Method for communicating a value over a transmission medium and for decoding same

• US 5,617,506 A
• Filed: 06/29/1994
• Issued: 04/01/1997
• Est. Priority Date: 06/29/1994
• Status: Expired due to Term

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: The 3DO Company
• Inventors: Hayes, Steven E., Burk, Phil L., Platt, David C., Mical, Robert J.
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): CHOWDHURY, INDRINAL
• Application Number:
  US08/267,978
• Time in Patent Office: 1,007 Days
• Field of Search: 395/2.1, 395/2.13, 395/2.2, 281/29, 281/30, 281/33, 281/34, 281/35
• US Class Current: 704/201
• CPC Class Codes: G11B 20/00007 Time or data compression or expansion audio compression based on psychoacoustics G10L19/00; data processing for reproducing audio data at different playback speeds G10L21/04; video compression H04N19/00; data compression per se H03M7/30
• CPC Class Codes: H04B 14/048 Non linear compression or expansion

Abstract

A method and apparatus for a 2:1 compression of a digital audio signal including 1) a compression operation for compressing a sixteen bit sample of the audio signal as both an exact value and a delta value, and 2) a regeneration operation for regenerating the compressed exact and delta values once again into a sixteen bit word. The regenerated exact and delta values are then compared against the original sample. Whichever regenerated value most closely matches the original sample has its corresponding compressed value encoded onto a storage medium.

Claims

1. A method for communicating a plurality of starting values s_i over a transmission medium, i=0, 1, 2, . . . , all of said starting values s_i being in a starting value space, comprising the steps of, for each of at least two i'"'"'th starting values s_i :

calculating a transmission value y_i given by y_i =f_i (s_i), where the value f_i (s_i) is selected depending on a selection decision from a predefined group of candidate values including e₁ (s_i) and d₁ (s_i -f^-1_i-1 (f_i-1 (s_i-1))), where the function e₁ ( ) is a predefined mapping from said starting value space to a first transmission value space, and where the function d₁ ( ) is a predefined mapping from said starting value space to a second transmission value space said first and second transmission value spaces containing discrete values; and

transmitting over said transmission medium said transmission value y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not of f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1));

wherein the function e₁ ( ) is defined by e₁ (s)=sign(s)*|s|^1/n, rounded to the nearest discrete value in said first transmission value space, and the function d₁ ( ) is defined by d₁ (s)=sign(s)*|s|^1/n, rounded to the nearest discrete value in said second transmission value space, n being an integer greater than 1.

2. A method for communicating a plurality of starting values s_i over a transmission medium, i=0, 1, 2, . . . , all of said starting values s_i being in a starting value space, comprising the steps of, for each of at least two i'"'"'th starting values s_i :

calculating a transmission value y_i given by y_i =f_i (s_i), where the value f_i (s_i) is selected depending on a selection decision from a predefined group of candidate values including e₁ (s_i) and d₁ (s_i -f^-1_i-1 (f_i-1 (s_i-1))), where the function e₁ ( ) is a predefined mapping from said starting value space to at first transmission value space, and where the function d₁ ( ) is a predefined mapping from said starting value space to a second transmission value space said first and second transmission value spaces containing discrete values; and

transmitting over said transmission medium said transmission value y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not of f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1));

wherein the function e₁ ( ) is defined by e₁ (s)=sign(s)*|s|^1/2, rounded to the nearest discrete value in said first transmission value space, and the function d₁ ( ) is defined by d₁ (s)=sign(s)*|s|^1/2, rounded to the nearest discrete value in said second transmission value space.

3. A method for communicating a plurality of starting values s_i over a transmission medium, i=0, 1, 2, . . . , all of said starting values s_i being in a starting value space, comprising the steps of, for each of at least two i'"'"'th starting values s_i :

calculating a transmission value y_i given by y_i =f_i (s_i), where the value f_i (s_i) is selected depending on a selection decision from a predefined group of candidate values including e₁ (s_i) and d₁ (s_i -f^-1_i-1 (f_i-1 (s_i-1))), where the function e₁ ( ) is a predefined mapping from said starting value space to a first transmission value space, and where the function d₁ ( ) is a predefined mapping from said starting value space to a second transmission value space said first and second transmission value spaces containing discrete values; and

transmitting over said transmission medium said transmission value y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not of f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1));

wherein said selection decision selects the value f_i (s_i) as whichever of the candidate values in said group of candidate values minimizes |s_i -f_i^-1 (f_i (s_i))|.

4. A method for communicating a plurality of starting values s_i over a transmission medium, i=0, 1, 2, . . . , all of said starting values s_i being in a starting value space, comprising the steps of, for each of at least two i'"'"'th starting values s_i :

calculating a transmission value y_i given by y_i =f_i (s_i), where the value f_i (s_i) is selected depending on a selection decision from a predefined group of candidate values including e₁ (s_i) and d₁ (s_i -f^-1_i-1 (f_i-1 (s_i-1))), where the function e₁ ( ) is a predefined mapping from said starting value space to a first transmission value space, and where the function d₁ ( ) is a predefined mapping from said starting value space to a second transmission value space said first and second transmission value spaces containing discrete values; and

transmitting over said transmission medium said transmission value y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not of f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1));

wherein said first transmission value space is the space of all 8-bit values having a low order bit of 0, wherein said second transmission value space is the space of all 8-bit values having a low order bit of 1, and wherein said indication of whether y_i ;

(a) is a function of s_i and not f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1)), comprises the low order bit of y_i.

