| 講演抄録/キーワード |
| 講演名 |
2014-01-28 11:45
Study on speech water marking based on formant enhancement ○Shengbei Wang・Masashi Unoki(JAIST) EMM2013-98 |
| 抄録 |
(和) |
(まだ登録されていません) |
| (英) |
There have been serious issues concerning the protection of speech signals due to the illegal use of modern digital technologies. Digital watermarking technique can effectively solve this problem by embedding some information into the host signals without degrading the original sound quality. This paper proposes a speech watermarking method based on the concept of formant enhancement. Formants were first derived from a speech signal by linear prediction (LP) analysis, and then were enhanced by symmetrically controlling the line spectral frequencies (LSFs). We evaluated the proposed method by two kinds of objective experiments regarding inaudibility and robustness. The results indicated that the proposed method could satisfy inaudibility. It could also satisfy the robustness against different speech codecs and general processing except some particular processing. To improve the robustness against these particular processing, we then slightly adjusted and optimized the proposed method. The evaluation results of the improved method suggested that our method could finally provide good performance for both inaudibility and robustness. |
| キーワード |
(和) |
/ / / / / / / |
| (英) |
Speech watermarking / formant enhancement / line spectral frequencies / inaudibility / robustness / / / |
| 文献情報 |
信学技報, vol. 113, no. 415, EMM2013-98, pp. 57-62, 2014年1月. |
| 資料番号 |
EMM2013-98 |
| 発行日 |
2014-01-20 (EMM) |
| ISSN |
Print edition: ISSN 0913-5685 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 |
EMM2013-98 |