• The Journal of the Convergence on Culture Technology
• /
• v.8 no.3
• /
• pp.469-475
• /
• 2022

To train the model of the deep learning-based single-speaker TTS system, a speech DB of tens of hours and a lot of training time are required. This is an inefficient method in terms of time and cost to train multi-speaker or personalized TTS models. The voice cloning method uses a speaker encoder model to make the TTS model of a new speaker. Through the trained speaker encoder model, a speaker embedding vector representing the timbre of the new speaker is created from the small speech data of the new speaker that is not used for training. In this paper, we propose a multi-speaker TTS system to which voice cloning is applied. The proposed TTS system consists of a speaker encoder, synthesizer and vocoder. The speaker encoder applies the d-vector technique used in the speaker recognition field. The timbre of the new speaker is expressed by adding the d-vector derived from the trained speaker encoder as an input to the synthesizer. It can be seen that the performance of the proposed TTS system is excellent from the experimental results derived by the MOS and timbre similarity listening tests.

• 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.

• Proceedings of the KIEE Conference
• /
• 2008.10b
• /
• pp.258-259
• /
• 2008

기존의 유선 연결을 이용한 Hometheater System은 점차 다채널로 발전하는Sound Format의 지원에 따라 복잡한 Speaker 연결 구조를 가지게 되어 오히려 사용자의 불편함을 만들고 이에 따라 일부 Speaker만을 사용하거나, 전면에 모든 Speaker를 배치하여 사용함에 따라 다채널 Speaker의 이점을 활용하지 못하고 있다. 이에 따라 점차 Wireless 기반의 다채널 Speaker에 대한 요구가 증가하고 있는데 현재 Wireless 기반의 각 Speaker Unit에 Data를 전송하는 과정에서 전송 delay가 발생하고 이 문제로 인해 Wireless Speaker의 보급에 빠르게 이루어지지 못하고 있는 상태에다. 이에 따라 본 논문에서는 무선랜을 이용한 무선 홈씨어터 시스템 구현에서 문제가 되는 전송Delay에 대해 보정 알고리즘을 통한 개선 방법을 제안한다.

• MALSORI
• /
• no.63
• /
• pp.125-137
• /
• 2007

Recently, many techniques have been proposed to improve the noise robustness for speaker verification. In this paper, we consider the feature recombination technique in multi-band approach. In the conventional feature recombination for speaker verification, to compute the likelihoods of speaker models or universal background model, whole feature components are used. This computation method is not effective in a view point of multi-band approach. To deal with non-effectiveness of the conventional feature recombination technique, we introduce a subband likelihood computation, and propose a modified feature recombination using subband likelihoods. In decision step of speaker verification system in noise environments, a few very low likelihood scores of a speaker model or universal background model cause speaker verification system to make wrong decision. To overcome this problem, a reliable feature selection method is proposed. The low likelihood scores of unreliable feature are substituted by likelihood scores of the adaptive noise model. In here, this adaptive noise model is estimated by maximum a posteriori adaptation technique using noise features directly obtained from noisy test speech. The proposed method using subband-based reliable feature selection obtains better performance than conventional feature recombination system. The error reduction rate is more than 31 % compared with the feature recombination-based speaker verification system.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.7
• /
• pp.387-394
• /
• 2005

In this paper, we propose a new transmission technique of audio signals for the serial connection of the speakers of multiple-channel audio systems. Analog audio signals from a multi-channel audio system are converted into digital signals with signal processing steps and transferred to each speaker through a serial line. The signal processing steps contain data compression and packet generation in association with audio signal characteristics. Each speaker gets its corresponding digital audio signals from the transmitted packets and converts the signals into analog audio signals to make sounds with the speaker All the proposed functions in this paper are modeled in VHDL. implemented with FPGA chips, and tested for actual multi-channel audio systems.

• 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.

• The Journal of the Acoustical Society of Korea
• /
• v.26 no.6
• /
• pp.269-275
• /
• 2007

Pervasive Multi-channel audio systems are being realized due to advances in digital technology. This paper proposes an efficient system that serially connects individual speakers with bidirectional digital communication capability by means of SoC design. In particular, each speaker can identify the bit stream assigned to the speaker and convert it into analog audio. Furthermore, the speaker can self-diagnose the speaker functionality by utilizing the designed capability to measure frequencies of various square wave test signals. The proposed system running on 200MHz clock yielded restoration of analog output signal with latency of only $500{\mu}s$ compared to directly driving the speakers in a traditional way.

• Speech Sciences
• /
• v.12 no.3
• /
• pp.125-136
• /
• 2005

Automatic speech recognition is much more difficult in real world. Speech recognition according to SIR (Signal to Interface Ratio) is difficult in situations in which noise of surrounding environment and multi-speaker exists. Therefore, study on main speaker's voice extractions a very important field in speech signal processing in binaural sound. In this paper, we used directional filter and harmonic filter among other existing methods to extract the main speaker's information in binaural sound. The main speaker's voice was extracted using directional filter, and other remaining speaker's information was removed using harmonic filter through main speaker's pitch detection. As a result, voice of the main speaker was enhanced.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.29 no.10
• /
• pp.601-607
• /
• 2016

Piezoelectric thick films of soft $Pb(Zr,Ti)O_3$ (PZT) based commercial material (S55) were fabricated using a conventional tape casting method. Ag-Pd electrodes were printed on the piezoelectric film at room temperature and all 5 layered films with a dimension of $12mm{\times}16mm$ were successfully laminated for a multi-layered piezoelectric ceramic actuator. The laminated specimens were co-fired at $1,100^{\circ}C$ for 1 h. A flat layered and dense microstructure was obtained for the $112{\mu}m$ thick piezoelectric actuator after sintering process. Thereafter, a prototype piezoelectric speaker was fabricated using the multi-layered piezoelectric ceramic actuator which can operate as a bimorph. Its SPL (sound pressure level) characteristic was also evaluated for speaker application. Frequency response revealed that the output SPL with a root mean square voltage of 10 V increased gradually to the highest peak of 87.5 dB for 1.5 kHz and exhibited a relatively stable behavior over the measured frequency range (${\leq}20kHz$) at a distance of 10 cm, implying that the fabricated piezoelectric speaker is potential for speaker applications.

• International journal of advanced smart convergence
• /
• v.5 no.4
• /
• pp.21-25
• /
• 2016

Recent development in multi-channel surround is emerging in various formats to provide better stereoscopic and sound effects to consumers in recent broadcasting. The ability sound localize the sound sources in space is most considerable design factor on multi-channel surround system for human earing perception model. However, this paper propose the change of the sound localization according to the spacing of the speakers, which is not covered in the existing research focus on sound system design. Presently the sound system uses the position and number of the speakers to localize the sound. In the multi-channel surround environment, the proposed design uses the sound localization is caused by the directional characteristics of the speaker, the distance between the speakers and the distance between the listener and the speaker according to the directivity is required. The proposed design is simulated using virtual measurement with MATLAB simulation environment and performances are measured.