR Corpus an average character error rate of 1.9% is obtained using the BBN BYBLOS continuous speech recognition system with no modifications. A first experiment using the identical system and procedures, trained and tested on a subset of the English Univ. of Washington OCR corpus resulted in 1.4% character error.

We present a framework for maximum a posteriori adaptation of large scale HMM speech recognizers. In this framework, we introduce mechanisms that take advantage of correlations present among HMM parameters in order to maximize the number of parameters that can be adapted by a limited number of observations. We are also separately exploring the feasibility of instantaneous adaptation techniques. Instantaneous adaptation attempts to improve recognition on a single sentence, the same sentence that is used to estimate the adaptation. We show that sizable gains (20-40% reduction in error rate) can be achieved by either batch or incremental adaptation for large vocabulary recognition of native speakers. The same techniques cut the error rate for recognition of non-native speakers by factors of 2 to 4, bringing their performance much closer to the native speaker performance. We also demonstrate that good improvements in performance (25-30%) are realized when instantaneous adaptation is used for recognition of non-native speakers.

We incorporate the hierarchical mixtures of experts (HME) method of probability estimation, developed by Jordan (see Neural Computation, 1994), into an HMM-based continuous speech recognition system. The resulting system can be thought of as a continuous-density HMM system, but instead of using Gaussian mixtures, the HME system employs a large set of hierarchically organized but relatively small neural networks to perform the probability density estimation. The hierarchical structure is reminiscent of a decision tree except for two important differences: each "expert" or neural net performs a "soft" decision rather than a hard decision, and, unlike ordinary decision trees, the parameters of all the neural nets in the HME are automatically trainable using the EM algorithm. We report results on the ARPA 5,000-word and 40,000-word Wall Street Journal corpus using HME models.

In this paper, we present several approaches designed to increase the robustness of BYBLOS, the BBN continuous speech, hidden Markov model (HMM) recognition system. We address the problem of increased degradation in performance when there is mismatch in the characteristics of the training and the test microphones. First we compare RASTA processing and cepstrum mean subtraction as preprocessing methods, to compensate for unknown channel transfer function effects, when we have no information about the new microphone. Then we introduce a new algorithm that computes a probabilistic transformation from the training microphone codebook to that of a new microphone, given some information about the new microphone. We test this algorithm in supervised mode and, combined with a microphone selection method, in unsupervised mode. We present experimental results which show that the proposed algorithm combined with cepstrum mean subtraction, improves the recognition accuracy when the system is tested on a microphone with different characteristics than the one on which it was trained.>

We describe recent changes to the BYBLOS system's training and recognition algorithms and report on numerous experiments in large vocabulary speech recognition. In earlier work, we performed five key experiments that were designed to answer questions related to different training scenarios. We investigated (1) the effect of varying the number of training speakers if the total amount of training data remains constant, (2) data pooling versus model averaging for generating speaker-independent (SI) HMMs, (3) the benefit of doubling the acoustic training data, (4) SI versus SD performance when the SI training data is twelve times greater, (5) the effect of cross-domain training for both the acoustic and language models. Our recent work was focused on four specific problem areas sharing the common thread that the test condition exposes the recognizer to phenomena not observed in the training data. Here we investigated (1) words outside the vocabulary, (2) spoken language effects due to subject variability and spontaneous dictation, (3) non-native dialects of the language, and (4) new microphones not used in training.>

A hidden Markov model (HMM) based continuous speech recognition system is applied to on-line cursive handwriting recognition. The base system is unmodified except for using handwriting feature vectors instead of speech. Due to inherent properties of HMMs, segmentation of the handwritten script sentences is unnecessary. A 1.1% word error rate is achieved for a 3050 word lexicon, 52 character, writer-dependent task and 3%-5% word error rates are obtained for six different writers in a 25,595 word lexicon, 86 character, writer-dependent task. Similarities and differences between the continuous speech and on-line cursive handwriting recognition tasks are explored; the handwriting database collected over the past year is described; and specific implementation details of the handwriting system are discussed.>

In this paper, we incorporate the Hierarchical Mixtures of Experts (HME) method of probability estimation, developed by Jordan [1], into an HMM(cid:173) based continuous speech recognition system. The resulting system can be thought of as a continuous-density HMM system, but instead of using gaussian mixtures, the HME system employs a large set of hierarchically organized but relatively small neural networks to perform the probability density estimation. The hierarchical structure is reminiscent of a decision tree except for two important differences: each "expert" or neural net performs a "soft" decision rather than a hard decision, and, unlike ordinary decision trees, the parameters of all the neural nets in the HME are automatically trainable using the EM algorithm. We report results on the ARPA 5,OOO-word and 4O,OOO-word Wall Street Journal corpus using HME models.

