Dual processor digital audio decoder with shared memory data transfer and task partitioning for decompressing compressed audio data, and systems and methods using the same

• US 6,081,783 A
• Filed: 11/14/1997
• Issued: 06/27/2000
• Est. Priority Date: 11/14/1997
• Status: Expired due to Term

• Current Assignee: Crystal Semiconductor Corporation (Cirrus Logic Incorporated)
• Original Assignee: Cirrus Logic Incorporated
• Inventors: Luo, Zheng, Rao, Raghunath, Pacourek, John, Dokic, Miroslav, Scott, Baker III, Divine, James, Niehaus, Jeffrey, Ritchie, Terry
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Storm, Donald L.
• Application Number:
  US08/970,979
• Time in Patent Office: 956 Days
• Field of Search: 704/200, 704/201, 704/221, 704/222, 704/500, 704/501, 395/800.35, 714/747
• US Class Current: 704/500
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

An audio decoder 100 for operating on a received compressed audio data stream compressed using an algorithm employing transform encoding and a bit allocation routine. A first processor 200 performs a first set of operations on the received compressed audio data stream including parsing the compressed audio data stream, recovering data fields within the compressed audio data stream, calculating a bit allocation, and passing frequency domain coefficients to shared memory. A second digital signal processor 100b performs a second set of operations on data passed from the first digital signal processor to shared memory including performing inverse transform operations on the data passed from the first digital signal processor.

Claims

1. An audio decoder comprising a first digital signal processor for performing a first set of operations on a received compressed audio data stream compressed using an algorithm employing transform encoding and a bit allocation routine including:

parsing said compressed audio data stream;

recovering data fields within said compressed audio data stream;

calculating a bit allocation; and

passing frequency domain transform coefficients to shared memory; and

a second digital signal processor for performing a second set of operations on data passed from said first digital processor to shared memory including performing inverse transform operations on said data passed from said first digital signal processor.

2. The audio decoder of claim 1 wherein said first and second digital signal processors process said compressed audio data as channels, said first digital signal processor processing channel n and said second digital signal processor processing in parallel channel n-1.

3. The audio decoder of claim 1 wherein said first and second digital signal processors are operable to exchange command and control messages via a set of registers.

4. The audio decoder of claim 1 wherein said first and second digital signal processors are fabricated on a single chip and further comprising:

first memory associated with said first digital signal processor including data RAM, program RAM, and program ROM; and

second memory associated with said second digital signal processor including data RAM, program RAM, and program ROM.

5. The audio decoder of claim 4, wherein a selected one of said digital signal processors comprises an execution unit coupled to first and second busses;

a program address unit coupled to said first bus for controlling an exchange of data between program memory and said first bus;

a data address unit coupled to said second bus for controlling an exchange of data between data memory and said second bus;

and further comprising a RAM repair unit, said RAM repair unit including a registers file, a plurality of remap registers and address match logic.

6. A dual-digital signal processor (dual-DSP) single chip audio decoder comprising:

a first signal processor operable to parse a received compressed stream of digital audio data and extract frequency domain transform coefficients represented by exponents and mantissas;

a shared memory system, said first processor loading said transform coefficients into said shared memory;

interprocessor communications registers including a register used by said first processor to indicate that said coefficients have been loaded into said memory; and

a second signal processor operable to retrieve said transform coefficients from said shared memory and perform reverse transform operations thereon to recover pulse-code-modulated samples.

7. The dual-DSP audio decoder of claim 6 wherein said first signal processor is further operable to perform a bit allocation for unpacking and dequantizing said mantissas representing said transform coefficients.

8. The dual-DSP audio decoder of claim 6 wherein said compressed stream of digital audio data is compressed using a high coding gain algorithm.

9. The dual-DSP audio decoder of claim 6 wherein said first processor is operable to extract said frequency domain transform coefficients for a first channel substantially simultaneously with said second processor recovering pulse-code-modulated samples for a second channel.

10. The dual-DSP audio decoder of claim 6 wherein said first and second processors operate in parallel in conjunction with dedicated program and data memories.

11. A method of decompressing a stream of audio data compressed using a high coding gain algorithm and representing pulse code modulated (PCM) samples as frequency domain coefficients comprising the steps of:

extracting the frequency domain coefficients from the stream using a first processor comprising the substeps of;

parsing the compressed audio data stream to extract channels of data;

recovering data fields within the compressed audio data stream for each channel; and

calculating a bit allocation to determine a number of bits used to encode each PCM sample into frequency domain coefficients; and

performing an inverse transform on the frequency domain coefficients using a second processor to recover the encoded PCM samples.

12. The method of claim 11 wherein the high coding gain algorithm comprises and AC-3 encoding algorithm.

13. The method of claim 11 wherein the first and second processors comprise digital signal processors.

14. The method of claim 11 wherein said step of extracting is performed on a first channel of data substantially concurrently with said step of performing the inverse transform on a second channel of data.

15. The method of claim 11 and further comprising the step of transferring the frequency domain coefficients from the first processor to the second processor using shared memory and a semaphore.