Telefonaktiebolaget Lm Ericsson (publ)

Fisseha Mekuria

Joakim Persson

Hans Cavander

Per A Ljungberg

In a speech recognition system of a mobile telephone the words are organized in a trellis structure. Thus, at each instant the speech recognition system only needs to search a limited part of the entire vocabulary. Such an arrangement solves the problem of having to search many words at a time, which is time consuming and imposes a high computational load on the system, and will therefore significantly increase the accuracy for the speech recognition system.

Method and a system for voice dialling

A mobile terminal includes a receiver for receiving text messages over an RF channel. The mobile terminal also includes a text-to-speech (TTS) converter that converts the transmitted text messages to an audible form. In this way, the present invention takes advantage of the reduced bandwidth required for transmitting text messages to provide an audible message to subscribers that use the mobile terminals. In an exemplary embodiment, the mobile terminal operates in a GSM communication system and receives text messages that are defined under short message service (SMS) protocol. Also, the TTS converter in the mobile terminal can be used to output the text menus of the mobile terminal's interface in speech (voice) format.

Mobile terminal with a text-to-speech converter

A voice activity detector is disclosed for use with a radio transmitter to continuously sense the presence of speech in an audio signal. Initially, the audio signal is processed to produce a train of signal samples. Signal peaks are identified therefrom, which are used to compute respective values for a succession of quasi-pitch periods associated with the signal sample train. The quasi-pitch period values are then selectively compared with one another, in order to determine the presence or absence of a speech component.

Reduced complexity voice activity detector

A decoder includes an indexing and thresholding stage that receives a compressed image matrix and produces therefrom a signal that is based on a comparison of at least one element in the compressed image matrix with a predetermined threshold value. A transform stage is coupled to receive the signal from an output of the indexing and thresholding stage, wherein the transform stage selectively decodes the compressed image matrix based on the signal. A row column index for each coefficient in a DCT matrix is used, together with a threshold value, to access each coefficient, and decide if the entry is greater than the threshold value. This information is used by an IDCT decoder to avoid unnecessary MAC operations for those entries with values below the threshold value. By not performing the IDCT on the DCT matrix values below the threshold value, the computational requirement of the decoder is significantly reduced, thereby reducing the computational speed requirement for the decoder.

Low complexity JPEG decoder

In a voice communication system having a transmitter and receiver on opposing sides of an interface, the transmitter is switched on to transmit speech components and is switched off during speech pauses. To provide comfort noise at the receiver, and thus avoid annoying effects caused by continual switching of the transmitter, a comfort noise generator disposed to produce comfort noise of an adjustable amplitude is located on the receiver side of the interface. A first subsystem responsive to operation of the transmitter provides a flag to the receiver to commence operation of the comfort noise generator, when the transmitter discontinues transmission in response to a speech pause. A second subsystem transmits a succession of amplitude parameters through the interface to selectively adjust the amplitude of the generator in corresponding relationship with a noise level at the transmitter.

Method and apparatus for providing comfort noise in communication system with discontinuous transmission

A communication device is provided having a speech encoder and a speech decoder and being able to retrieve and store voice messages in a memory. The messages are stored in the memory according to a message format. This format is more compressed than the speech encoding format which is provided by the speech encoder. The device includes a frame interpolation block for decompressing a stored message and thereby creating a signal according to the speech encoding format. A frame decimation block is also provided for compressing a speech encoded signal thereby allowing a corresponding voice message to be stored in the memory according to the message format.

Communication device and method of operation

Apparatuses for noise reduction and noise processing methods for reducing noise in audio signals are presented. The noise level of an input signal at an input terminal is measured and the noise-to-signal ratio is established. A reduced voice activity detector is used to determine whether the input signal comprises speech or not. If the measured noise level exceeds a threshold level a switch connects the input signal to means for noise reduction. However, if the measured noise level does not exceed the threshold level, i.e. when noise reduction is not needed, the switch disconnects the means for noise reduction and the input signal is passed unchanged. Power is saved by powering off the means for noise reduction when it is not needed.

Noise reduction method and apparatus

A voice activity detector that implements a fast wavelet transformation using filter pairs. A quadrature high pass filter provides an output signal corresponding to the upper half of the Nyquist frequency and a quadrature low pass filter provides an output signal corresponding to the lower half of the Nyquist frequency. The quadrature high pass filter is useful for catching and isolating transients in the input signal and the quadrature low pass filter is useful for fine frequency analysis. The voice activity detector can utilize multiple decomposition levels that are arranged in a pyramid or tree formation to increase the reliability of the voice activity decision. For example, the output of the quadrature low pass filter can be further decomposed using a second pair of filters. The voice activity decision can be generated by comparing a signal power estimate for the output of the filter pairs to threshold levels that are specific for each filter or frequency range. The reliability of the voice activity decision is maximized by training the system to determine the optimum threshold levels and by basing the decision on a combination of the signal outputs. While increasing the number of decomposition levels increases the reliability of the voice activity decision, three decomposition levels is usually sufficient for detecting speech activity.

Method and apparatus for detecting voice activity

Speech or voice activity in an audio signal is defected without using a speech coder. Pitch period information and signal energy information are extracted from the audio signal, and a decision regarding the presence or absence of voice is made based on that information.

Non-parametric voice activity detection

A waveform-based technique for generating periodicity information from an input signal includes generating a pre-processed signal by applying low pass and non-linear filtering to the input signal, wherein the pre-processed signal has highlighted speech pitch tracks. An adaptive threshold algorithm is applied to the pre-processed signal to generate a detection signal having waveform segments whose peaks are separated by a pitch period of the input signal. A period between peaks in the detection signal is determined that indicates the periodicity information. Information about the period between the peaks in the detection signal is then used to adapt a scaling value to be used by the adaptive threshold algorithm in a subsequent step. The periodicity information may be utilized in a voice activity detector in a telephonic communications system.

Growing community of inventors

Lund, Sweden

Fisseha Mekuria