Method of efficiently recording an audio signal in semiconductor memory

• US 5,752,221 A
• Filed: 09/18/1996
• Issued: 05/12/1998
• Est. Priority Date: 04/20/1992
• Status: Expired due to Term

• Current Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Original Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Inventors: Sugiyama, Kazuhiro, Ishida, Yoshinobu, Ono, Yukari
• Primary Examiner(s): HAFIZ, TARIQ R
• Application Number:
  US08/716,566
• Time in Patent Office: 601 Days
• Field of Search: 395/2.1, 395/2.2, 395/2.15-2.17, 395/2.12, 395/2.67, 395/2.92, 375/240
• US Class Current: 704/201
• CPC Class Codes: G11C 2207/16 Solid state audio
• CPC Class Codes: G11C 7/16 Storage of analogue signals in digital stores using an arrangement comprising analogue/digital [A/D] converters, digital memories and digital/analogue [D/A] converters

Abstract

An audio signal is recorded in a semiconductor memory in a plurality of hierarchical levels, with the lowest level sufficing for reproduction with a certain reduced degree of fidelity. Successively higher hierarchical levels provide successively greater fidelity when reproduced. When the memory has been determined to have reached maximum capacity recording continues by overwriting the highest hierarchical level of data currently stored in the memory with lower hierarchical levels of new data. A code is recorded in the memory, indicating the number of hierarchical levels recorded therein, for subsequent reproduction. The audio signal can furthermore be recorded in variable-length frames and reproduced at high speed by reading every N-th frame, N being a positive integer, or by reading only frames having at least a certain minimum length.

Claims

1. A hierarchical encoder, comprising:

frequency transformation means for transforming an input digital audio signal into transformation coefficients for a plurality of frequency bands; and

hierarchical quantizing means for extracting audible signal components from the transformation coefficients based on characteristics for a psycho-acoustic curve, and for generating, from the extracted audible signal components, first encoded data to n-th encoded data with priorities based on human auditory characteristics, n being an integer.

2. The hierarchical encoder of claim 1, further comprising a dividing means for dividing the input digital audio signal into blocks of a predetermined size, wherein said frequency transformation means separately transforms each of said blocks.

3. A hierarchical encoder, comprising:

frequency transformation means for transforming an input digital audio signal into transformation coefficients for a plurality of frequency bands; and

hierarchical dividing and quantizing means for extracting audible signal components from the transformation coefficients based on psycho-acoustic characteristics, for dividing the audible signal components into hierarchical levels, and for quantizing the audible signal components to produce encoded data divided into a plurality of hierarchical levels,

a first hierarchical level of encoded data S1 being obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a first predetermined frequency f₁, with a quantization level from the most significant bit equal to or less than a first predetermined number of bits b₁,

a second hierarchical level of encoded data S2 being obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a second predetermined frequency f₂ (f₂ ≧

f₁), with a quantization level from the most significant bit equal to or less than a second predetermined number of bits b₂ (b₂ ≧

b₁), and subtracting out the transformation coefficients of the first hierarchical level, and

an n-th hierarchical level of encoded data Sn being obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than an nth predetermined frequency f_n (f_n ≧

f_n-1), with a quantization level from the most significant bit equal to or less than an nth predetermined number of bits b_n (b_n ≧

b_n-1), and subtracting out the transformation coefficients of each of the first hierarchical level to the (n-1)-th hierarchical level, wherein n is an integer.

4. The hierarchical encoder of claim 3, wherein the division into hierarchical levels is carried out in relation to a psycho-acoustic-characteristic curve.

5. The hierarchical encoder of claim 4, wherein the psycho-acoustic-characteristic curve includes the human threshold of audibility as a characteristic thereof.

6. The hierarchical encoder of claim 3, further comprising a dividing means for dividing the input digital audio signal into blocks of a predetermined size, wherein said frequency transformation means separately transforms each of said blocks.