The current state-of-the-art in large-vocabulary, continuous speech recognition is based on the use of hidden Markov models (HMM). In an attempt to improve over HMM performance, the authors developed a hybrid system that combines the advantages of neural networks and HMM using a multiple hypothesis (or N-best) paradigm. The connectionist component of the system, the segmental neural net (SNN), models all the frames of a phonetic segment simultaneously, thus overcoming the well-known conditional-independence limitation of the HMM. They describe the hybrid system and discuss various aspects of SNN modeling, including network architectures, training algorithms and context modeling. Finally, they evaluate the hybrid system by performing several speaker-independent experiments with the DARPA Resource Management (RM) corpus, and demonstrate that the hybrid system shows a consistent improvement in performance over the baseline HMM system.

Up till recently, state-of-the-art, large vocabulary, continuous speech recognition (CSR) had employed hidden Markov modeling (HMM) to model speech sounds. In an attempt to improve over HMM we developed a hybrid system that integrates HMM technology with neural networks. We present the concept of a Segmental Neural Net (SNN) for phonetic modeling in CSR. By taking into account all the frames of a phonetic segment simultaneously, the SNN overcomes the well-known conditional-independence limitation of HMMs. We have developed a novel hybrid SNN/HMM system that combines the advantages of SNNs and HMMs using a multiple hypothesis (or N-best) paradigm. In this system, we generate likely phonetic segmentations from the HMM N-best list of word sequences, which are scored by the SNN. The HMM and SNN scores are then combined to optimize performance. In several speaker-independent, 1000-word CSR tests, the error rate for the hybrid system dropped 20% from that of a state-of-the-art HMM system alone.

The authors present the concept of a segmental neural net (SNN) for phonetic modeling in continuous speech recognition (CSR) and demonstrate how this can be used with a multiple hypothesis (or N-Best) paradigm to combine different CSR systems. In particular, the authors developed a system that combines the SNN with a hidden Markov model (HMM) system. In a speaker-independent, 1000-word CSR test using a word-pair grammar, the error rate for the hybrid system dropped 25% from that of a state-of-the-art HMM system alone. By taking into account all the frames of a phonetic segment simultaneously, the SNN overcomes the well-known conditional-independence limitation of HMMs. The hybrid SNN/HMM system generates likely phonetic segmentations from the HMM N-best list, which are scored by the SNN. The HMM and SNN scores are then combined to optimize performance.>

The authors describe four different ways in which they used the N-Best paradigm within the BYBLOS system. The most obvious use is for the efficient integration of speech recognition with a linguistic natural language understanding module. However, the authors have extended this principle to several other acoustic knowledge sources. They also describe a simple and efficient means for investigating and incorporating arbitrary knowledge sources. The N-Best hypotheses are used to provide close alternatives for discriminative training. Finally, the authors have developed a simple technique that allows them to optimize several weights and parameters within a system in a way that directly minimizes word error rate. Examples of each of these uses within the BYBLOS system are described.>

The authors report on the detection of new words for the speaker-dependent and speaker-independent paradigms. A useful operating point in a speaker-dependent paradigm is defined at 71% detection rate and 1% false alarm rate. The authors present a novel technique for obtaining a phonetic transcription for a new word, which is needed to add the new word to the system. The technique utilizes DECtalk's text-to-sound rules to obtain an initial phonetic transcription for the new word. Since these text-to-sound rules are imperfect, a probabilistic transformation technique is used that produces a phonetic pronunciation network of all possible pronunciations given DECtalk's transcription. The network is used to constrain a phonetic recognition process that results in an improved phonetic transcription for the new word. The resulting transcriptions are sufficient for speech recognition purposes.>

Two results are derived concerning the peak (i.e., maximum amplitude) of a causal signal with a given average delay. The first result is that, for an average delay of tau , the maximum possible location of the signal peak is on the order of tau ( tau +3)/2. (This bound can also be interpreted as providing the maximum integer at which the most probable value of a discrete nonnegative random variable could occur, given that the random variable has a known mean.) The second result is that the signals that minimize the peak amplitude, subject to unit energy and average delay tau , have a peak value of the order of 1/ square root (2 tau +1). Causal signals for which the derived bounds are attained for any given real-valued delay are constructed. The derived bounds are compared to the corresponding ones for all-pass signals.>

