Acoustic Modeling in Speech Recognition: A Systematic Review

Abstract

The paper presents a systematic review of acoustic modeling (AM) techniques in speech recognition(SR). Acoustic modeling establishes a relationship between acoustic information and language construct in SR. Over the past decades, researchers presented studies addressing specific concerns in AM. However, all previous research works lack a systematic and comprehensive review of acoustic modeling issues. A systematic review is introduced to understand the acoustic modeling issues in speech recognition. This paper provides an extensive and comprehensive inspection of various researches that have been performed since 1984. The extensive investigation and analysis into AM was performed by getting the relevant data from 73 research works chose after the screening process between the years from 1984 to 2020. The systematic review process was divided into different parts to investigate acoustic modeling issues. Main issues in acoustic modeling such as feature extraction techniques, acoustic modeling units, speech corpora, classification methods, different tools used, language issues applied, and evaluation parameters were investigated. This study helps the reader to understand various acoustic modeling issues with comprehensive details. The research outcomes presented in this study depict research trends and shed light on new research topics in AM. The result of this review can be used to build a better speech recognition system by choosing a suitable acoustic modeling construct in SR.

Deep Models for Low-Resourced Speech Recognition: Livvi-Karelian Case

“…In simple terms, lack of data hampers the direct application of "classical" approaches to automatic speech recognition and translation, which usually imply the use of acoustic, lexical, and grammatical (language) models. Usually, an automatic speech recognition (ASR) system (the "standard" approach) consists of an acoustic model (AM) that establishes the relationship between acoustic information and allophones of a language at issue [21], a language model (LM) necessary for building hypotheses of a recognized utterance, and a vocabulary of lexical units with phonetic transcriptions. The training of acoustic models involves utilizing a speech corpus, while the development of the language model draws upon probabilistic modeling using available target language texts (as illustrated in Figure 1).…”

Performance Comparison between Lab-VIEW and MATLAB on Feature Matching-based Speech Recognition System

“…The combination of the vocal strings with verbalization can create an assortment of discourse. The discourse quality, counting unforgiving, tense, breathy, or whispery voice, can be influenced by emotion and temperament [2]. In the last decade, there has been an automated method of identifying words spoken by the human voice and converting them into readable text.…”

ASRPU: A Programmable Accelerator for Low-Power Automatic Speech Recognition

“…It comes, however, at the cost of performing inferences with huge models that require on the order of billions of arithmetic operations per second of speech and expensive searches in large graphs, such as lexicon and language models graphs. Another challenge for on-edge ASR comes from the wide diversity of ASR systems [13]. ASR systems come in one of two major flavors: Hybrid DNN-HMM systems [14], [15], [16] and End-to-End systems [11], [17], [18], [19], [20].…”