| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程系 === 97 === This thesis is mainly devoted to the application of adaptive filter technology to the design and realization of an active noise control (ANC) headphone system built with TI’s TMS320C6416T fixed-point digital signal processor. The pertinent adaptive algorithms for the ANC headphone are developed based upon an integrated feedback ANC (IFBANC) system model which is basically a commonly-used feedback ANC model that accommodates stereo audio input signals. In this ANC headphone, only one microphone is used in each ear-cup to acquire the residual noise component to produce the corresponding anti-noise signal, leading to a more compact and less costly system architecture that is well suited for the ANC headphone applications.
In the initialization phase of the ANC headphone system, the characteristics of the secondary path are first estimated by using artificially generated white random inputs. After the noise cancelling function is activated, slight variations in the characteristics of the secondary path due to a small change in the headphone position are still tracked through the use of an online adaptation algorithm. When the ANC headphone is used for listening to music, the adaptation algorithm can isolate the noise components from the combined signal measured with the microphone in each ear-cup, which leads to the enhancement of the listening quality of the ANC headphone; when not listening to music, the ANC headphone essentially becomes a digital active protection earmuff. We have verified the effectiveness of the ANC algorithm by using MATLAB simulation. Then, the performance of the actual ANC headphone system is evaluated through the use of single-tone sinusoidal noise, dual-tone sinusoidal noise, and the noise of an AC cooling fan as the interfering noise signals. The results of the above experiments has clearly illustrated that the interfering outside noise signals can be significantly reduced in strength without noticeably affecting the audio input signals intended for listening. In particular, the ANC headphone has been shown to be most effective for reducing stationary and low-frequency noise signals.
|
|---|
