Tone frame detector for digital speech

A speech decoder configured to detect and extract tone data embedded in a voice bit stream that includes frames of bits, with some of the frames of bits being frames of non-tone bits and some of the frames of bits being frames of tone bits, the speech decoder being operable to:

• select a frame of bits from the voice bit stream;
• analyze the selected frame of bits to determine whether the selected frame of bits is a frame of tone bits;
• and
• when the selected frame of bits is a frame of tone bits, extract tone data from the selected frame of bits;
• wherein the speech decoder analyzes the selected frame of bits by:
• comparing bits of the selected frame of bits to sets of tone data to produce error criteria representative of differences between the selected frame of bits and each of multiple sets of tone data,
• based on the error criteria, selecting a set of tone data that most closely corresponds to the bits of the selected frame of bits, and
• when the error criteria corresponding to the selected set of tone data satisfies a set of thresholds, designating the selected frame of bits as a frame of tone bits, and
• wherein the speech decoder extracts tone data from the selected frame of bits by providing the selected set of tone data as the extracted tone data.

A method for detecting and extracting non-voice data embedded in a voice bit stream that includes frames of bits, with some of the frames of bits being frames of voice bits and some of the frames of bits being frames of non-voice bits, the method comprising:

• selecting a frame of bits from the voice bit stream;
• analyzing the selected frame of bits to determine whether the selected frame of bits is a frame of non-voice bits;
• and
• when the selected frame of bits is a frame of non-voice bits, extracting non-voice data from the selected frame of bits;
• wherein analyzing the selected frame of bits comprises:
• comparing bits of the selected frame of bits to sets of non-voice data to produce error criteria representative of differences between the selected frame of bits and each of multiple sets of non-voice data,
• based on the error criteria, selecting a set of non-voice data that most closely corresponds to the bits of the selected frame of bits, and
• when the error criteria corresponding to the selected set of non-voice data satisfies a set of thresholds, designating the selected frame of bits as a frame of non-voice bits, and
• wherein extracting non-voice data from the selected frame of bits comprises providing the selected set of non-voice data as the extracted non-voice data.