Digital signal encoding with quantizing based on masking from multiple frequency bands

• US 5,157,760 A
• Filed: 04/16/1991
• Issued: 10/20/1992
• Est. Priority Date: 04/20/1990
• Status: Expired

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Akagiri, Kenzo
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): Knepper, David D.
• Application Number:
  US07/685,745
• Time in Patent Office: 553 Days
• Field of Search: 381/29-50, 381/71, 381/94, 395/2
• US Class Current: 704/233
• CPC Class Codes: G06T 9/007 Transform coding, e.g. discrete cosine transform
• CPC Class Codes: H04B 1/665 using psychoacoustic properties of the ear, e.g. masking effect
• CPC Class Codes: H04B 1/667 using a division in frequency subbands for TV signals H04N19/63

Abstract

An allowable noise level is set for each frequency band of the input digital signal on the basis of the energy level of the respective frequency band and that of another frequency band. The energy level and the allowable noise level are non-linearly related so as to enable high data compression and to provide a satisfactory sound. The signal components of each of the frequency bands are quantized with a number of bits which correspond to the difference between the energy level of each frequency band and the allowable noise level. If the input digital signal is preemphasized, the signal is corrected for preemphasis prior to the setting of the allowable noise level, so that the quantization is performed on a signal which is free from the effects of preemphasis.

Claims

1. An apparatus for encoding a digital signal, comprising:

dividing means for receiving said digital signal and for dividing said digital signal into a plurality of frequency bands to produce a plurality of respective frequency band signals having respective energy levels,

noise level setting means for generating an allowable noise level for each of said frequency band signals based on the energy level of the respective frequency band signal and the energy level of at least another one of said frequency band signals in which each of the allowable noise levels is non-linearly related to the energy level of the respective frequency band signal,

means for producing a difference signal for each of said frequency band signals which represents a difference between the energy level of the respective frequency band signal and the allowable noise level for said respective frequency band signal, and

quantization means for producing a quantized signal by quantizing components of each of said frequency band signals with a respective number of bits corresponding to the respective difference signal for each of said frequency band signals.

2. An apparatus for encoding a digital signal according to claim 1, wherein said dividing means divides said digital signal into said plurality of frequency bands having respective bandwidths in which each of said bandwidths is selected so as to increase in width as the frequency increases.

3. An apparatus for encoding a digital signal according to claim 2, wherein said quantization means includes memory means for receiving said difference signal and for generating said number of bits.

4. An apparatus for encoding a digital signal according to claim 2, further comprising pre-emphasis means for emphasizing predetermined components of said digital signal relative to other components and de-emphasis means for de-emphasizing said quantized signal, and wherein said dividing means receives said digital signal as pre-emphasized by said pre-emphasizing means.

5. An apparatus for encoding a digital signal according to claim 4, wherein said noise level setting means includes correcting means for correcting said plurality of frequency band signals produced by said dividing means so as to cancel any pre-emphasis effects produced by said pre-emphasis means.