Abstract: In an example embodiment, there is disclosed herein an automatic speech recognition (ASR) system that employs speaker clustering (or speaker type) for transcribing audio. A large corpus of audio with corresponding transcripts is analyzed to determine a plurality of speaker types (e.g., dialects). The ASR system is trained for each speaker type. Upon encountering a new user, the ASR system attempts to map the user to a speaker type. After the new user is mapped to a speaker type, the ASR employs the speaker type for transcribing audio from the new user.

Abstract: According to certain embodiments, training a transcription system includes accessing recorded voice data of a user from one or more sources. The recorded voice data comprises voice samples. A transcript of the recorded voice data is accessed. The transcript comprises text representing one or more words of each voice sample. The transcript and the recorded voice data are provided to a transcription system to generate a voice profile for the user. The voice profile comprises information used to convert a voice sample to corresponding text.

Abstract: A system for facilitating network performance analysis. In an illustrative embodiment, the system includes a first module capable of quantifying network performance associated with a first portion of the network and providing a signal in response thereto. A second module is capable of ascertaining performance degradations in a communications link traversing plural portions of the network that are attributable to the first portion based on the signal. In a more specific embodiment, the second module further includes a third module adapted to ascertain degradations in network performance associated with a second portion of the network based on the signal and based on a performance parameter associated with both the first portion and the second portion of the network. In this embodiment, the first portion includes a wireless portion, and the second portion includes a wired portion.

Abstract: In one method embodiment, receiving from the network device a multiplex of a compressed video stream and a compressed audio stream, the multiplex comprising a succession of intervals corresponding to a video program corresponding to a first playout rate; and at the start of each interval, replacing the compressed audio stream with a compressed, pitch-preserving audio stream corresponding to a second playout rate different than the first.

Abstract: In one method embodiment, providing a multiplex of compressed versions of a first video stream and a first audio stream, each corresponding to an audiovisual (A/V) program, the first video stream and the first audio stream each corresponding to a first playout rate and un-synchronized with each other for an initial playout portion; and providing a compressed version of a second audio stream, the second audio stream corresponding to a pitch-preserving, second playout rate different than the first playout rate, the second audio stream synchronized to the initial playout portion of the first video stream when the first video stream is played out at the second playout rate, the first audio stream replaceable by the second audio stream for the initial playout portion.

Abstract: In an example embodiment, a control connection is introduced between an adaptive jitter buffer (AJB) and an adaptive bulk delay (ABD) buffer of an echo canceller (ECAN) in an Internet Protocol (IP) conference bridge. The control connection allows the AJB to control the amount of delay inserted by the ABD in the ECAN convolution processor (CP) signal path. The adjustment in ABD delay restores the time alignment of the ECAN internal echo estimate and offsets variations in echo tail delay introduced by network induced AJB delay changes. Time-invariance is preserved in the echo tail path.

Abstract: A method is disclosed for providing fault tolerance of an application over a Stream Control Transmission Protocol (SCTP) connection. A method for moving a transport connection from a first host to a second host on a network is also described. After establishing a transport connection between the first host and a peer host, first information about the transport connection is transferred to the second host. Upon the first host becoming unavailable, the second host sends to the peer host a request to update the state of the transport connection. The peer host then updates the transport connection. Without resetting the transport connection, the second host then establishes the same transport connection with the peer host.

Abstract: Classifying signals includes receiving signals from communication devices, where each signal is received from a communication device. The following is repeated for each signal: a received signal from the communication device is encoded, whether the encoded signal is an undesirable auditory signal is determined, and the encoded signal is modified if the encoded signal comprises an undesirable auditory signal. The received signals are combined.

Abstract: A method in one embodiment includes embedding a unique event tag into media files recorded at an event by a first media recording device, which is operable to wirelessly discover a second media recording device at the event. The first media recording device wirelessly communicates the unique event tag to the second media recording device for embedding into media files recorded at the event by the second media recording device. It is emphasized that this abstract is provided to comply with the rules requiring an abstract that will allow a searcher or other reader to quickly ascertain the subject matter of the technical disclosure.