5. A method for communicating a plurality of starting values s_i over a transmission medium, i=0, 1, 2, . . . , all of said starting values s_i being in a starting value space, comprising the steps of, for each of at least two i'"'"'th starting values s_i :

calculating a transmission value y_i given by y_i =f_i (s_i), where the value f_i (s_i) is selected depending on a selection decision from a predefined group of candidate values including e₁ (s_i) and d₁ (s_i -f^-1_i-1 (f_i-1 (s_i-1))), where the function e₁ ( ) is a predefined mapping from said starting value space to a first transmission value space, and where the function d₁ ( ) is a predefined mapping from said starting value space to a second transmission value space said first and second transmission value spaces containing discrete values; and

transmitting over said transmission medium said transmission value y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not of f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1));

wherein said group of candidate values further includes values e₂ (s_i) and d₂ (s_i -f^-1_i-1 (f_i-1 (s_i-1))), where, for a predefined function g( );

the value e₁ (s)=g(s) rounded to the next lower discrete value in said first transmission value space;

the value e₂ (s)=g(s) rounded to the next higher discrete value in said first transmission value space;

the value d₁ (s)]=g(s) rounded to the next lower discrete value in said second transmission value space; and

the value d₂ (s)]=g(s) rounded to the next higher discrete value in said second transmission value space.

6. A method according to claim 5, wherein g(s)=sign(s)*|s|^1/n, n being an integer greater than 1.

7. A method according to claim 5, wherein g(s)=sign(s)*|s|^1/2.

8. A method according to claim 5, wherein said selection decision selects the value f_i (s) as whichever of said candidate values minimizes |s_i -f_i^-1 (f_i (s_i))|.

9. A method according to claim 5, wherein said selection decision selects the value f_i (s) by determining the one of said candidate values f_i (s) and the one of said candidate values f_i+1 (s) which together minimize [s_i -f_i^-1 (f_i (s_i))]² +[s_i+1 -f_i+1^-1 (f_i+1 (s_i+1))]².

10. A method for communicating a plurality of starting values s_i over a transmission medium, i=0, 1, 2, . . . , comprising the steps of:

calculating a value y_i =sign(s_i)*|s_i |^1/n for each i, n being an integer greater than 1; and

transmitting over said transmission medium values responsive to said y_i.

11. A method according to claim 10, for use with a plurality of input values x_i each corresponding to a respective one of said starting values s_i, wherein said step of transmitting comprises the step of transmitting over said transmission medium a respective value responsive to each of said y_i, further comprising the step of performing one of the following steps selectively for each s_i :

(a) setting s_i equal to the corresponding x_i ; and

(b) setting s_i equal to the difference between x_i and an inverse function of the value transmitted over said transmission medium in response to y_i-1,

selectably in dependence on a selection decision.

12. A method according to claim 10, further comprising a step of mapping each of said y_i to an element in a transmission value space selected from a plurality of predefined transmission value spaces, wherein said step of transmitting comprises a step of transmitting over said transmission medium both the mapped y_i and an indication of the transmission value space selected for y_i.

13. A method for decoding a plurality of transmission values y_i received over a transmission medium, i=0, 1, 2, . . . , comprising the steps of:

receiving each of said transmission values y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1)), where each of the s_i are starting values to be reconstructed; and

calculating a reconstructed value s'"'"'_i for each y_i as s'"'"'_i =g^-1 (y_i) if y_i is a function of s_i and not f^-1_i-1 (f_i-1 (s_i-1)), and calculating a reconstructed value s'"'"'_i for each y_i as s'"'"'_i =s'"'"'_i-1 +g^-1 (y_i) if y_i is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1)), wherein the function g^-1 ( )=sign(y)*|y|ⁿ, n being a predefined integer greater than 2.

14. A method for decoding a plurality of transmission values y_i received over a transmission medium i=0, 1, 2, . . . , comprising the steps of, for i>

0;

receiving each of said transmission values y_i in conjunction with an indication of whether y_i ;

(a) is a function of s_i and not f^-1_i-1 (f_i-1 (s_i-1)), or (b) is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1)), where each of the s_i are starting values to be reconstructed; and

calculating a reconstructed value s'"'"'_i for each y_i as s'"'"'_i =g^-1 (y_i ) if y_i is a function of s_i and not f^-1_i-1 (f_i-1 (s_i-1)), and calculating a reconstructed value s'"'"'_i for each y_i as s'"'"'_i =s'"'"'_i-1 +g^-1 (y_i) if y_i is a function of s_i -f^-1_i-1 (f_i-1 (s_i-1)), wherein the function g^-1 ( )=sign(y)*y².

15. A method for decoding a plurality of transmission values y_i received over a transmission medium, i=0, 1, 2, . . . , comprising the steps of, for i>

0;

receiving each of said transmission values y_i ; and

calculating a value z_i for each of said y_i, where for each i, z_i =sign(y_i)*|y_i |ⁿ, n being a predetermined integer greater than 1.

16. A method according to claim 15, wherein n=2.

17. A method according to claim 15, wherein n=3.

18. A method according to claim 15, further comprising the steps of:

receiving in conjunction with each of said y_i an indication of whether y_i is a function of s_i and not of the difference between s_i and a reconstruction of an encoded version of s_i-1, or is a function of the difference between s_i and a reconstruction of an encoded version of s_i-1, each of the s_i being starting values to be reconstructed; and

determining a reconstructed value s'"'"'_i for each y_i as s'"'"'_i =z_i if y_i is a function of s_i and not of the difference between s_i and a reconstruction of an encoded version of s_i-1, and as s'"'"'_i =s'"'"'_i-1 +z_i if y_i is a function of the difference between s_i and a reconstruction of an encoded version of s_i-1.