A preliminary investigation of techniques that automatically detect when a speaker has used a word that is not in the vocabulary of a continuous-speech recognition system is described. A technique that uses a general model for the acoustics of any word to recognize the existence of new words is developed. Using this general word model, the correct detection of new words versus the false alarm rate is measured. Experiments were run using the DARPA 1000-word Resource Management Database indicate a detection rate for new words of 74% with a false alarm rate of 3.4%.>

Several attempts to improve recognition accuracy with the use of supervised clustering techniques are described. These techniques modify the distance metric and/or the clustering procedure in a discrete hidden Markov model recognition system in an attempt to improve phonetic modeling. Three techniques considered are linear discriminant analysis, a hierarchical supervised vector quantization technique, and Kohonen's LVQ2 technique. All experiments were performed on the DARPA resource management speech corpus using the BBN BYBLOS system. Even though the techniques improved the phonetic recognition capability of the vector quantization, the overall word and sentence recognition accuracy did not improve.>

The authors introduce an error criterion for training which improves the performance of neural nets as posterior probability estimators, as compared to using least squares. The proposed criterion is similar to the Kullback-Leibler information measure and is simple to use. A straightforward iterative algorithm for the minimization of the error criterion which has been shown to have good convergence properties is described. The authors applied the proposed technique to some classification examples and showed it to produce better posterior probability estimates than least squares, especially for low probabilities

A method is introduced for pole-zero modeling of spectral envelopes when only a discrete set of spectral values is given. This method, which is called discrete pole-zero (DPZ) modeling, uses a discrete version of the Itakura-Saito distortion measure as its error criterion. The authors introduce an efficient iterative algorithm for the estimation of the pole-zero parameters that is guaranteed to converge to a locally optimal model. They apply DPZ modeling to the estimation of speech spectral envelopes and to the design of digital filters.>

The authors present several techniques to improve an algorithm presented last year for speaker-adaptive training in continuous speech recognition. The previous method uses a transformation matrix to modify the hidden Markov model (HMM) parameters of a prechosen prototype speaker to model a target speaker. To estimate the transformation matrix, it aligns a set of target speech with the same set of speech uttered by the prototype speaker using dynamic time warping. The authors focus on improving the previous method in the modeling of the spectral differences between two speakers, and the accuracy of the alignment. To improve the modeling of the spectral differences, they implemented a phoneme-dependent mapping procedure which transforms the prototype HMMs to the estimated target HMMs using a set of phoneme-dependent matrices. To improve the alignment, the authors developed a modeling of the silence, a linear duration normalization, and an iterative normalization procedure. They tested the new methods in the standard DARPA database with a grammar of perplexity 60. The performance shows a 30% word-error reduction compared to the previous algorithm.>

The authors consider the classification capabilities of feedforward two-layer neural nets with a single hidden layer and having threshold units only; that is they consider the type of decision regions that two-layer nets are capable of forming in the input space. It had been asserted previously that such nets are capable of forming only convex decision regions or nonconvex but connected regions. The authors show that two-layer nets are capable of forming disconnected decision regions as well. In addition to giving examples of the phenomena, they explain why and how disconnected decision regions are formed. They also derive an expression for the number of cells in the input space that are to be grouped together to form the decision regions. This expression can be useful in deciding how many nodes to have in the first layer. The results have bearing on neural networks where the nonlinear elements are smooth (sigmoid) functions rather than threshold functions.>

Three methods for smoothing discrete probability functions in discrete hidden Markov models for large-vocabulary continuous-speech recognition are presented. The smoothing is based on deriving a probabilistic co-occurrence matrix between the different vector-quantized spectra. Each estimated probability density is then multiplied by this matrix, ensuring that none of the probabilities are severely underestimated due to lack of training data. Experimental results show a 20-30% reduction in error rate when this smoothing is used. A word error rate of 3.0% is achieved with the DARPA 1000-word continuous speech recognition database and a word-pair grammar with a perplexity of 60.>

A novel text-dependent probabilistic spectral mapping method is presented for rapid speaker adaptation. The algorithm has been tested on the DARPA 1000-word resource management database with a grammar perplexity of 60. It results in significant better performance than the previous algorithms, and also provides recognition performance which is less than two times the word error rate for speaker-dependent training, using two minutes of adaptation speech.>

