| 電子情報通信学会 研究会発表申込システム 講演論文 詳細 |
技報閲覧サービス [ログイン] 技報アーカイブ |
| トップに戻る | 前のページに戻る | [Japanese] / [English] |
| 講演抄録／キーワード | ||
| 講演名 | 2026-07-13 10:00 Speech Leakage Suppression Using Spatial Speech Envelope Coding and Active Auditory Masking for Audio-Spot Design with Parametric Array Loudspeakers ○Tomoki Hashida・Yuting Geng・Takanobu Nishiura（Ritsumeikan Univ.）  US2026-16 EA2026-6 |
|
| 抄録 | （和） | A parametric array loudspeaker (PAL) represents the target audible sound only within the propagation area of ultrasound, forming an audio-spot. However, unintended listeners in this area may also perceive the sound, causing speech leakage when the target signal is speech. To address this issue, an auditory masking-based audio-spot design method has been proposed. In this method, the target speech signal is mixed with masker and anti-masker signals and emitted from two intersecting PALs. By spatially controlling auditory masking, audio-spot is formed only at and around the intersection of the normal axes of the two PALs. However, it requires a high masker sound pressure level, which may be perceived as noise. To overcome this limitation, we propose an audio-spot design method that combines spatial speech envelope coding and active auditory masking. The target speech signal is encoded into two complementary signals based on its temporal and spectral envelope information. These signals are individually mixed with the masker and anti-masker signals. At the audio-spot, auditory masking is canceled and the speech envelope is spatially decoded, allowing target speech perception. Outside the audio-spot, intelligibility is degraded by both auditory masking and envelope encoding. Evaluation results based on dictation tests and objective metrics show that the proposed method reduces the required masker sound pressure level while maintaining speech leakage suppression. |
| （英） | A parametric array loudspeaker (PAL) represents the target audible sound only within the propagation area of ultrasound, forming an audio-spot. However, unintended listeners in this area may also perceive the sound, causing speech leakage when the target signal is speech. To address this issue, an auditory masking-based audio-spot design method has been proposed. In this method, the target speech signal is mixed with masker and anti-masker signals and emitted from two intersecting PALs. By spatially controlling auditory masking, audio-spot is formed only at and around the intersection of the normal axes of the two PALs. However, it requires a high masker sound pressure level, which may be perceived as noise. To overcome this limitation, we propose an audio-spot design method that combines spatial speech envelope coding and active auditory masking. The target speech signal is encoded into two complementary signals based on its temporal and spectral envelope information. These signals are individually mixed with the masker and anti-masker signals. At the audio-spot, auditory masking is canceled and the speech envelope is spatially decoded, allowing target speech perception. Outside the audio-spot, intelligibility is degraded by both auditory masking and envelope encoding. Evaluation results based on dictation tests and objective metrics show that the proposed method reduces the required masker sound pressure level while maintaining speech leakage suppression. |
|
| キーワード | （和） | / / / / / / / |
| （英） | Audio-Spot / Speech Leakage / Parametric Array Loudspeaker / Auditory Masking / Speech Envelope / / / | |
| 文献情報 | 信学技報, vol. 126, no. 106, EA2026-6, pp. 1-6, 2026年7月. | |
| 資料番号 | EA2026-6 | |
| 発行日 | 2026-07-06 (US, EA) | |
| ISSN | Online edition: ISSN 2432-6380 | |
| 著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します．(許諾番号：10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) | |
| PDFダウンロード | US2026-16 EA2026-6 |
|
| 研究会情報 | |
| 研究会 | EA US ASJ-H |
| 開催期間 | 2026-07-13 - 2026-07-14 |
| 開催地（和） | オークランド大学 (City Campus) |
| 開催地（英） | |
| テーマ（和） | 応用／電気音響，超音波，聴覚，音声，騒音・振動,一般 |
| テーマ（英） | |
| 講演論文情報の詳細 | |
| 申込み研究会 | EA |
| 会議コード | 2026-07-EA-US-H |
| 本文の言語 | 英語 |
| タイトル（和） | |
| サブタイトル（和） | |
| タイトル（英） | Speech Leakage Suppression Using Spatial Speech Envelope Coding and Active Auditory Masking for Audio-Spot Design with Parametric Array Loudspeakers |
| サブタイトル（英） | |
| キーワード(1)（和/英） | / Audio-Spot |
| キーワード(2)（和/英） | / Speech Leakage |
| キーワード(3)（和/英） | / Parametric Array Loudspeaker |
| キーワード(4)（和/英） | / Auditory Masking |
| キーワード(5)（和/英） | / Speech Envelope |
| キーワード(6)（和/英） | / |
| キーワード(7)（和/英） | / |
| キーワード(8)（和/英） | / |
| 第1著者 氏名（和/英/ヨミ） | 橋田 朋希 / Tomoki Hashida / ハシダ トモキ |
| 第1著者 所属（和/英） | 立命館大学 (略称： 立命館大) Ritsumeikan University (略称： Ritsumeikan Univ.) |
| 第2著者 氏名（和/英/ヨミ） | 耿 毓庭 / Yuting Geng / コウ イクテイ |
| 第2著者 所属（和/英） | 立命館大学 (略称： 立命館大) Ritsumeikan University (略称： Ritsumeikan Univ.) |
| 第3著者 氏名（和/英/ヨミ） | 西浦 敬信 / Takanobu Nishiura / ニシウラ タカノブ |
| 第3著者 所属（和/英） | 立命館大学 (略称： 立命館大) Ritsumeikan University (略称： Ritsumeikan Univ.) |
| 講演者 | 第1著者 |
| 発表日時 | 2026-07-13 10:00:00 |
| 発表時間 | 25分 |
| 申込先研究会 | EA |
| 資料番号 | US2026-16, EA2026-6 |
| 巻番号（vol） | vol.126 |
| 号番号（no） | no.105(US), no.106(EA) |
| ページ範囲 | pp.1-6 |
| ページ数 | 6 |
| 発行日 | 2026-07-06 (US, EA) |
[研究会発表申込システムのトップページに戻る]