Compression of stored waveforms for artificial speech

• US 4,833,718 A
• Filed: 02/12/1987
• Issued: 05/23/1989
• Est. Priority Date: 11/18/1986
• Status: Expired due to Fees

• Current Assignee: Sierra Entertainment, Inc. (Vivendi SE)
• Original Assignee: First Byte
• Inventors: Sprague, Richard P.
• Primary Examiner(s): Kemeny, Emanuel S.
• Application Number:
  US07/013,966
• Time in Patent Office: 831 Days
• Field of Search: 381/29-53, 379/88, 379/196, 379/211, 382/56, 358/263, 358/136, 358/283
• US Class Current: 704/229
• CPC Class Codes: G10L 13/07 Concatenation rules
• 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
• CPC Class Codes: H03M 7/42 using table look-up for the coding or decoding process, e.g. using read-only memory H03M7/4006 takes precedence
• CPC Class Codes: H03M 7/50 Conversion to or from non-linear codes, e.g. companding

Abstract

In a digital text-to-speech conversion system of the type usually contained in all-software form on a floppy disk, memory requirements for the storage of digitized waveform samples are reduced while speech quality is improved, by providing compression techniques and anti-distortion techniques which interact to provide clear speech at widely varying speeds with a minimum of memory. These techniques include using Huffman coding of first- or second-order differences, encoding only differences between successive waveforms where feasible, using a demi-diphone organization of the speech to allow use of the same instruction lists for several sounds, selectively deleting or repeating waveforms in the concatenation to vary speed without affecting pitch, and encoding waveforms linearly or anti-logarithmically for storage while converting the stored linear or anti-logarithmic codes to logarithmic codes such as μ-law codes upon retrieval.

Claims

1. In a real-time text-to-speech conversion system in which waveforms are encoded for storage in terms of digital samples, the frequency of occurrence of the various possible sample levels being non-uniform, the improvement comprising encoding corresponding samples of a pair of successive waveforms for storage by a Huffman coding in which the sample values themselves are encoded for the first waveform of said pair, and the value of the difference between each sample and the corresponding sample of said first waveform is encoded for the second waveform of said pair.

2. The improvement of claim 1, in which the shortest codes of said Huffman coding represent the sample or difference values occurring most frequently in said pair of waveforms.

3. The improvement of claim 2, in which, when said first and second waveforms have different numbers of samples, the shorter waveform is treated in the computation of said difference values as if it were padded with sufficient zero value samples to equal the number of samples in the longer waveform.