Automatic emphasis of spoken words
AssigneeAdobe Systems Incorporated
InventorsYang Zhang, Gautham J. Mysore, Floraine Berthouzoz
Life of this patent
Estimating the term…
Abstract
Embodiments of the present invention provide systems, methods, and computer storage media directed towards automatic emphasis of spoken words. In one embodiment, a process may begin by identifying, within an audio recording, a word that is to be emphasized. Once identified, contextual and lexical information relating to the emphasized word can be extracted from the audio recording. This contextual and lexical information can be utilized in conjunction with a predictive model to determine a set of emphasis parameters for the identified word. These emphasis parameters can then be applied to the identified word to cause the word to be emphasized. Other embodiments may be described and/or claimed.
Claims20
One or more non-transitory computer-readable media having a plurality of executable instructions embodied thereon, which, when executed by a processor of a computing device cause the computing device to:
- receive an audio recording of spoken content for audibly emphasizing an identified word with respect to other words included in the spoken content;
- extract contextual and lexical information, from the audio recording, relating to the identified word;
- determine emphasis parameters based on the contextual and lexical information and a predictive model that correlates the contextual and lexical information with the determined emphasis parameters;
- and
- apply the emphasis parameters to modify the identified word such that when modified, the identified word is audibly emphasized within the audio recording of the spoken content and with respect to the other words included in the spoken content.
A computer-implemented method to facilitate audibly emphasizing an identified word, the method comprising:
- receiving a first audio recording of first spoken content that includes a plurality of audibly emphasized words;
- determining contextual and lexical information based on the first spoken content, wherein the contextual and lexical information is associated with each of the plurality of audibly emphasized words;
- generating a predictive pitch model by correlating the contextual and lexical information associated with each of the plurality of audibly emphasized words included in the first spoken content with pitch parameters, the pitch parameters including a starting pitch level, a peak pitch level, and an ending pitch level for each of the plurality of audibly emphasized words;
- modifying a second audio recording of second spoken content that includes the identified word, by automatically applying emphasis parameters, based on the predictive pitch model, to the second audio recording, such that the identified word is audibly emphasized with respect to other words included in the second spoken content.
A computing device to facilitate a user of the computing system in automatically emphasizing spoken words, the computing device comprising:
- one or more processors;
- and
- a memory, coupled with the one or more processors, having a plurality of executable instructions embodied thereon, which, when executed by the one or more processors cause the computing device to:
- identify, within an audio recording of spoken content, a word to be audibly emphasized, with respect to other words included in the spoken content;
- extract contextual and lexical information from the audio recording relating to the identified word;
- determine predicted pitch parameters based on the contextual and lexical information and a predictive model that correlates the contextual and lexical information with the predicted pitch parameters, the predicted pitch parameters including a predicted starting pitch level, a predicted peak pitch level, and a predicted ending pitch level for the identified word;
- and
- apply the predicted pitch parameters to modify the identified word by calculating a pitch contour based on the predicted starting pitch level, the predicted peak pitch level, and the predicted ending pitch level, such that when modified, the identified word is audibly emphasized with respect to the other words included in the spoken content.
Dependencies are inferred from the claim text.