著者

牧野 奨平 坂野 秀樹 旭 健作

出版者

一般社団法人電子情報通信学会

雑誌

電子情報通信学会技術研究報告 = IEICE technical report : 信学技報 (ISSN:09135685)

巻号頁・発行日

vol.113, no.404, pp.61-66, 2014-01-23

---

音声分析合成による鼻声の声質改善を目的とし,鼻声発声の分析を行っている.なお,本研究において,鼻声発声をノーズクリップを装着した状態での発声とし,異なるノーズクリップの装着位置で鼻声発声1・2とした.発声継続時間長が長めの鼻子音を収録し,鼻子音の発声区間での長時間平均スペクトルと,帯域分割した長時間平均スペクトルのパワーの時系列を観察・比較した.その結果,鼻声発声における鼻子音の発声区間は,通常発声に比べ630Hz以上の帯域でパワーが減衰すること,その帯域において,時間経過に伴いパワーが減衰することがわかった.また,鼻声発声することによって,通常発声で見られる400Hz付近の山が100〜200Hzに移動することがわかった.これらの結果から,鼻声発声1の周波数特性の傾向が通常発声と鼻声発声2のものの間にあると考えられる.

著者

旭 健作 小川 明

出版者

一般社団法人 電気学会

雑誌

電気学会論文誌C(電子・情報・システム部門誌) (ISSN:03854221)

巻号頁・発行日

vol.126, no.12, pp.1483-1489, 2006 (Released:2007-03-01)

参考文献数

17

被引用文献数

1

---

Speech recognition systems have come to be used widely. When any speech recognition system is disturbed by surrounding noises, considerable reduction in the recognition rate is inevitable. It is much desired to develop noise reduction methods so that any speech recognition system can be used in realistic environments. We propose a novel scheme especially effective for reducing the noise generated in the vehicles. The reduction is achieved through image processing techniques applied to the corresponding spectrograms. Experiments have been conducted on speech sounds in the vehicles. The performances have been evaluated in terms of the output signal-to-noise ratio (SNR). The proposed scheme has been compared with the conventional spectral subtraction method, and found to be promising especially for speeches corrupted with great amount of car noises.

著者

樋口 寛晃 旭 健作 佐川 雄二 杉江 昇

出版者

一般社団法人 電気学会

雑誌

電気学会論文誌C(電子・情報・システム部門誌) (ISSN:03854221)

巻号頁・発行日

vol.124, no.12, pp.2439-2445, 2004 (Released:2005-03-01)

参考文献数

18

被引用文献数

1 2

---

We propose a method for separating speeches using two spectrograms. First, two spectrograms are generated from voices recorded with a pair of microphones. The onsets and the offsets of the frequency components are extracted as the features using image processing techniques. Then the correspondences of the features between the spectrograms are determined and the intermicrophone time differences are calculated. Each of frequency components with the common onset/offset occurrences and time difference are grouped together as originating one of the speech signals. A set of band-pass filters are generated corresponding to each group of frequency components. Finally, each of the separated speech signals is extracted by applying the set of band-pass filters to the voice signal recorded by a microphone. Experiments were conducted with the mixture of a male speech sound and a female speech sound consisting of Japanese vowel and contain consonant. The evaluation results demonstrated that the separation was done reasonably well with the proposed method.