摘要:
Disclosed herein are various aspects of a toolkit used for generating a TTS voice for use in a spoken dialog system. The embodiments in each case may be in the form of the system, a computer-readable medium or a method for generating the TTS voice. An embodiment of the invention relates to a method of tracking progress in developing a text-to-speech (TTS) voice. The method comprises insuring that a corpus of recorded speech contains reading errors and matches an associated written text, creating a tuple for each utterance in the corpus and tracking progress for each utterance utilizing the tuple. Various parameters may be tracked using the tuple but the tuple provides a means for enabling multiple workers to efficiently process a database of utterance in preparation of a TTS voice.
摘要:
The present invention provides various elements of a toolkit used for generating a TTS voice for use in a spoken dialog system. The embodiments in each case may be in the form of the system, a computer-readable medium or a method for generating the TTS voice. One embodiment of the invention relates to a method of correcting a database associated with the development of a text-to-speech (TTS) voice. The method comprises generating a pronunciation dictionary for use with a TTS voice, generating a TTS voice to a stage wherein it is prepared to be tested before being deployed, identifying mislabeled phonetic units associated with the TTS voice, for each identified mislabeled phonetic unit, linking to an entry within the pronunciation dictionary to correct the entry and deleting utterances and all associated data for unacceptable utterances.
摘要:
Disclosed are various elements of a toolkit used for generating a TTS voice for use in a spoken dialog system. The invention in each case may be in the form of the system, a computer-readable medium or a method for generating the TTS voice. An embodiment of the invention relates to a method for preparing a text-to-speech (TTS) voice for testing and verification. The method comprises processing a TTS voice to be ready for testing, synthesizing words utilizing the TTS voice, presenting to a person a smallest possible subset that contains at least N instances of a group of units in the TTS voice, receiving information from the person associated with corrections needed to the TTS voice and making corrections to the TTS voice according to the received information.
摘要:
A coin discrimination device which provides for positive transport of coins past a sensor is provided. In one embodiment, a rotating disk, slightly inclined to the vertical, includes a plurality of pockets for receiving coins from a hopper bowl. The disk and bowl are configured to efficiently load the pockets with no more than one coin per pocket. As the disk rotates, the coins, which are thus registered in a desired location in the pockets, are transported by the rotating disk past a sensor. When, based on the sensor data, the item in a pocket is recognized as an acceptable coin, a controllable ramp is lowered to divert the coin out of the pocket and into a path to place the coin in an acceptance bin or bag.
摘要:
Disclosed herein are various aspects of a toolkit used for generating a TTS voice for use in a spoken dialog system. The embodiments in each case may be in the form of the system, a computer-readable medium or a method for generating the TTS voice. An embodiment of the invention relates to a method of tracking progress in developing a text-to-speech (TTS) voice. The method comprises insuring that a corpus of recorded speech contains reading errors and matches an associated written text, creating a tuple for each utterance in the corpus and tracking progress for each utterance utilizing the tuple. Various parameters may be tracked using the tuple but the tuple provides a means for enabling multiple workers to efficiently process a database of utterance in preparation of a TTS voice.
摘要:
One embodiment of the present invention provides a method for facilitating asymmetric line rates in an Ethernet passive optical network (EPON) which includes a central node and at least one remote node. During operation, the system provides a downstream code-group clock, wherein each cycle thereof corresponds to a code group transmitted from the central node to a remote node. The system also provides an upstream code-group clock, wherein each cycle thereof corresponds to a code group received at the central node from a remote node. In addition, the system provides a multi-point control protocol (MPCP) clock, wherein the frequency ratio of the MPCP clock to the downstream code-group clock is different from the frequency ratio of the MPCP clock to the upstream code-group clock, thereby allowing the downstream transmission to be performed at a faster line rate than the upstream transmission line rate.
摘要:
One embodiment of the present invention provides a system that decrypts downstream data in an Ethernet passive optical network (EPON). During operation, the system receives a data frame which is encrypted based on a remote input block and a session key, wherein the remote input block is constructed based on a remote cipher counter and a remote block counter. The system adjusts a local cipher counter based on a received checksum located in a preamble of the data frame, wherein the local cipher counter is substantially synchronized with the remote cipher counter. In addition, the system truncates the local cipher counter by discarding n least significant bits thereof. The system then constructs a local input block based on the truncated cipher counter and a local block counter for the received data frame. Next, the system decrypts the data frame based on the local input block and the session key.
摘要:
One embodiment of the present invention provides a system that facilitates bandwidth-efficient multicast in EPONs. The system includes a central node and at least one remote node. Downstream data from the central node is broadcast through a passive optical coupler to remote nodes. In the other direction, upstream data from a remote node is transmitted through the passive optical coupler to the central node in a unicast manner. During operation, the system first selects a common multicast LLID to represent a number of remote nodes that comprise a multicast group. Upon receiving a multicast message destined to the multicast group, the system sends the multicast message along with the multicast LLID for the multicast group, whereby the multicast message is broadcast through the passive optical coupler in the downstream direction. This allows each remote node belonging to the multicast group to receive the multicast message by matching the multicast LLID.
摘要:
One embodiment of the present invention provides a system that facilitates registration of remote nodes in an Ethernet passive optical network (EPON). The system includes a central node and at least one remote node, wherein a number of virtual remote nodes are coupled to a common physical remote node and transmit upstream data through a common transmitter within the common physical remote node. During an initial discovery cycle, the system receives a solicitation message from the central node at a remote node, wherein the solicitation message assigns a discovery slot in which an unregistered remote node may transmit a response message to the central node for registration. The system then transmits a response message from an unregistered remote node to register within the assigned discovery slot. If multiple virtual remote nodes coupled to the common physical remote node seek registration concurrently, the system controls the transmission of response messages from these virtual remote nodes to the central node so as to avoid collisions between response messages from different remote nodes.