This application claims the benefit of United Kingdom Patent Application No. 0414711.2 filed Jul. 1, 2004.
This invention relates to speech recognition, and more particularly to a microphone initialization enhancement for automated speech recognition systems.
In the field of this invention it is known that many speaker phones and mobile phone car installations suppress the microphone while sound is being played out in order to eliminate echo and feedback. According to conventional teachings, the microphone remains in an off state until a relatively high volume audio signal is received. This microphone enablement technique causes the microphone to remain in an off state when a user speaks in a low to medium volume. Because the speaker phone or mobile kit does not react until after a high energy audio signal is received, any low level sounds at the start of the utterance can be lost. Accordingly, the system only receives a truncated part of the utterance resulting in a speech recognition of the utterance being unsuccessful.
This approach has the disadvantage that the spoken phrase is highly likely to be rejected outright or possibly confused with other candidate phrases. A need therefore exists for method of increasing speech recognition performance in such systems wherein the abovementioned disadvantage(s) may be alleviated.
In accordance with a first aspect of the present invention there is provided a method of speech recognition for use with a system having a microphone. The method includes receiving a signal from the microphone representative of speech to be recognized, performing detection of a transition in the signal indicative of switch on of the microphone, and, in response to the detection, performing speech recognition on the signal with reduced contribution from an initial portion thereof.
In accordance with a second aspect of the present invention there is provided a speech recognition system for use with a system having a microphone. The system includes means for receiving a signal from the microphone representative of speech to be recognized, means for performing detection of a transition in the signal indicative of switch on of the microphone, and means for performing, in response to the detection, speech recognition on the signal with reduced contribution from an initial portion thereof.
It should be noted that various aspects of the invention can be implemented as a program for controlling computing equipment to implement the functions described herein, or a program for enabling computing equipment to perform processes corresponding to the steps disclosed herein. This program may be provided by storing the program in a magnetic disk, an optical disk, a semiconductor memory, any other recording medium, or can also be provided as a digitally encoded signal conveyed via a carrier wave. The described program can be a single program or can be implemented as multiple subprograms, each of which interact within a single computing device or interact in a distributed fashion across a network space.
One method of improved speech recognition with speaker phones and car kits incorporating the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:
Referring firstly to
Referring now to
The speech detector 301 serves to automatically detect and enable the following mode of operation: the speech detector 301 identifies if the profile of the audio energy received from the speakerphone arrangement 200 is silence or close to silence followed by a high energy edge (rapid transition caused by the microphone 203 switching on); then, this situation will be notified by the speech recognition controller 302 to the speech recognition engine 303.
The speech recognition engine 303 will process the utterance but will, in light of the signal from the speech detector 301, automatically modify its behavior to not expect to match the initial sounds of the utterance. In effect, the initial sounds will become an optional part of the utterance.
Another example is to say that the recognition engine will accept deletions of sounds at the start of the utterance. Exactly how much (time) of the utterance might be allowed to be missing may be established by tuning the system for optimum recognition.
Alternatively, the grammars and call flows of the speech recognition engine may be constructed to optionally accept a redundant word (such as “please”) at the start of every utterance. This word is defined with special purpose “sound sequences” for this method which allow all possible tail ends of the utterance to match.
When the system detects repeated failures, it suggests to the user that they use the “microphone enabling” word which has the effect that the first thing the user says to interrupt the system is not required to match.
Good recognition can proceed with the significant portion of the grammar after this redundant word.
The core of this alternative mode of operation is illustrated with reference to
It will be understood that this method can be implemented without modification to speech recognition system software, only the speech recognition controller being new in this alternative.
It will be understood that in this alternative, the special word (“please” in the above example) can be redefined to match alternate “baseform” sequences (as known in WVS—WebSphere™ Voice Server—terminology) which, in this case, are the various truncated endings of the sound. The special word will not be usable in other grammars where it is not permissible to accept truncated utterances.
It will be understood that the above-described preferred embodiment's method of increasing speech recognition performance with speaker phones and car kits provides the following advantages.
While the reliability of recognition may be reduced when compared with a good (microphone unsuppressed or full duplex) telephony situation, especially with short utterances, recognition will be made possible when it is often impossible without this method.
Recognition match scoring will be increased as the low weighting given by deleted initial sounds will be eliminated and therefore confusion of the recognized phrase will be reduced.
The method described with reference to
The present invention may be realized in hardware, software, or a combination of hardware and software. The present invention may be realized in a centralized fashion in one computer system or in a distributed fashion where different elements are spread across several interconnected computer systems. Any kind of computer system or other apparatus adapted for carrying out the methods described herein is suited. A typical combination of hardware and software may be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein.
The present invention also may be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which when loaded in a computer system is able to carry out these methods. Computer program in the present context means any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following: a) conversion to another language, code or notation; b) reproduction in a different material form.
This invention may be embodied in other forms without departing from the spirit or essential attributes thereof. Accordingly, reference should be made to the following claims, rather than to the foregoing specification, as indicating the scope of the invention.
Number | Date | Country | Kind |
---|---|---|---|
0414711.2 | Jul 2004 | GB | national |