Abstract: In one embodiment, a signal transmission authentication apparatus includes an input operable to receive a changing signal, a first extractor operable to extract first phoneme data from the received changing signal, a first generator including logic operable to generate first data representative of the extracted first phoneme data, a first output operable to communicate output data corresponding to the received changing signal and the first data from an associated transmitter to an associated receiver, a second extractor associated with the receiver including logic operable to extract second phoneme data from the received output data via the receiver and regenerate the first phoneme data from the received first data, a comparator including logic operable to generate a comparison signal in accordance with a comparison of the first and second phoneme data, and a second output operable to generate a signal corresponding to authenticity of the received changing signal in accordance with an output of the comparator

Abstract: Particular embodiments provide for attenuating one or more microphone signals in a teleconferencing system upon detecting a non-audible signal. A far end voice signal is received from a sound source. A non-audible signal is added to the far end voice signal to create a composite signal, which is provided to one or more speakers. The speakers output the composite signal and the non-audible signal are detected in the composite signal after the composite signal is received at one or more microphones. The non-audible signal allows an attenuator to attenuate a microphone signal including the composite signal from a particular microphone in response to the detected non-audible signal to reduce far end echo.

Abstract: A system and technique for analyzing incoming packets. The path taken by incoming packets can be analyzed from a near end without having access to the far end. This can be done even though packets sent to the far end from the near end, take a different path than that used by the incoming packets. The process is started by an RTP no op packet sent from the near end to the far end. The no op packet request the far end to initiate a trace using a series of media packets that have sequential values in the TTL field. The results of the trace are either send from the far end to the near end and a report is complied at the near end, or the results of the trace are compiled into a report at the far end and the report is sent to the near end.

Abstract: An interface receives a packet stream associated with a source and a destination. The packet stream includes multiple packets, each packet including a header and a payload. A system generates a modulated stream for inclusion in the packet stream and performs a low deviation modification to a header field in each of the multiple packets to include the modulated stream in the packet stream. The system then transmits the packet stream including the modulated stream.

Abstract: An echo canceller apparatus comprises a receive side attenuator coupled in a receive side signal path that is configured to couple from a conference call bridge to a caller; a convolution processor coupled to the receive side signal path at a convolution processor pick-off point; a double-talk detector coupled to the receive side signal path and to a sending side signal path that is configured to couple from the caller to the conference call bridge; and logic coupled to the receive side attenuator which when executed is responsive to a double-talk condition detected by the double-talk detector and operable to determine a level of echo canceled by the convolution processor, to determine an additional amount of attenuation to introduce, and to activate the receive side attenuator to introduce the additional attenuation.

Abstract: A method in one embodiment includes embedding a unique event tag into media files recorded at an event by a first media recording device, which is operable to wirelessly discover a second media recording device at the event. The first media recording device wirelessly communicates the unique event tag to the second media recording device for embedding into media files recorded at the event by the second media recording device. It is emphasized that this abstract is provided to comply with the rules requiring an abstract that will allow a searcher or other reader to quickly ascertain the subject matter of the technical disclosure.

Abstract: In one embodiment, an echo canceller configured to cancel echo in a wideband voice conference is provided. A double-talk condition may be when a plurality of users are speaking substantially simultaneously. When a double-talk condition is detected in the wideband conference, a high-frequency process is enabled and used to process signals in the high band to reduce echo. Accordingly, echo in the high band may not be produced by end devices being used by the users' speaking. Also, the users speaking have the echo cancelled in the low band and substantial echo does not result. This results in the users speaking experiencing the conference in the narrowband. The other users that are not speaking, however, continue to receive wideband signals. The users not speaking also continue to have echo cancellation performed for the high band and low band because these users are not speaking and thus attenuation of their voices is not a consideration.

Abstract: A method for handling media in a make-before-break handoff between wireless networks is provided that includes communicating with a mobile station to facilitate a session handoff between two call legs of a session. The method also includes identifying a relevant frame of one of the legs, performing a cross correlation operation, generating a delay difference estimate between the two legs, and performing a level difference estimation for the two legs.

Abstract: An interface receives a packet stream associated with a source and a destination. The packet stream includes multiple packets, each packet including a header and a payload. A system generates a modulated stream for inclusion in the packet stream and performs a low deviation modification to a header field in each of the multiple packets to include the modulated stream in the packet stream. The system then transmits the packet stream including the modulated stream.

Abstract: Methods and apparatus are provided for compressing delay sensitive signals. Frames including multiple samples of the delay sensitive signal are analyzed to determine characteristics associated with the frames, such as the range of quantization levels represented by the samples of the frames. The delay sensitive signal is then compressed by providing an anchor point is sent along with information for determining the variation of each sample from the anchor point.

Abstract: A mobile device supports concurrent registrations of different wireless personalities with one or more service provider networks. Potentially using a single wireless interface, the mobile device can register and maintain multiple personalities and monitor for notifications targeted to any one of these personalities.