The system was trained in a speaker dependent mode on 28 minutes of speech from each of 8 speakers, and was tested on independent test material for each speaker. The system was tested with three artificial grammars spanning a broad perplexity range. The average performance of the system measured in percent word error was: 1.4% for a pattern grammar of perplexity 9, 7.5% for a word-pair grammar of perplexity 62, and 32.4% for a null grammar of perplexity 1000.>

One measure of the additional constraints placed on autocorrelation sequences is the volume of the space they occupy. The ratio of that volume to the volume of the hypercube circumscribing autocorrelation space can be interpreted as the probability that a random sequence restricted to less than one in magnitude is in fact a valid autocorrelation sequence. The author derives a closed-form expression for the volume of the space of an autocorrelation sequence of any order. The derivation throws additional light on the spectral sensitivity of small variations in an autocorrelation sequence near the positive semidefinite boundary.>

In this paper, we describe BYBLOS, the BBN continuous speech recognition system. The system, designed for large vocabulary applications, integrates acoustic, phonetic, lexical, and linguistic knowledge sources to achieve high recognition performance. The basic approach, as described in previous papers [1, 2], makes extensive use of robust context-dependent models of phonetic coarticulation using Hidden Markov Models (HMM). We describe the components of the BYBLOS system, including: signal processing frontend, dictionary, phonetic model training system, word model generator, grammar and decoder. In recognition experiments, we demonstrate consistently high word recognition performance on continuous speech across: speakers, task domains, and grammars of varying complexity. In speaker-dependent mode, where 15 minutes of speech is required for training to a speaker, 98.5% word accuracy has been achieved in continuous speech for a 350-word task, using grammars with perplexity ranging from 30 to 60. With only 15 seconds of training speech we demonstrate performance of 97% using a grammar.

A new method is introduced for parametric modeling of spectral envelopes when only a discrete set of spectral points is given. This method, which we call Discrete All-Pole (DAP) modeling, uses a discrete version of the Itakura-Saito distortion measure as its error criterion. DAP overcomes the limitations of linear prediction and produces better fitting spectral envelopes for spectra that are representable by a relatively small discrete set of values, such as in voiced speech. The method is especially relevant when the pitch frequency is high as in female speech.

The correlation matching property of all-pole models does not extend readily to pole-zero models. We prove that, if the number of poles in a pole-zero model is strictly less than the number of given correlations, a pole-zero model that matches the given correlations may not exist, irrespective of the number of zeroes in the model. We also show that, if we add a set of cepstral constraints, a maximum-entropy pole-zero model that matches the correlation and cepstral constraints may not exist. We conclude that pole-zero modeling based on minimizing some error criterion might be preferable to exact constraint matching.

This paper describes the results of our work in designing a system for large-vocabulary word recognition of continuous speech. We generalize the use of context-dependent Hidden Markov Models (HMM) of phonemes to take into account word-dependent coarticulatory effects, Robustness is assured by smoothing the detailed word-dependent models with less detailed but more robust models. We describe training and recognition algorithms for HMMs of phonemes-in-context. On a task with a 334-word vocabulary and no grammar (i.e., a branching factor of 334), in speaker-dependent mode, we show an average reduction in word error rate from 24% using context-independent phoneme models, to 10% when using robust context-dependent phoneme models.

Time-scale modification (TSM) of speech is one potential method for speech data reduction at medium to low bit rates. In this application, the speech waveform is compressed in time prior to coding, then expanded to its original speed at the receiver. In this paper, we apply the synchronized-overlap-add (SOLA) method of TSM, developed at BBN, to speech coding at 9.6 and 4.8 kbits/s. These systems were obtained by combining TSM with other established speech coding methods.

This paper describes the results of our work in designing a system for phonetic recognition of unrestricted continuous speech. We describe several algorithms used to recognize phonemes using context-dependent Hidden Markov Models of the phonemes. We present results for several variations of the parameters of the algorithms. In addition, we propose a technique that makes it possible to integrate traditional acoustic-phonetic features into a hidden Markov process. The categorical decisions usually associated with heuristic acoustic-phonetic algorithms are replaced by automated training techniques and global search strategies. The combination of general spectral information and specific acoustic-phonetic features is shown to result in more accurate phonetic recognition than either representation by itself.