7. A hierarchical decoder, comprising:

a decoding means, receiving input audio data hierarchically encoded with hierarchical priorities based on characteristics from a psycho-acoustic curve and receiving an identification code identifying a hierarchical priority level of the encoded audio data, for decoding said hierarchically encoded audio data responsive to a hierarchical level thereof, determined based on said received identification code, to produce transformation coefficients corresponding to frequency bands; and

an inverse frequency transform means for performing an inverse frequency transformation on the produced transformation coefficients to thereby produce an original digital audio signal.

8. A hierarchical decoder, comprising:

a decoding means, receiving input audio data hierarchically encoded with hierarchial priorities based on human auditory characteristics and receiving an identification code identifying a hierarchical priority level of the encoded audio data, for decoding said hierarchically encoded audio data responsive to a hierarchical level thereof, determined based on said received identification code, to produce transformation coefficients corresponding to frequency bands; and

an inverse frequency transform means for performing an inverse frequency transformation on the produced transformation coefficients to thereby produce an original digital audio signal,

wherein said hierarchically encoded audio data is data which has been previously divided into hierarchical levels and quantized such that,

a first hierarchical level of encoded data S1 is obtained by selecting, from among the transformation coefficients of audible signal components extracted according to human psycho-acoustic characteristics, transformation coefficients of frequency bands equal to or less than a first predetermined frequency f₁, including a quantization level from the most significant bit equal to or less than a first predetermined number of bits b₁,

a second hierarchical level of encoded data S2 is obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a second predetermined frequency f₂ (f₂ ≧

f₁) including a quantization level from the most significant bit equal to or less than a second predetermined number of bits b₂ (b₂ ≧

b₁), and subtracting out the transformation coefficients of the first hierarchical level, and

an n-th hierarchical level of encoded data Sn is obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than an n-th predetermined frequency f_n (f_n ≧

f_n-1), including a quantization level from the most significant bit equal to or less than an nth predetermined number of bits b_n (b_n ≧

b_n-1), and subtracting out the transformation coefficients of each of the first hierarchical level to the (n-1)-th hierarchical level, wherein n is an integer.

9. The hierarchical decoder of claim 8, wherein the division into hierarchical levels is carried out in relation to a psycho-acoustic-characteristic curve.

10. The hierarchical decoder of claim 9, wherein the psycho-acoustic characteristic curve includes the human threshold of audibility as a characteristic thereof.

11. The hierarchical decoder of claim 9, wherein the psycho-acoustic characteristic curve includes the human threshold of audibility as a characteristic thereof.

12. The hierarchical decoder of claim 8, wherein the division into hierarchical levels is carried out in relation to a psycho-acoustic-characteristic curve.

13. A hierarchical encoder, comprising:

frequency transformation means for transforming an input digital audio signal into transformation coefficients for a plurality of frequency bands; and

hierarchical dividing and quantizing means for extracting audible signal components from the transformation coefficients based on psycho-acoustic characteristics, for dividing the audible signal components into hierarchical levels, and for quantizing the audible signal components to produce encoded data divided into a plurality of hierarchical levels,

a first hierarchical level of encoded data S1 being obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a first predetermined frequency f₁, with a quantization level from the most significant bit equal to or less than a first predetermined number of bits b₁, and

an n-th hierarchical level of encoded data Sn being obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than an nth predetermined frequency f_n (f_n ≧

f_n-1), with a quantization level from the most significant bit equal to or less than an nth predetermined number of bits b_n (b_n ≧

b_n-1), and subtracting out the transformation coefficients of each of the first hierarchical level to the (n-1)-th hierarchical level, wherein n is an integer.

14. The hierarchical encoder of claim 13, wherein the division into hierarchical levels is carried out in relation to a psycho-acoustic-characteristic curve.

15. The hierarchical encoder of claim 14, wherein the psycho-acoustic-characteristic curve includes the human threshold of audibility as a characteristic thereof.

16. The hierarchical encoder of claim 13, further comprising a dividing means for dividing the input digital audio signal into blocks of a predetermined size, wherein said frequency transformation means separately transforms each of said blocks.

