• The Journal of the Acoustical Society of Korea
• /
• v.13 no.3
• /
• pp.33-40
• /
• 1994

Code Excited Linear Prediction(CELP) speech coders exhibit good performance at data rates below 4.8 kbps. This major drawback of CELP type coders is required much computation. In this paper, we propose a new pitch search method that preserves the quality of the CELP vocoder with reducing complexity. In the pitch searching, we detect the segments of high correlation by a simple preprocessing, and then carry out the pitch searching only for the segments obtained by the preprocessing. By using the proposed method, we can get approximately $77\%$ complexity reduction in the pitch search.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.38 no.5
• /
• pp.576-582
• /
• 2001

The Multi-Band Excitation(MBE) speech coder uses a different approach for the representation of the excitation signal. It replaces the frame-based single voiced/unvoiced classification of a classical speech coder with a set of such decision over harmonic intervals in the frequency domain. This enables each speech segment to be a mixture of voiced and unvoiced, and improves the synthetic speech quality by reducing decision errors that might occur on the frame-based single voiced and unvoiced decision process when input speech is degraded with noise. The IMBE-LP, improved version of MBE with linear prediction, represents the spectral information of MBE model with linear prediction coefficients to obtain low bit rate of 2.4 kbps. In this Paper, we proposed a variable rate IMBE-LP vocoder that has lower bit rate than IMBE-LP without degrading the synthetic speech quality. To determine the LP order, it uses the spectral band information of the MBE model that has something to do with he input speech's characteristics. Experimental results are riven with our findings and discussions.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.2
• /
• pp.173-181
• /
• 2002

Generally the one of serious problems in Voice Activity Detection (VAD) is speech region detection in noise environment. Therefore, this paper propose the new method using energy, lsp varation. As a result of processing time and speech quality of the proposed algorithm, the processing time is reduced due to the accurate detection of inactive period, and there is almot no difference in the subjective quality test. As a result of bit rate, proposed algorithm measures the number of VAD=1 and the result shows predominant reduction of bit rate as SNR of noisy speech is low (about 5∼10 dB).

• The Journal of the Acoustical Society of Korea
• /
• v.14 no.2E
• /
• pp.33-36
• /
• 1995

Based on the Characteristics of the correlation function of speech signal, the skipping technique can reduced the computation time considerably with a little degradation of speech quality. To improve the speech quality of the skipping technique, we use the reduced form of the correlation function to check the sign of the correlation value before the match score is calculated. The experimental results show that this modified skipping technique can reduce the computation time in pitch search over 35% compared with the traditional full search method without quality degradation.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.453-455
• /
• 2020

본 논문은 시각 장애인을 위한 감정 음성 자막 서비스를 생성하는 종단 간(end-to-end) 감정 음성 합성 시스템(emotional text-to-speech synthesis system, TTS)의 음성 합성 속도를 높이면서도 합성음의 음질을 향상시키는 방법을 제안한다. 기존에 사용했던 전역 스타일 토큰(Global Style Token, GST)을 이용한 감정 음성 합성 방법은 다양한 감정을 표현할 수 있는 장점을 갖고 있으나, 합성음을 생성하는데 필요한 시간이 길고 학습할 데이터의 동적 영역을 효과적으로 처리하지 않으면 합성음에 클리핑(clipping) 현상이 발생하는 등 음질이 저하되는 양상을 보였다. 이를 보안하기 위해 본 논문에서는 새로운 데이터 전처리 과정을 도입하였고 기존의 보코더(vocoder)인 웨이브넷(WaveNet)을 웨이브알엔엔(WaveRNN)으로 대체하여 생성 속도와 음질 측면에서 개선됨을 보였다.

• 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.

• Phonetics and Speech Sciences
• /
• v.12 no.3
• /
• pp.55-63
• /
• 2020

Voice conversion can be applied to various voice processing applications. It can also play an important role in data augmentation for speech recognition. The conventional method uses the architecture of voice conversion with speech synthesis, with Mel filter bank as the main parameter. Mel filter bank is well-suited for quick computation of neural networks but cannot be converted into a high-quality waveform without the aid of a vocoder. Further, it is not effective in terms of obtaining data for speech recognition. In this paper, we focus on performing voice-to-voice conversion using only the raw spectrum. We propose a deep learning model based on the transformer network, which quickly learns the voice conversion properties using an attention mechanism between source and target spectral components. The experiments were performed on TIDIGITS data, a series of numbers spoken by an English speaker. The conversion voices were evaluated for naturalness and similarity using mean opinion score (MOS) obtained from 30 participants. Our final results yielded 3.52±0.22 for naturalness and 3.89±0.19 for similarity.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.4
• /
• pp.675-681
• /
• 2021

With the recent growth of the AI speaker market, the demand for speech synthesis technology that enables natural conversation with users is increasing. Therefore, there is a need for a multi-speaker speech synthesis system that can generate voices of various tones. In order to synthesize natural speech, it is required to train with a large-capacity. high-quality speech DB. However, it is very difficult in terms of recording time and cost to collect a high-quality, large-capacity speech database uttered by many speakers. Therefore, it is necessary to train the speech synthesis system using the speech DB of a very large number of speakers with a small amount of training data for each speaker, and a technique for naturally expressing the tone and rhyme of multiple speakers is required. In this paper, we propose a technology for constructing a speaker encoder by applying the deep learning-based x-vector technique used in speaker recognition technology, and synthesizing a new speaker's tone with a small amount of data through the speaker encoder. In the multi-speaker speech synthesis system, the module for synthesizing mel-spectrogram from input text is composed of Tacotron2, and the vocoder generating synthesized speech consists of WaveNet with mixture of logistic distributions applied. The x-vector extracted from the trained speaker embedding neural networks is added to Tacotron2 as an input to express the desired speaker's tone.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.6
• /
• pp.61-70
• /
• 2003

In this paper, a transcoding algorithm for the Selectable Mode Vocoder (SMV) and the G.723.1 speech coder via direct parameter transformation is proposed. In contrast to the conventional tandem transcoding algorithm, the proposed algorithm converts the parameters of one coder to the other without going through the decoding and encoding process. The proposed algorithm is composed of four parts: the parameter decoding, line spectral pair (LSP) conversion, pitch period conversion, excitation conversion and rate selection. The evaluation results show that the proposed algorithm achieves equivalent speech quality to that of tandem transcoding with reduced computational complexity and delay.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.6
• /
• pp.1430-1437
• /
• 1996

A low delay code excited linear predictive speech coder(LD-CELP) at bit rates under 8kbps is considered. We try to improve the perfomance of speech coder with frame type dependent modification of synthesis filter. We first classify frames into 3 groups: voiced, unvoiced and onset. For voicedand unvoiced frame, the spectral envelope of the synthesis filter is adapted to the phonetic characteristics. For transition frame from unvoiced to voiced, the synthesis filter which has been interpolated with the bias filter is used. The proposed vocoder produced more clear sound with similar delay level than other pre-existing LD-CELP vocoders.