Quantization, the process of approximating continuous-amplitude signals by digital (discrete-amplitude) signals, is an important aspect of data compression or coding, the field concerned with the reduction of the number of bits necessary to transmit or store analog data, subject to a distortion or fidelity criterion. The independent quantization of each signal value or parameter is termed scalar quantization, while the joint quantization of a block of parameters is termed block or vector quantization. This tutorial review presents the basic concepts employed in vector quantization and gives a realistic assessment of its benefits and costs when compared to scalar quantization. Vector quantization is presented as a process of redundancy removal that makes effective use of four interrelated properties of vector parameters: linear dependency (correlation), nonlinear dependency, shape of the probability density function (pdf), and vector dimensionality itself. In contrast, scalar quantization can utilize effectively only linear dependency and pdf shape. The basic concepts are illustrated by means of simple examples and the theoretical limits of vector quantizer performance are reviewed, based on results from rate-distortion theory. Practical issues relating to quantizer design, implementation, and performance in actual applications are explored. While many of the methods presented are quite general and can be used for the coding of arbitrary signals, this paper focuses primarily on the coding of speech signals and parameters.

This paper discusses the use of the Hidden Markov Model (HMM) in phonetic recognition. In particular, we present improvements that deal with the problems of modeling the effect of phonetic context and the problem of robust pdf estimation. The effect of phonetic context is taken into account by conditioning the probability density functions (pdfs) of the acoustic parameters on the adjacent phonemes, only to the extent that there are sufficient tokens of the phoneme in that context. This partial conditioning is achieved by combining the conditioned and unconditioned pdfs models with weights that depend on the confidence in each pdf estimate. This combination is shown to result in better performance than either model by itself. We also show that it is possible to obtain the computational advantages of using discrete probability densities without the usual requirement for large amounts of training data.

In this paper we investigate several methods for reducing the bit rate of a segment vocoder [1] by 35% to 150 b/s. In the original vocoder we used a random sample of vectors as a set of templates for vector quantization. We demonstrate in this paper that this random quantizer is near-optimal by comparing it with quantizers that use clustering algorithms for quantizing speech segments. The reduction of the bit rate of the segment vocoder was achieved primarily by using a segment network, i.e., not all segment templates are allowed to follow a given segment template. The spectral continuity of speech is used to determine the subset of templates, that can be used to quantize an input segment. To achieve the low rate of 150 b/s, we also reduced the bit rate for coding pitch, gain, and segment duration. Finally, we present the bit allocation used for transmitting speech at 150 b/s as a single speaker segment vocoder.

We present a method that reduces the bit rate of a low rate LPC vocoder by modelling the sequence of quantized spectra by a Markov chain. To minimize the bit rate, one would want to use a high-order chain. Unfortunately, a high-order chain would require an inordinate amount of data for training. We describe in this paper the use of a variable order Markov chain that maximizes the effective use of a given amount of speech data. To reduce the number of states of a high-order chain, we define an equivalence relation on the states, i.e., "similar" states are grouped together in an equivalence class and a single conditional distribution is associated with the equivalence class. We introduce two equivalence relations. In the first, called variable order Markov chain, the equivalence classes represent the most probable states of any order. In the second method, called variable resolution, the equivalence class is obtained by decreasing the quantization accuracy in representing a spectrum that belongs to a more remote past. For an LPC vocoder with 64 possible spectra (using 6-bit vector quantization), the second method is superior to the first and decreases the entropy from 6 bits to 4 bits per spectrum with 256 equivalence classes.

We introduce a new method for very-low-rate vocoding that the input speech as a sequence of variable-length segments. A segment is a by a spectrum of frames, where each frame is represented by a spectrum, pitch and gain. We use an automatic segmentation algorithm to obtain segments with an average duration comparable to that of a phoneme. A segment is quantized as a single block. The distance measure used for quantization incooporates the appropriate time alignment of two segments. We employ a computationally efficient metric that does not use the usual dynamic programming time warping. Two basic vocoders using the above approach of block quantization have been used to transmit intelligible speech at 200 b/s.