17. A hierarchical decoder, comprising:

a decoding means, receiving input audio data hierarchically encoded with hierarchial priorities based on human auditory characteristics and receiving an identification code identifying a hierarchical priority level of the encoded audio data, for decoding said hierarchically encoded audio data responsive to a hierarchical level thereof, determined based on said received identification code, to produce transformation coefficients corresponding to frequency bands; and

an inverse frequency transform means for performing an inverse frequency transformation on the produced transformation coefficients to thereby produce an original digital audio signal,

wherein said hierarchically encoded audio data is data which has been previously divided into hierarchical levels and quantized such that,

a first hierarchical level of encoded data S1 is obtained by selecting, from among the transformation coefficients of audible signal components extracted according to human psycho-acoustic characteristics, transformation coefficients of frequency bands equal to or less than a first predetermined frequency f₁, including a quantization level from the most significant bit equal to or less than a first predetermined number of bits b₁, and

an n-th hierarchical level of encoded data Sn is obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than an n-th predetermined frequency f_n (f_n ≧

f_n-1), including a quantization level from the most significant bit equal to or less than an nth predetermined number of bits b_n (b_n ≧

b_n-1) and subtracting out the transformation coefficients of each of the first hierarchical level to the (n-1)-th hierarchical level, wherein n is an integer.

18. The hierarchical decoder of claim 17, wherein the division into hierarchical levels is carried out in relation to a psycho-acoustic-characteristic curve.

19. The hierarchical decoder of claim 18, wherein the psycho-acoustic characteristic curve includes the human threshold of audibility as a characteristic thereof.

20. A hierarchical encoder, comprising:

frequency transformation means for transforming an input digital audio signal into transformation coefficients for a plurality of frequency bands; and

hierarchical dividing means for extracting audible signal components from the transformation coefficients based on psycho-acoustic characteristics and for dividing the audible signal components into hierarchical levels,

a first hierarchical level of encoded data S1 being obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a first predetermined frequency f₁,

a second hierarchical level of encoded data S2 being obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a second predetermined frequency f₂ (f₂ ≧

f₁), and subtracting out the transformation coefficients of the first hierarchical level, and

an n-th hierarchical level of encoded data Sn being obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than an nth predetermined frequency f_n (f_n ≧

f_n-1), and subtracting out the transformation coefficients of each of the first hierarchical level to the (n-1)-th hierarchical level, wherein n is an integer.

21. The hierarchical encoder of claim 20, wherein the division into hierarchical levels is carried out in relation to a psycho-acoustic-characteristic curve.

22. The hierarchical encoder of claim 21, wherein the psycho-acoustic-characteristic curve includes the human threshold of audibility as a characteristic thereof.

23. The hierarchical encoder of claim 20, further comprising a dividing means for dividing the input digital audio signal into blocks of a predetermined size, wherein said frequency transformation means separately transforms each of said blocks.

24. A hierarchical decoder, comprising:

a decoding means, receiving input audio data hierarchically encoded with hierarchial priorities based on human auditory characteristics and receiving an identification code identifying a hierarchical priority level of the encoded audio data, for decoding said hierarchically encoded audio data responsive to a hierarchical level thereof, determined based on said received identification code, to produce transformation coefficients corresponding to frequency bands; and

an inverse frequency transform means for performing an inverse frequency transformation on the produced transformation coefficients to thereby produce an original digital audio signal,

wherein said hierarchically encoded audio data is data which has been previously divided into hierarchical levels such that,

a first hierarchical level of encoded data S1 is obtained by selecting, from among the transformation coefficients of audible signal components extracted according to human psycho-acoustic characteristics, transformation coefficients of frequency bands equal to or less than a first predetermined frequency f₁,

a second hierarchical level of encoded data S2 is obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than a second predetermined frequency f₂ (f₂ ≧

f₁), and subtracting out the transformation coefficients of the first hierarchical level, and

an n-th hierarchical level of encoded data Sn is obtained as a residual signal obtained by selecting, from among the transformation coefficients of said extracted audible signal components, transformation coefficients of frequency bands equal to or less than an n-th predetermined frequency f_n (f_n ≧

f_n-1), and subtracting out the transformation coefficients of each of the first hierarchical level to the (n-1)-th hierarchical level, wherein n is an integer.