摘要
介绍了一种针对低采样率音频信号的压缩方案。该方案通过混合滤波器组分解音频信号,并根据心理声学原理给不同的子带分配不同的比特数和量化阶距,最后还利用了哈夫曼编码技术对量化后数据进行编码。实验证明,这种基于Mp3框架的低采样率音频压缩方案能够对低至8kHz采样率的音频信号达到20kbps的压缩效果。
A scheme of low-frequency sampled audio compression is proposed here. Audio signal is divided into 8 subbands by hybrid filterbanks, and different numbers of bits and quantizers are allocated to different subbands based on result of psychoacoustic analysis. At last, samples after being quantized are coded with Huffman codes. By using Mp3 compress frame, this scheme can provide perceptually lossless compression of 20kbps with 8kHz sampled audio signal.
出处
《计算机工程》
CAS
CSCD
北大核心
2003年第19期146-148,共3页
Computer Engineering
基金
西北工业大学博士论文创新基金项目
关键词
低采样率
MP3
子带编码
音频压缩
Low-frequence sample
Mp3(MPEG Lsyer-3)
Subband coding
Audio compression