We review the variable frame rate (VFR) transmission methodology that we developed, implemented, and tested during the period 1973-1978 for efficiently transmitting LPC vocoder parameters extracted from the input speech at a fixed frame rate. In the VFR method, parameters are transmitted only when their values have changed sufficiently over the interval since their preceding transmission. We explored two distinct approaches to automatic implementation of the VFR method. The first approach bases the transmission decisions on comparisons of the parameter values of the present frame and the last transmitted frame. The second approach, which is based on a functional perceptual model of speech, compares the parameter values of all the frames that lie in the interval between the present frame and the last transmitted frame against a linear model of parameter variation over that interval. The application of VFR transmission to the design of narrow-band LPC speech coders with average bit rates of 2000-2400 bits/s is also considered. The transmission decisions are made separately for the three sets of LPC parameters, pitch, gain, and spectral parameters, using separate VFR schemes. A formal subjective spccch quality test of six selected LPC coders is described, and the results are presented and analyzed in detail. It is shown that a 2075 bit/s VFR coder produces speech quality equal to or better than that of a 5700 bit/s fixed frame rate coder.

Adaptive predictive coding (APC) is a useful technique for high-quality digital encoding of speech signals at medium band data rates. A particularly annoying degradation in APC systems, however, is the presence of "glitches" or "beeps" in the output speech. These correspond to frames with signal-to-quantization-noise ratios (S/Q) that are often less than unity, i.e., negative in dB. We show that these degradations are caused by instabilities of the APC system. This paper provides a method for analyzing the stability of APC systems as a function of the encoding bit rate, the quantization algorithm, and the power gain of the APC loop feedback filter. Two methods of improving system performance based on the results of the stability analysis are then developed. Experimental evaluation by listening tests confirm the validity of the techniques.

We present a general framework for the coding of signals with a time-varying short-term spectrum. The total quantization distortion is minimized by a dynamic bit allocation in time and in frequency. The resulting distortion is then white noise with fixed power. Such noise, however, is not optimal perceptually when coding speech signals. Spectral and temporal noise shaping are then used to minimize the perception of noise in frequency and in time. Initial experiments at 16 kb/s demonstrate the validity of shaping the noise in time as well as in frequency.

In embedded multirate speech coding, it is desired to have a transmitter/receiver system that operates efficiently over a wide range of channel transmission rates. We are currently investigating a system based on adaptive transform coding of speech, in which we code and transmit the system parameters and the discrete cosine transform (DCT) coefficients of the fullband linear prediction residual waveform. The multirate property of the system is achieved by allowing the channel to discard some of the bits generated by the high data rate transmitter. Stripping off bits results in an absence of DCT components, which the receiver regenerates by a spectral duplication method. An inverse DCT at the receiver yields the time domain residual waveform to be used as input to the linear prediction synthesis filter. The lowest data rate achievable by the system is about 2.5 kb/s, in which case the system reduces to a narrowband LPC pitch-excited vocoder.

We report on the initial development of a phonetic vocoder operating at 100 b/s. With each phoneme, the vocoder transmits the duration and a single pitch value. The synthesizer uses a large inventory of diphone templates to synthesize a desired phoneme string. To determine a phoneme string from input speech, the analyzer takes into account the synthesis model by using the same inventory of diphone templates, augmented by additional diphone templates to account for alternate pronunciations. The phoneme string is chosen to minimize the difference between the diphone templates and the input speech according to a distance measure.

This paper presents the results of our investigation of the various aspects of baseband LPC coders with the goal of maximizing the speech quality at a transmission bit-rate of 9.6 kb/s and for channel bit-error rates of up to 1%. Important among these aspects are: baseband width, coding of baseband, high-frequency regeneration, and error protection of important transmission parameters. The paper discusses these and other issues, presents the results of speech-quality tests conducted during the various stages of optimization, and describes the details of the optimized speech coder.

This paper considers the optimization of speech quality of adaptive predictive coding (APC) systems for transmission over a synchronous 16 kb/s channel. Among the important issues included in this on-going optimization study are: comparative evaluation of several methods for adaptive coding of the APC residual; comparative testing of several methods for adaptive shaping of the spectrum of the quantization noise; and optimization of various parameter values and their bit allocation. In addition to reporting the results from this study, we report on the occurrence of "limit cycles" or regions of excessive quantization noise build-up, offer an explanation for their cause in terms of the feedback gain (or "loop gain") of the APC transmitter, and present experimental results obtained using several remedial means for this problem. Also, we report on the relative properties of different APC coder configurations. Based on the results of our optimization study to date, we present several optimized APC systems.

