| 摘要 |
PURPOSE: To provide a system for compressing and decompressing numerical audio information without any loss. CONSTITUTION: A first audio sample is retrieved from a numerical audio signal, and is entirely stored in a non-compressed form. A second audio sample is retrieved, and a signal difference between the second voice sample and the first sample is calculated. When this difference can be expressed in fewer data segments than would be required to represent the second audio sample, this difference is stored. Otherwise, the second audio sample is stored. In decompressing a compressed data stream, the stored sample is retrieved, and examined whether it is the difference value or a flag value indicating that the following data are a coded amplitude value. When this is the difference value, it is added to the preceding value (taking into account the sign), and the sum is inserted into an output data stream. This sum is preserved for calculation in case that the next output value is also the difference value. When this is the flag value, it is discarded, and the following sample is extracted and inserted into the output data stream. |
