• The Journal of the Acoustical Society of Korea
• /
• v.40 no.5
• /
• pp.496-502
• /
• 2021

Recently, deep speaker embedding approach has been widely used in text-independent speaker verification, which shows better performance than the traditional i-vector approach. In this work, to improve the deep speaker embedding approach, we propose a novel method called group-based speaker embedding which incorporates group information. We cluster all speakers of the training data into a predefined number of groups in an unsupervised manner, so that a fixed-length group embedding represents the corresponding group. A Group Decision Network (GDN) produces a group weight, and an aggregated group embedding is generated from the weighted sum of the group embeddings and the group weights. Finally, we generate a group-based embedding by adding the aggregated group embedding to the deep speaker embedding. In this way, a speaker embedding can reduce the search space of the speaker identity by incorporating group information, and thereby can flexibly represent a significant number of speakers. We conducted experiments using the VoxCeleb1 database to show that our proposed approach can improve the previous approaches.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.3
• /
• pp.24-33
• /
• 2001

this paper describes a baseline for an implementation of a corpus-based Korean TTS system. The conventional TTS systems using small-sized speech still generate machine-like synthetic speech. To overcome this problem we introduce the corpus-based TTS system which enables to generate natural synthetic speech without prosodic modifications. The corpus should be composed of a natural prosody of source speech and multiple instances of synthesis units. To make a phone level synthesis unit, we train a speech recognizer with the target speech, and then perform an automatic phoneme segmentation. We also detect the fine pitch period using Laryngo graph signals, which is used for prosodic feature extraction. For break strength allocation, 4 levels of break indices are decided as pause length and also attached to phones to reflect prosodic variations in phrase boundaries. To predict the break strength on texts, we utilize the statistical information of POS (Part-of-Speech) sequences. The best triphone sequences are selected by Viterbi search considering the minimization of accumulative Euclidean distance of concatenating distortion. To get high quality synthesis speech applicable to commercial purpose, we introduce a domain specific database. By adding domain specific database to general domain database, we can greatly improve the quality of synthetic speech on specific domain. From the subjective evaluation, the new Korean corpus-based TTS system shows better naturalness than the conventional demisyllable-based one.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.1
• /
• pp.9-18
• /
• 2024

The importance of active sonar systems is emerging due to the quietness of underwater targets and the increase in ambient noise due to the increase in maritime traffic. However, the low signal-to-noise ratio of the echo signal due to multipath propagation of the signal, various clutter, ambient noise and reverberation makes it difficult to identify underwater targets using active sonar. Attempts have been made to apply data-based methods such as machine learning or deep learning to improve the performance of underwater target recognition systems, but it is difficult to collect enough data for training due to the nature of sonar datasets. Methods based on mathematical modeling have been mainly used to compensate for insufficient active sonar data. However, methodologies based on mathematical modeling have limitations in accurately simulating complex underwater phenomena. Therefore, in this paper, we propose a sonar signal synthesis method based on a deep neural network. In order to apply the neural network model to the field of sonar signal synthesis, the proposed method appropriately corrects the attention-based encoder and decoder to the sonar signal, which is the main module of the Tacotron model mainly used in the field of speech synthesis. It is possible to synthesize a signal more similar to the actual signal by training the proposed model using the dataset collected by arranging a simulated target in an actual marine environment. In order to verify the performance of the proposed method, Perceptual evaluation of audio quality test was conducted and within score difference -2.3 was shown compared to actual signal in a total of four different environments. These results prove that the active sonar signal generated by the proposed method approximates the actual signal.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.359-366
• /
• 2022

In this paper, we present the Korean menu-ordering Sentence Text-to-Speech (TTS) system using conformer-based FastSpeech2. Conformer is the convolution-augmented transformer, which was originally proposed in Speech Recognition. Combining two different structures, the Conformer extracts better local and global features. It comprises two half Feed Forward module at the front and the end, sandwiching the Multi-Head Self-Attention module and Convolution module. We introduce the Conformer in Korean TTS, as we know it works well in Korean Speech Recognition. For comparison between transformer-based TTS model and Conformer-based one, we train FastSpeech2 and Conformer-based FastSpeech2. We collected a phoneme-balanced data set and used this for training our models. This corpus comprises not only general conversation, but also menu-ordering conversation consisting mainly of loanwords. This data set is the solution to the current Korean TTS model's degradation in loanwords. As a result of generating a synthesized sound using ParallelWave Gan, the Conformer-based FastSpeech2 achieved superior performance of MOS 4.04. We confirm that the model performance improved when the same structure was changed from transformer to Conformer in the Korean TTS.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.6
• /
• pp.536-543
• /
• 2023

Speaker diarization, which labels for "who spoken when?" in speech with multiple speakers, has been studied on a deep neural network-based end-to-end method for labeling on speech overlap and optimization of speaker diarization models. Most deep neural network-based end-to-end speaker diarization systems perform multi-label classification problem that predicts the labels of all speakers spoken in each frame of speech. However, the performance of the multi-label-based model varies greatly depending on what the threshold is set to. In this paper, it is studied a speaker diarization system using single-label classification so that speaker diarization can be performed without thresholds. The proposed model estimate labels from the output of the model by converting speaker labels into a single label. To consider speaker label permutations in the training, the proposed model is used a combination of Permutation Invariant Training (PIT) loss and cross-entropy loss. In addition, how to add the residual connection structures to model is studied for effective learning of speaker diarization models with deep structures. The experiment used the Librispech database to generate and use simulated noise data for two speakers. When compared with the proposed method and baseline model using the Diarization Error Rate (DER) performance the proposed method can be labeling without threshold, and it has improved performance by about 20.7 %.

• Phonetics and Speech Sciences
• /
• v.16 no.1
• /
• pp.67-76
• /
• 2024

Recent advances in text-to-speech (TTS) technology have significantly improved the quality of synthesized speech, reaching a level where it can closely imitate natural human speech. Especially, TTS models offering various voice characteristics and personalized speech, are widely utilized in fields such as artificial intelligence (AI) tutors, advertising, and video dubbing. Accordingly, in this paper, we propose a one-shot multi-speaker TTS system that can ensure acoustic diversity and synthesize personalized voice by generating speech using unseen target speakers' utterances. The proposed model integrates a speaker encoder into a TTS model consisting of the FastSpeech2 acoustic model and the HiFi-GAN vocoder. The speaker encoder, based on the pre-trained RawNet3, extracts speaker-specific voice features. Furthermore, the proposed approach not only includes an English one-shot multi-speaker TTS but also introduces a Korean one-shot multi-speaker TTS. We evaluate naturalness and speaker similarity of the generated speech using objective and subjective metrics. In the subjective evaluation, the proposed Korean one-shot multi-speaker TTS obtained naturalness mean opinion score (NMOS) of 3.36 and similarity MOS (SMOS) of 3.16. The objective evaluation of the proposed English and Korean one-shot multi-speaker TTS showed a prediction MOS (P-MOS) of 2.54 and 3.74, respectively. These results indicate that the performance of our proposed model is improved over the baseline models in terms of both naturalness and speaker similarity.