This paper describes a method for enhancing speech corrupted by broadband noise. The method is based on the spectral noise subtraction method. The original method entails subtracting an estimate of the noise power spectrum from the speech power spectrum, setting negative differences to zero, recombining the new power spectrum with the original phase, and then reconstructing the time waveform. While this method reduces the broadband noise, it also usually introduces an annoying "musical noise". We have devised a method that eliminates this "musical noise" while further reducing the background noise. The method consists in subtracting an overestimate of the noise power spectrum, and preventing the resultant spectral components from going below a preset minimum level (spectral floor). The method can automatically adapt to a wide range of signal-to-noise ratios, as long as a reasonable estimate of the noise spectrum can be obtained. Extensive listening tests were performed to determine the quality and intelligibility of speech enhanced by our method. Listeners unanimously preferred the quality of the processed speech. Also, for an input signal-to-noise ratio of 5 dB, there was no loss of intelligibility associated with the enhancement technique.

Recently we described a variable-frame-rate LPC vocoder designed to transmit good quality speech over 2400 bps fixed-rate noisy channels with bit-error probabilities ranging up to 5% [3]. The basic idea was to lower the data rate by transmitting LPC parameters only when speech characteristics have changed sufficiently since the last transmission, and to employ the resulting bit-rate savings for protecting important transmission data against channel noise. This paper describes our continuing efforts which have concentrated on minimizing loss of synchronization between the receiver and the transmitter. In one approach, we emphasize heavy protection of header, and rapid resynchronization. Alternatively, we apply constraints which guarantee synchronization at a cost of some freedom in the selection of data for transmission. Results from the first approach are presented; results from both methods will be compared at the conference.

The traditional method of high-frequency regeneration (HFR) of the excitation signal in baseband coders has been to rectify the transmitted baseband, followed by spectral flattening. In addition, a noise source is added at high frequencies to compensate for lack of energy during certain sounds. In this paper, we reexamine the whole HFR process. We show that the degree of rectification does not affect the output speech, and that, with proper processing, the high-frequency noise source may be eliminated. We introduce a new type of HFR based on spectral duplication of the baseband. Two types of spectral duplication are presented: spectral folding and spectral translation. Finally, in order to eliminate the problem of breaking the harmonic structure due to spectral duplication, we propose a pitch-adaptive spectral duplication scheme in the frequency domain by using adaptive transform coding to code the baseband.

We report on the synthesis of speech in the context of a phonetic vocoder operating at 100 b/s. With each phoneme, the vocoder transmits the duration and a single pitch value. The synthesizer uses a large inventory of diphone "models" to synthesize a desired phoneme string. The diphone inventory has been selected to differentiate between prevocalic and postvocalic allophones of sonorants, to account for changes in vowel color conditioned by postvocalic liquids, to allow exact specification of voice onset time, and to permit synthesis of glottal stops alveolar flaps and syllabic consonants. The diphones are extracted from carefully constructed short utterances and are stored as a sequence of LPC parameters. During synthesis, the requisite diphone models are time-warped, abutted and smoothed to produce a complete sequence of LPC parameters that are used in the synthesis. The algorithms used are described and compared with more conventional methods. Examples of the synthesized speech will be played.

This paper considers the use of voice-excited linear predictive (LPC) coders for speech transmission at a bit-rate of 9.6 kbps. In our on-going work, we study in detail the various aspects of this class of speech coders, with the goal of maximizing the speech quality at the above rate. Important among these aspects are: baseband residual versus baseband speech transmission, coding of the baseband signal, and high-frequency regeneration from the baseband. We provide a discussion of these and other issues, and indicate a number of variables that have been included in our speech-quality optimization study. Experimental results obtained to date are summarized in the paper. More complete results and conclusions will be presented at the conference.

We report on the results of research to code speech at 16 kbps under the condition that the quality of the transmitted speech be equal to that of the original. Some of the original speech had been corrupted by noise and distortions typical of long distance telephone lines. The rigorous requirements of this work led to a new outlook on adaptive predictive coding. We have found that the pitch predictor is not useful on balance and should be eliminated, and that the residual should be quantized with no clipping and encoded using a variable-length code. A single coding scheme seems to be adequate for all speech and all conditions. In addition, the adaptive predictive coding system has been modified to include a noise spectral shaping filter that effectively eliminates the perception of background granular noise.

