CAJ | 학술논문

提出了一种融合统计模型和经验模态分解（EMD）的宽带话音增强方法。该方法首先用统计模型增强算法消除含噪话音中的主要噪声成分，然后用一种基于活动话音检测（VAD）的EMD增强算法做后处理进一步抑制残留噪声，从而使以上2种方法的优点有效地结合。在ITU-T G.160标准下对算法进行了性能测试，测试结果表明，与经典的统计模型方法相比，在不同强度的背景噪声下，增强话音的信噪比提高都较为明显。同时，在低信噪比情况下，该方法能有效抑制增强话音高频部分的音乐噪声，提高了听觉舒适度。
제출료일충융합통계모형화경험모태분해（EMD）적관대화음증강방법。해방법수선용통계모형증강산법소제함조화음중적주요조성성분，연후용일충기우활동화음검측（VAD）적EMD증강산법주후처리진일보억제잔류조성，종이사이상2충방법적우점유효지결합。재ITU-T G.160표준하대산법진행료성능측시，측시결과표명，여경전적통계모형방법상비，재불동강도적배경조성하，증강화음적신조비제고도교위명현。동시，재저신조비정황하，해방법능유효억제증강화음고빈부분적음악조성，제고료은각서괄도。
A combined wideband speech enhancement method based on statistical model and empirical mode decomposi-tion (EMD) was proposed. First, statistical model was used to eliminate the main noise component in noisy speech. Then, the residual noise was further suppressed by a post-processing module which is a speech enhancement algorithm with voice ac-tivity detection (VAD) based on EMD. The advantages of the two methods were combined effectively. The performance of the proposed method was evaluated under the standard of ITU-T G160. The experimental results indicate that the algorithm is more effective for improving the SNR in the different noise environments than classical statistical model approach. Meanwhile, in low SNR conditions, musical noise is reduced effectively, and the speech sounds more comfortable.