Study And Implementation Of Virtual Bass

The decrease of the size and thickness of mobile multimedia devices has strictly limited the size of the speaker. Since the ability of low-frequency replay of the speaker is directly related to its size, the low-frequency performance of the small-sized speaker is not satisfied. Sometimes, it is useful to enhance the loudness of the low-frequency signal. Whereas, it will not only increase the power consumption of the device, but it will also bring about negative impact to the longevity of the device. Another effective method to enhance the low-frequency, which utilizes the specific phenomena of the "lose of base frequency" in psychoacoustics, is "virtual bass". Through producing and replaying the harmonics of the low-frequency signal, the base tone, which is actually not played, could be listened.The main research aspects of the paper are as follows:1, The basic principle and the defects of two kinds of representative virtual base algorithm are analyzed. The MaxxBass algorithm in time-domain produces harmonics via only one feedback multiplication circuit, which will bring about inter-modulation distortion. The VB PhaseVocoder algorithm in frequency-domain could directly calculate the base frequency by changing frequency so as to get multiple harmonics. Obviously, even though the component of the harmonics is "pure", the final effect is susceptible to the calculation accuracy of the base frequency.2, According to the disadvantages of the above two algorithms, a virtual bass algorithm based on ZFFT is proposed. This algorithm zoom the spectrum according to the problem that the frequency component of the intensive spectrum will mutually interfered and further influence the accuracy. The high frequency resolution could be acquired without adding the sample-points of TFT and decreasing the bandwidth. The adaptability of the algorithm is improved and finally, the acquired base frequency is closer to the true value.3, A rapidly-reduce-slowly-rise AGC algorithm is proposed to realize the control of automatic gain according to the saturation distortion of composite signal. The simulation of the algorithm is conducted and the comparison effect is presented between the unprocessed audio and the audio processed by the AGC algorithm in time domain. The final result shows that the AGC algorithm could not only solve the saturation distortion but also preserve the information of the audio itself.4, The paired comparison principle is adopted to make subjective assessment about the loudness of bass, soundness, dynamics and the quality of the tone of the MaxxBass algorithm, VB Phase Vocoder algorithm and virtual bass algorithm based on ZFFT. The subjective assessment result shows that the virtual bass algorithm based on ZFFT is more balanced in each indicator compared with the other two algorithms and the whole quality of the tone is improved.5, The FPGA hardware platform is used to realize this virtual bass algorithm. Since the procession of audio does not require high speed, some approaches which could save more resource is adopted. Meanwhile, the distributed DA algorithm is utilized to realize the multiplications and additions of FIR filter. The virtual bass algorithm consumes small amount of hardware resource and has a broad application prospects.