USING SEPARATE RECORDING CHANNELS FOR SPEECH-TO-SPEECH TRANSLATION SYSTEMS

• US 20090306957A1
• Filed: 10/02/2007
• Published: 12/10/2009
• Est. Priority Date: 10/02/2007
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Gu, Liang, Zhang, Wei, GAO, YUQING
• Granted Patent: US 7,953,590 B2
• Time in Patent Office: Days
• US Class Current: 704/2
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation

Abstract

A system and method for speech-to-speech translation using a translation system includes designating separate input channels for each of a plurality of speakers. In response to speech from a first channel in a first language, the speech from the first channel is translated to a second language, and in response to speech from the second channel in a second language, the speech from the second channel is translated to the first language. Translating the speech from the first channel and translating the speech from the second channel are performed concurrently.

Claims

1. A method for speech-to-speech translation using a translation system, comprising:

designating separate input channels for each of a plurality of speakers;

in response to speech from a first channel in a first language, translating the speech from the first channel to a second language; and

in response to speech from the second channel in a second language, translating the speech from the second channel to the first language;

wherein the steps of translating the speech from the first channel and translating the speech from the second channel are performed concurrently.

2. The method as recited in claim 1, further comprising in response to speech from the first channel in the first language, outputting the speech from the first channel in the second language.

3. The method as recited in claim 2, further comprising in response to speech from the second channel in the second language, outputting the speech from the second channel in the first language.

4. The method as recited in claim 3, wherein the steps of outputting the speech from the first channel and outputting the speech from the second channel are performed concurrently.

5. The method as recited in claim 1, further comprising determining signal to noise ratio (SNR) for each channel.

6. The method as recited in claim 5, wherein the SNR is employed to improve speech recognition accuracy.

7. The method as recited in claim 1, wherein designating separate input channels for each speaker includes designating a separate microphone for each speaker.

8. The method as recited in claim 1, further comprising determining a language spoken by a speaker if the language spoken is other than an expected language, warning a user that the language has changed.

9. A computer readable medium comprising a computer readable program, wherein the computer readable program when executed on a computer causes the computer to perform the steps of:

designating separate input channels for each of a plurality of speakers;

in response to speech from a first channel in a first language, translating the speech from the first channel to a second language; and

in response to speech from the second channel in a second language, translating the speech from the second channel to the first language;

wherein the steps of translating the speech from the first channel and translating the speech from the second channel are performed concurrently.

10. The computer readable medium as recited in claim 9, further comprising in response to speech from the first channel in the first language, outputting the speech from the first channel in the second language.

11. The computer readable medium as recited in claim 10, further comprising in response to speech from the second channel in the second language, outputting the speech from the second channel in the first language.

12. The computer readable medium as recited in claim 11, wherein the steps of outputting the speech from the first channel and outputting the speech from the second channel are performed concurrently.

13. A translation system, comprising:

a plurality of input channels where each input channel is configured to receive speech from a single speaker;

a memory configured to record speech from each channel to be translated and to store one or more training models and rules for translating the input speech; and

an automatic speech recognition engine and machine translator configured to concurrently translate speech for each channel separately and independently from each of any other channel from an input language to another language.

14. The system as recited in claim 13, further comprising a synthesizer configured to output translated speech to a speaker.

15. The system as recited in claim 14, wherein the synthesizer is configured to output translated speech from a plurality of speakers concurrently.

16. The system as recited in claim 13, further comprising a language detector configured to alert a user that an unexpected language is being spoken or an incorrect channel is being employed.

17. The system as recited in claim 13, wherein the plurality of input channels each includes a separate microphone.

18. The system as recited in claim 13, further comprising a display and a graphical user interface configured for user interaction with the system.

19. The system as recited in claim 13, wherein speech recognition accuracy is improved by accounting for noise on each channel.

20. The system as recited in claim 13, wherein the machine translator translates speech in an order that the speech is received over the plurality of channels.