GCI Detection from Raw Speech Using a Fully-Convolutional Network

Abstract

Glottal Closure Instants (GCI) detection consists in automatically detecting temporal locations of most significant excitation of the vocal tract from the speech signal. It is used in many speech analysis and processing applications, and various algorithms have been proposed for this purpose. Recently, new approaches using convolutional neural networks have emerged , with encouraging results. Following this trend, we propose a simple approach that performs a regression from the speech waveform to a target signal from which the GCI are easily obtained by peak-picking. However, the ground truth GCI used for training and evaluation are usually extracted from EGG signals, which are not reliable and often not available. To overcome this problem, we propose to train our network on high-quality synthetic speech with perfect ground truth. The performances of the proposed algorithm are compared with three other state-of-the-art approaches using publicly available datasets, and the impact of using controlled synthetic or real speech signals in the training stage is investigated. The experimental results demonstrate that the proposed method obtains similar or better results than other state-of-the-art algorithms and that using large synthetic datasets with many speaker offers better generalization ability than using a smaller database of real speech and EGG signals.

Extraction and Utilization of Excitation Information of Speech: A Review

“…In [279], the GCI detection was posed as a temporal event detection problem, relaxing the constraints used in [278]. In [279] and [280], the GCI detection was formulated using a representation learning perspective, where an appropriate representation is implicitly learned from the raw signal. In [281] and [282], a deep CNN-based GCI detection method was proposed by fusing raw speech and LP residual features.…”

Neural Analysis and Synthesis: Reconstructing Speech from Self-Supervised Representations

“…Recently, it has been reported that the representation from different layers of wav2vec 2.0 exhibit different characteristics. Especially, Shah et al [39] showed that it is the output from the middle layer that has the most relevant characteristics to pronunciation 2 . In light of this empirical observation, we decided to use the intermediate features of XLSR-53.…”

“…Pitch Due to the irregular periodicity of the glottal pulse, we often hear creaky voice in speech, which is usually manifested as jitter or sub-harmonics in signals. This makes hard for f 0 trackers to estimate f 0 because the f 0 itself is not well defined in such cases [16,1,2]. We take a hint from the popular Yin algorithm to address this issue.…”

“…Note that in every voice conversion experiment, we shifted the median pitch of a source utterance to the median pitch of a target utterance by shifting the scope of Yingram. We measured naturalness with 5-scale mean opinion score (MOS [1][2][3][4][5]). Speaker similarity (SSIM (%)) were measured with a binary decision and uncertainty options, following [50].…”

“…In addition, we propose a new feature that can effectively represent and control pitch information. Although it is the fundamental frequency (f 0 ) that is mostly used to represent the pitch information, f 0 is sometimes ill-defined when there exists sub-harmonics in the signal (e.g., vocal fry) [16,1,2]. We address this issue by proposing a controllable but more abstract feature than f 0 that still includes information such as sub-harmonics.…”