A general method for adaptive updating of lattice coefficients in the linear predictive analysis of nonstationary signals is presented. The method is given as one of two sequential estimation methods, the other being a block sequential estimation method. The fast convergence of adaptive lattice algorithms is seen to be due to the orthogonalization and decoupling properties of the lattice. These properties are useful in adaptive Wiener filtering. As an application, a new fast start-up equalizer structure is presented. In addition, a one-multiplier form of the lattice is presented, which results in a reduction of computations.

This paper presents an excitation source model for speech compression and synthesis, which allows for a degree of voicing by mixing voiced (pulse) and unvoiced (noise) excitations in a frequency-selective manner. The mix is achieved by dividing the speech spectrum into two regions, with the pulse source exciting the low-frequency region and the noise source exciting the high-frequency region. A parameter Fcdetermines the degree of voicing by specifying the cut-off frequency between the voiced and unvoiced regions. For speech compression applications, Fccan be extracted automatically from the speech spectrum and transmitted. Experiments using the new model indicate its power in synthesizing natural sounding voiced fricatives, and in largely eliminating the "buzzy" quality of vocoded speech. A functional definition of buzziness and naturalness is given in terms of the model.

Several methods are presented for the objective speech quality evaluation of narrowband LPC vocoders, based on a framework that we proposed at the 1976 ICASSP conference. In each method, the error in short-term spectral behavior between vocoded speech and the original is computed once every 10 ms. These errors are appropriately weighted and averaged over an utterance to produce a single objective score. Several short-term error measures, and time-weighting and averaging techniques are investigated. We evaluate the objective methods by correlating the resulting objective scores with formal subjective speech quality judgments. High correlations obtained indicate the usefulness of these methods.

The spectral distortion of speech signals, without affecting the pitch or the speed of the signal, has met with some difficulty due to the need for pitch extraction. This paper presents a general analysis-synthesis scheme for the arbitrary spectral distortion of speech signals without the need for pitch extraction. Linear predictive warping, cepstral warping, and autocorrelation warping, are given as examples of the general scheme. Applications include the unscrambling of helium speech, spectral compression for the hard of hearing, bit rate reduction in speech compression systems, and efficiency of spectral representation for speech recognition systems.

This paper presents a new formulation for linear prediction, which we call the covariance lattice method. The method is viewed as one of a class of lattice methods which guarantee the stability of the all-pole filter, with or without windowing of the signal, with finite wordlength computations, and with the number of computations being comparable to the traditional autocorrelation and covariance methods. In addition, quantization of the reflection coefficients can be accomplished within the recursion for retention of accuracy in representation.

In ordinary linear prediction the speech spectral envelope is modeled by an all-pole spectrum. The error criterion employed guarantees a uniform fit across the whole frequency range. However, we know from speech perception studies that low frequencies are more important than high frequencies for perception. Therefore, a minimally redundant model would strive to achieve a uniform perceptual fit across the spectrum, which means that it should be able to represent low frequencies more accurately than high frequencies. This is achieved in the LPCW vocoder: an LPC vocoder employing our recently developed method of linear predictive warping (LPW). The result is improved speech quality for the same bit rate.

While there exist methods in the literature for objectively evaluating the intelligibility of speech in the presence of stationary noise, little has been done regarding the objective evaluation of either the intelligibility or the quality of vocoded speech. We present a framework within which we have begun a step-by-step program to develop objective measures for vocoded speech quality that are consistent with results from subjective tests.

This paper considers distance measures for determining the deviation between two smoothed short-time speech spectra. Since such distance measures are employed in speech processing applications that either involve or relate to human perceptual judgment, the effectiveness of these measures will be enhanced if they provide results consistent with human speech perception. As a first step, we suggest Flanagan's results on difference limens for formant frequencies as one basis for checking the perceptual consistency of a measure. A general necessary condition for perceptual consistency is derived for a class of spectral distance measures. A class of perceptually consistent measures obtained through experimental investigations is then described, and results obtained using one such measure under Flanagan's test conditions are presented.

This study examines the feasibility and limitations of speaker adaptation in improving the performance of a fixed (speaker-independent) automatic speech recognition system. A fixed vocabulary of 55 [ƏCVd] syllables is used in the recognition system, where C is 1 of 11 stops and fricatives, and V is 1 of 5 tense vowels. The results of the experiment on speaker adaptation, performed with 6 male and 6 female adult speakers, show that speakers can learn to change their articulations to appreciably improve recognition scores. The preliminary results obtained also indicate that most of the necessary adaptation can be achieved in a relatively short time, provided that the speakers are instructed in how to change their articulations to produce the desired effects.