Performance evaluation of power-saving audio reproduction

International audienceThe trends of audio reproduction can be divided into three categories: high fidelity reproduction with high resolution for audio enthusiasts, immersive audio reproduction in the professional audio industry, and casual audio reproduction with portable audio players, smartphones, and tablets. Casual music listeners have concerned with the battery life of the mobile device while audio reproduction rather than the quality of audio reproduction. The authors have proposed a power-saving audio playback algorithm based on our auditory characteristics. In this paper, the relationship between perceptual evaluation and power consumption is investigated for both music and speech signals. For the music sources, general attributes used for audio quality assessment are employed to quantified the distortion caused by the proposed method. Concerning the speech signals, the diagnostic rhyme test is carried out to characterize the speech distortion when the reduction rates of power consumption are set at approximately 20 %, 50 %, and 70 %. It is confirmed that the proposed method could reduce the power consumption by 25 % and 50 % within the acceptable distortion for music and speech signals, respectively