Digital audio watermarking algorithm based on masking threshold and complex cepstrum;
基于掩蔽阈值与复倒谱的数字音频水印算法
Choosing the frequency from 14 to 16 Bark as the embedding positions which is below the masking threshold,this paper changes the DCT coefficient with the certain strength of the embedded which is determined by the statistic of the DCT coefficient.
文中选用中低频14~16Bark中,在掩蔽阈值以下的频率分量作为嵌入点,对该嵌入点的DCT系数以一定嵌入强度作修改来实现秘密信息的嵌入。
The masking threshold for each audio signal segment is calculated on the basic psychoacoustic model,watermarking is embedded by quantizing DCT low-band and mid-band frequency coefficients of digital audio signal.
提出一种基于听觉掩蔽效应和量化的DCT域音频水印嵌入方案,根据人耳的生理听觉特性,计算音频信号的掩蔽阈值对获取的中低频系数量化后嵌入数字水印图像,量化步长由掩蔽阈值自适应控制,并且水印的提出和检测都不需要原始数字音频信号的参与。
The masking threshold for each audio signal segment was calculated on the basis of psycho-acoustic model.
首先根据心理声学模型 ,计算载体音频信号的掩蔽门限 ,并利用人耳听觉的临界频率与小波包子带间的相似特性 ,将DFT域掩蔽门限映射到小波包域 ;然后将水印信号嵌入到中低频小波包系数中 ,嵌入强度由掩蔽门限自适应控制。
In this algorithm,both the psychoacoustic model and the auditory masking threshold are used to determine the embedding positions of watermarking for the purpose of ensuring the imperceptibility of watermarking,the synchronization code is embedded in the audio signal in time domain to withstand synchronous attacks,and the audio signal is segmented and the watermar.
该算法基于心理声学模型,并利用载体听觉掩蔽阈值选择嵌入位置,以保证水印的隐藏性;同时,在时域嵌入同步码以抵制同步攻击;最后,将音频载体进行分段,在每段音频中自适应嵌入水印,并在每个段首嵌入段号以区别不同音频段。
First, VAD is done to noisy speech, then spectral subtraction based on the perceptual mask threshold follows to achieve noise reduction.
这种新算法源于VAD技术及谱减法,它先对含噪声信号进行VAD处理,而后是基于感知掩蔽阈值的谱减法运算,降低噪声,获得清晰的增强信号。
