Automatic realtime speech impairment correction

• US 8,682,678 B2
• Filed: 03/14/2012
• Issued: 03/25/2014
• Est. Priority Date: 03/14/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Malkin, Peter K., Trewin, Sharon M.
• Primary Examiner(s): He, Jialong
• Application Number:
  US13/420,088
• Publication Number:
  US 20130246061A1
• Time in Patent Office: 741 Days
• Field of Search: 704/270, 704/271, 704/278, 600/23
• US Class Current: 704/278
• CPC Class Codes: G10L 2021/0575 Aids for the handicapped in speaking
• CPC Class Codes: G10L 21/00 Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility G10L19/00 takes precedence
• CPC Class Codes: G10L 21/057 for improving intelligibility
• CPC Class Codes: G10L 25/48 specially adapted for particular use

Abstract

Automatic correcting of user'"'"'s speech impairment in speech may include obtaining the audio signal of a given user'"'"'s speech, and analyzing the obtained audio signal to identify artifacts caused by the user'"'"'s impairment. The obtained audio signal may be modified by eliminating the identified artifacts from it. The modified audio signal may be provided, e.g., to be played or broadcast or transmitted.

Claims

1. A method for correcting effects of a user'"'"'s speech impairment, comprising:

obtaining audio signal of a speech;

analyzing, by a processor, the audio signal to identify audio signal artifacts caused by the user'"'"' s speech impairment;

modifying, by the processor, the audio signal by eliminating the identified audio signal artifacts from the audio signal based on a database of impairment-to-artifact associations; and

providing the modified audio signal;

wherein the database of impairment-to-artifact associations including repeated instances of phonemes associated with stuttering, inappropriate verbiage inserted into speech associated with Tourete Syndrome, and slurred pronunciation associated with lisping,

wherein the analyzing further comprises identifying the audio signal artifacts caused by the user'"'"'s speech impairment by employing identified differences between audio signals of the user reading a text with a predefined audio signal associated with the text.

2. The method of claim 1, wherein the method is performed in real time or near real time as the speech is being made.

4. The method of claim 1, wherein the method is performed in real time while the speech is being made, and the obtaining and the providing occur continuously, with a next speech audio being obtained while previously modified audio signal is provided.

5. The method of claim 1, wherein the speech impairment includes stuttering, Tourete Syndrome, or lisping, or combinations thereof.

6. The method of claim 1, further including:

receiving training of which audio signals are artifacts caused by the user'"'"'s speech impairment.

7. The method of claim 6, wherein the receiving training includes receiving user'"'"'s indication of which audio signals are artifacts associated with the user'"'"'s speech impairment.

8. The method of claim 6, wherein the receiving training includes:

receiving audio signals of a user reading a text;

comparing the audio signals of the user reading the text with a predefined audio signal associated with the text; and

identifying the audio signal artifacts associated with the user'"'"'s speech impairment by identifying differences between the audio signals of the user reading the text with a predefined audio signal associated with the text.

3. The method of 1, wherein the providing includes transmitting the modified signal or playing the modified audio signal or combinations thereof.