Claims
- 1. A predictive audio coder for encoding digital audio samples representative of an audio signal that has been sampled at an input sampling rate, the predictive audio coder comprising:
- a predictor filter to adaptively compute predictive parameters based upon the digital audio samples;
- an encoder to encode the digital audio samples into a compressed format for transmission, the encoder comprising:
- (a) multiple quantizers for encoding the digital audio samples into various numbers of bits to be contained in audio data blocks so that one quantizer produces a first number of bits for an audio data block of a first size and another quantizer produces a second number of bits for an audio data block of a second size; and
- (b) a quantizer selector to select one of the multiple quantizers.
- 2. A predictive audio coder for encoding digital audio samples representative of an audio signal that has been sampled at an input sampling rate, the predictive audio coder comprising:
- a predictor filter to adaptively compute predictive parameters based upon the digital audio samples, the predictor filter comprising both long and short term predictive filters;
- an encoder to encode the digital audio samples into a compressed format for transmission, the encoder comprising:
- (a) multiple quantizers for encoding the digital audio samples into various numbers of bits to be contained in audio data blocks so that one quantizer produces a first number of bits for an audio data block of a first size and another quantizer produces a second number of bits for an audio data block of a second size; and
- (b) a quantizer selector to select one of the multiple quantizers.
- 3. A predictive audio coder for encoding digital audio samples representative of an audio signal that has been sampled at an input sampling rate and further configured to conform to the Group Speciale Mobile (GSM) standard, the predictive audio coder comprising:
- a predictor filter to adaptively compute predictive parameters based upon the digital audio samples;
- an encoder to encode the digital audio samples into a compressed format for transmission, the encoder comprising:
- (a) nine quantizers for encoding the digital audio samples into various numbers of bits to be contained in audio data blocks so that one quantizer produces a first number of bits for an audio data block of a first size and another quantizer produces a second number of bits for an audio data block of a second size; and
- (b) a quantizer selector to select one of the nine quantizers.
- 4. A predictive audio coder as recited in claim 3 wherein:
- a first quantizer produces 60 bits for audio data blocks consisting of 164 bits;
- a second quantizer produces 72 bits for audio data blocks consisting of 176 bits;
- a third quantizer produces 84 bits for audio data blocks consisting of 188 bits;
- a fourth quantizer produces 96 bits for audio data blocks consisting of 200 bits;
- a fifth quantizer produces 108 bits for audio data blocks consisting of 212 bits;
- a sixth quantizer produces 120 bits for audio data blocks consisting of 224 bits;
- a seventh quantizer produces 132 bits for audio data blocks consisting of 236 bits;
- an eighth quantizer produces 144 bits for audio data blocks consisting of 248 bits; and
- a ninth quantizer produces 156 bits for audio data blocks consisting of 260 bits.
- 5. In an audio coder/decoder implemented using regular pulse excitation (RPE), an RPE encoder comprising:
- multiple quantizers for encoding digital audio samples sampled at a selected input sampling rate into a compressed format for transmission, the quantizers encoding the audio digital data into audio data blocks containing a defined number of samples with varying numbers of sample bits so that one quantizer produces an audio data block having a first number of sample bits and another quantizer produces an audio data block having a second number of sample bits different than the first number of sample bits; and
- a quantizer selector to select one of the multiple quantizers.
- 6. In an audio coder/decoder implemented using regular pulse excitation (RPE) and conforming to the Group Speciale Mobile (GSM) standard, an RPE encoder comprising:
- nine quantizers for encoding digital audio samples sampled at a selected input sampling rate into a compressed format for transmission, the quantizers encoding the audio digital data into audio data blocks containing a defined number of samples with varying numbers of sample bits so that one quantizer produces an audio data block having a first number of sample bits and another quantizer produces an audio data block having a second number of sample bits different than the first number of sample bits; and
- a quantizer selector to select one of the nine quantizers.
- 7. An RPE encoder as recited in claim 6 wherein:
- a first quantizer produces audio data blocks consisting of 164 bits including 60 sample bits;
- a second quantizer produces audio data blocks consisting of 176 bits including 72 sample bits;
- a third quantizer produces audio data blocks consisting of 188 bits including 84 sample bits;
- a fourth quantizer produces audio data blocks consisting of 200 bits including 96 sample bits;
- a fifth quantizer produces audio data blocks consisting of 212 bits including 108 sample bits;
- a sixth quantizer produces audio data blocks consisting of 224 bits including 120 sample bits;
- a seventh quantizer produces audio data blocks consisting of 236 bits including 132 sample bits;
- an eighth quantizer produces audio data blocks consisting of 248 bits including 144 sample bits; and
- a ninth quantizer produces audio data blocks consisting of 260 bits including 156 sample bits.
- 8. An RPE-LTP coder/decoder implemented using regular pulse excitation (RPE) incorporating an RPE encoder comprising:
- multiple quantizers for encoding digital audio samples sampled at a selected input sampling rate into a compressed format for transmission, the quantizers encoding the audio digital data into audio data blocks containing a defined number of samples with varying numbers of sample bits so that one quantizer produces an audio data block having a first number of sample bits and another quantizer produces an audio data block having a second number of sample bits different than the first number of sample bits; and
- a quantizer selector to select one of the multiple quantizers.
- 9. In a system for supplying digital audio files to a recipient having a modem to receive digital data at an effective bit rate, the audio files being stored in multiple different versions having audio data blocks of different numbers of data bits taken at different input sampling rates, a program embodied on a compute-readable medium, comprising:
- computer-executable instructions to select a version of the audio file that has an appropriate number of data bits and an input sampling rate to produce a bit stream bit rate that is less than or equal to the effective bit rate of the recipient's modem; and
- computer-executable instructions to supply the selected version of the audio file to the recipient.
- 10. In a system for supplying digital audio files to a recipient having a modem to receive digital data at an effective bit rate, the audio files being stored in an uncompressed format, a program embodied on a compute-readable medium comprising computer-executable instructions to encode an audio file into individual audio data blocks in real-time wherein each audio data block contains a selected number of bits of digital audio data sampled at a selected input sampling rate, the selected number of bits and the selected input sampling rate determining a bit stream bit rate that is less than or equal to the effective bit rate of the recipient's modem.
- 11. In a system for supplying digital audio files to a recipient having a modem to receive digital data at an effective bit rate, the audio files being configured into individual audio data blocks wherein each audio data block has a block size, which represents a number bits of digital audio data, and an input sampling rate, which represents the sample rate at which the digital audio data is sampled, a program embodied on a compute-readable medium, comprising:
- computer-executable instructions to index a size/rate lookup table listing various combinations of block sizes and sampling rates and select a combination that yields an encoded bit stream bit rate that is less than or equal to the effective bit rate of the recipient's modem; and
- computer-executable instructions to supply the audio file at the encoded bit stream bit rate.
RELATED APPLICATIONS
This is a divisional of U.S. patent application Ser. No. 08/540,818, filed Oct. 11, 1995.
US Referenced Citations (9)
Divisions (1)
|
Number |
Date |
Country |
Parent |
540818 |
Oct 1995 |
|