Patent
1994-01-28
1997-01-14
MacDonald, Allen R.
395 231, 395 232, 395 238, G10L 302, G10L 900
Patent
active
055948332
DESCRIPTION:
BRIEF SUMMARY
[TECHNICAL FIELD]
The present invention relates to a method for compressing sound data and an apparatus for reproducing the same, and more particularly to a sound data compressing method and a sound data reproducing apparatus with which sound data such as music data or voice data is coded at high efficiency and the coded sound data is reproduced.
[BACKGROUND ART]
Recently, with development of the digital technology, digital recording/storing media such as compact disks (CD's) have been put into practice and used widely. The widespread use of digital recording/storing media are based on development of the A/D conversion technology for converting analog data into digital data. In particular, the technology for A/D converting audio signals such as voice data or music data is prevalent in the fields of domestic audio products and communication data equipment.
Further, because digitizing audio signals increases an amount of data, the compressing/reproducing (coding/decoding) technology for compressing digital data during transmission and reproducing the digital data compressed and transmitted during reception has been developed and practiced in the field of communication data equipment for the purpose of reducing a communication cost. Also, the compressing/reproducing technology for recording and storing a larger amount of digital data has been researched and developed in the fields of communication data equipment and domestic audio products, aiming at not only a higher compression rate but also improved sound quality in auditory sense. As one high-efficient coding technique intended for a higher compression rate and improved sound quality in auditory sense for audio signals in digital recording/storing media, there is proposed a high-efficient coding technique for audio signals using subband coding and auditory characteristics, for example, as described in "Nikkei Electronics", Jul. 22, 1991 and Aug. 5, 1991. With this high-efficient coding technique, an input signal is sampled and divided to form frames each comprising 384 samples, these frames are classified one by one through band-pass filters (band width; 750 Hz) for each of bands, and bit allocation for each band is determined based on a threshold obtained from an auditory masking effect (using an amplitude spectrum with 512-point FFT) and a minimum audible limit characteristic (with simple tone of 1 kHz as reference), thereby compressing audio signals (sampling frequency; 48 kHz, quantization; 16 bits), comparable to those used in CD's, into 128 kbits/sec.
FIG. 11 shows a block diagram based on a basic algorithm of the high-efficient coding technique for audio signals using subband coding and auditory characteristics. In FIG. 11, it is assumed that an input signal is processed in units of frames each corresponding to 8-20 ms (about 384-1000 samples) and a group of 32 band-pass filters are used for the subband coding. A block denoted by 1 in the drawing is a subband coding block. The subband coding block 1 includes a group of band-pass filters (BPF1-BPF32) each comprising an FIR type polyphase filter with 64-512 taps. Output signals of the band-pass filter group are each converted into a base band signal through a thinning-out process in a thinning-out block, coded in a quantizing/coding block, and then output as coded data. In parallel to the above coding process, with in an auditory characteristic block denoted by 2 in the drawing, the input signal is subjected to 512 to 1024-point FFT (Fast Fourier Transform) in an FFT block for determining an amplitude spectrum. Subsequently, a threshold of the amplitude spectrum is obtained in a threshold calculating block within the auditory characteristic block 2 by applying models of an auditory masking effect and a minimum audible limit characteristic. Bit allocation for output signals of the band-pass filters are determined from the threshold. Then, the coded data is combined with auxiliary data (such as bit allocation and a scale factor for quantization) in a multiplexing block within the subband coding blo
Chowdhury Indranil
MacDonald Allen R.
LandOfFree
Rapid sound data compression in code book creation does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Rapid sound data compression in code book creation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Rapid sound data compression in code book creation will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1394953