Search

This paper presents a system of audio signal processing based on FPGA,the system uses audio codec chip LM4550 to A/D transform and D/A transform the input analog audio signal and output digital audio signal.Using FPGA as the high speed signal processor to realize volume adjustment and audio effect control,so it can output different style music.Meantime, the system designs a FFT computing module and control system of VGA display interface,to compute the digital audio signal which is A/D transformed,and real-time display the frequency spectrum of audio signal on VGA.

This paper discusses the issue of tradeoff between image size and performance in PSNR on bit plane image. A bit plane of a digital discrete signal is a set of bits indicate a presented bit plane in each of the binary numbers depicting the image. Image can be converted into a series of binary signal by shattering them up into their bit planes. The 0thbit plane is the least significant bit and the last bit plane is the most significant bit. Experimental results show that the best tradeoff is using the third to the seventh bit planes.

A kind of multifunctional audio digital watermark algorithm was put forward according to the copyright protection of audio problems. The audio signal is layered three by wavelet packet decomposition and robust watermark is embedded in the low frequency part, fragile watermark is embedded in the intermediate frequency part. The experimental results show that, the algorithm has no sound very good, excellent robustness and fragility. In the actual use of the process, which not only guarantees the safety of audio information effectively , but also guarantees good audio information fidelity.

When a digital watermark is directly embedded into low frequency part of audio, sensitive noises maybe occur. In the paper, a method of audio digital watermark based on quantization index modulation is proposed to solve the problem. At first, audio signal is divided to fragment by audio mask effect. Then low frequency part of discrete wavelet transform (DWT) is extracted. Scrambling watermark is embedded into the low frequency part by quantization index modulation (QIM). The experiment results show that the presented method has a good capacity to fight many attacks.

Aiming at content-based audio retrieval (CBAR) applications, a robust audio hashing scheme is proposed. First the audio is divided to frame by fixed length and then low-frequent and high-frequent components are obtained by three-level lifting-based wavelet transformation in every frame. Secondly the audio frame is approximately represented as a product of a base matrix and an encoding matrix, or coefficient matrix, using non-negative matrix factorization (NMF). Finally the sum of each column in the coefficient matrix is calculated, which is then quantized to produce one bit of the hash sequence. Experiment results show that the proposed scheme is robust against Mp3 compression, Real compression, filtering, amplitude compression, equalization, echo, etc. It is insensitive to small local change, and therefore is suitable for distinguishing different audios.