基于时基调制算法的音频密写标签技术研究

Research on Audio Hiding Written Label Technology Based on Time-scale Modification

研究基于时基调制的数字音频密写标签的嵌入及盲检测提取技术,该技术可用于广播播发广告的智能监听监管。首先将原始音频划分成音频数据段,然后将嵌入数据分成检测码和密写标签两部分,采用时基调制算法将两部分数据嵌入到音频段中。该方法可以实现密写标签的盲检测,提取密写标签时不需要原始音频。实验结果表明,隐藏有密写标签的载体音频听觉质量没有明显下降。

The embedded and blind testing extraction technology on digital audio hiding written label was studied, which was based on time - scale modification. It can be used in the intelligent monitoring supervision in ads, which was broadcasted by radio. Firstly, the origin digital audio was segmented. Secondly, embedded digital segment was cut into two sections, named detecting code and hiding written label. Finally, the two sections were embedded into audio segment which used the time - scale modification. Blind testing of hiding written labels can be realized by this method, and the origin digital audio wasn＇t needed when the labels were extracted. Experiment results show that the hearing quality of carried audio which was hided in the written label was not decreased.