Beatrice T. Oshika

The OGI Multi-language Telephone Speech Corpus is designed to support research on automatic language identification and multi-language speech recognition. The corpus consists of up to nine separate responses from each caller, ranging from single words to short topic-specific descriptions to 60 seconds of unconstrained spontaneous speech. The utterances were spoken over commercial telephone lines by speakers of English, Farsi (Persian), French, German, Japanese, Korean, Mandarin Chinese, Spanish, Tamil, and Vietnamese. We have completed the initial phase of our data acquisition effort: the recording and initial verification of utterances produced by 100 different speakers in each of the 10 languages. We describe the recording protocol, data collection procedure, ongoing corpus development, preliminary results of the statistical evaluation of the 10 languages, and plans to provide orthographic transcriptions of the speech. INTRODUCTION Research in multi-language recognition systems wou...

This paper presents some phonological rules which describe systematic pronunciation variation occurring in natural continuous speech. It is argued that a speech understanding system must account for such variation by incorporating phonological rules, either implicitly or explicitly, into the system. Spectrographic evidence for the phonological phenomena described by the rules is included.

The potential benefits of long-term parameter averaging for speaker recognition were investigated. Parameters studied were pitch, gain, and reflection coefficients. Parameter variability was computed over various averaging lengths from one frame averaging (in effect, no averaging) to 1000 frame averaging (about 70 s of speech). It was demonstrated that the between-to-within speaker variance ratio, measured over several speakers, was significantly increased by performing long-term averaging of the parameter sets. The reflection coefficient averages for k <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> and k <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</inf> , respectively, were shown to produce the highest variance ratios.

This paper analyzes inter-labeler agreement of label choice and boundary placement for human phonetic transcriptions of continuous telephone speech in different languages. In experiment one, English, German, Mandarin and Spanish are labeled by fluent speakers of the languages. In experiment two, German and Hindi are labeled by linguists who do not speak the languages. Experiment two uses a somewhat finer phonetic transcription set than experiment one. We compare the transcriptions of the utterances in terms of the minimum number of substitutions, insertions and deletions needed to map one transcription to the other. Native speakers agree on the average 67.52% of the time at the finest level of labeling, including diacritics. Non-native linguists agree 34.41% of the time. The implications of the results are discussed for evaluation of phonetic recognition algorithms. 1. INTRODUCTION Phonetically transcribed continuous speech databases are important for understanding the phonological st...