Recognition of Creaky Voice from Emergency Calls

Abstract

Although creaky voice, or vocal fry, is widely studied phonation mode, open questions still exist in creak's acoustic characterization and automatic recognition. Many questions are open since creak varies significantly depending on conversational context. In this study, we introduce an exploratory creak recognizer based on convolutional neural network (CNN), which is generated specifically for emergency calls. The study focuses on recognition of creaky voice from authentic emergency calls because creak detection could potentially provide information about the caller's emotional state or attempt of voice disguise. We generated the CNN recognition system using emergency call recordings and other out-of-domain speech recordings and compared the results with an already existing and widely used creaky voice detection system: using poor quality emergency call recordings as test data, this system achieved F1 of 0.41 whereas our CNN system accomplished an F1 of 0.64. The results show that the CNN system can perform moderately well using a limited amount of training data on challenging testing data and has the potential to achieve higher F scores when more emergency calls are used for model training.

Classification of Pitch and Gender of Speakers for Forensic Speaker Recognition from Disguised Voices Using Novel Features Learned by Deep Convolutional Neural Networks

“…Some works analyze the difference in feature variations due to disguise for different genders [5,7]. Tavi et al [17] suggests that the effect of speaker's sex on creakiness should be treated carefully. Gonzá lez Hautamä ki et al [31] did an extensive study of how certain features are affected in male and female speakers differently in three voice conditions; modal, intended old and intended child.…”

“…Identifying whether a given test speech is disguised or original is the first step in ASR from disguised voices. In some works, deep features and neural network classifiers are used for this classification [15][16][17][18]. This classification is done in literature using both prosodic and cepstral features [16,[18][19][20][21].…”

“…This classification is done in literature using both prosodic and cepstral features [16,[18][19][20][21]. Specific types of disguises are considered in most of the works like pitch disguised voices [16,[18][19][20], creaky voices [9,17], mimicked voices [15,21] etc. Some of the related works available in literature analyzed various voice features affected by disguise and also robust to disguise.…”

Hädänalaisen puheen prosodia

“…Kolmannessa tutkimuksessa (Tavi, Alumäe & Werner 2019) kehitettiin automaattinen narinaäänenlaadun tunnistin hätäpuheluille. Tunnistin perustui syväoppimismenetelmiin, jotka ovat koneoppimisen eräs kehittynyt alalaji.…”

Reconocimiento del habla con acento español basado en un modelo acústico

“…Por otro lado, el avance de la tecnología ha permitido reconocer el nivel de eficacia que estos sistemas de reconocimiento de voz obtienen en casos específicos. Así, los sistemas RAV basados en redes neuronales convolucionales (convolutional neural network-ConvNet/CNN) que se generan especialmente durante las llamadas de emergencia y cuyo fin es el de detectar el estado emocional y verificar la autenticidad intencional del hablante (Tavi et al, 2019).…”