Real-time accent conversion model

A non-transitory computer-readable medium provisioned with program instructions that, when executed by at least one processor of a computing device, cause the computing device to:

• train a first machine-learning algorithm with first audio data comprising speech content captured from a plurality of different speakers having a first accent, wherein the training comprises aligning and classifying each of a first plurality of frames of the captured speech content corresponding to respective ones of the plurality of different speakers;
• apply the first machine-learning algorithm to speech content received via at least one microphone, and comprising a set of phonemes associated with a first pronunciation of the received speech content, to derive a non-text linguistic representation of the set of phonemes associated with the first pronunciation of the received speech content;
• based on the derived non-text linguistic representation of the set of phonemes associated with the first pronunciation of the received speech content, synthesize, using a second machine-learning-algorithm trained with (i) second audio data comprising the first accent and (ii) third audio data comprising a second accent, fourth audio data representative of the received speech content having the second accent, wherein synthesizing the fourth audio data comprises mapping at least a first non-text linguistic representation of a first phoneme of the set of phonemes associated with the first pronunciation of the received speech content to a second non-text linguistic representation of a second phoneme of an updated set of phonemes associated with a second pronunciation of the received speech content that is different from the first pronunciation of the received speech content, wherein the first and second phonemes are different phonemes;
• and
• convert the synthesized fourth audio data into a synthesized version of the received speech content having the second accent, wherein the synthesized version of the received speech content having the second accent comprises the updated set of phonemes associated with the second pronunciation of the received speech content.

A method comprising:

• training a first machine-learning algorithm with first audio data comprising speech content captured from a plurality of different speakers having a first accent, wherein the training comprises aligning and classifying each of a first plurality of frames of the captured speech content corresponding to respective ones of the plurality of different speakers;
• applying the first machine-learning algorithm to speech content received via at least one microphone, and comprising a set of phonemes associated with a first pronunciation of the received speech content, to derive a non-text linguistic representation of the set of phonemes associated with the first pronunciation of the received speech content;
• based on the derived non-text linguistic representation of the set of phonemes associated with the first pronunciation of the received speech content, synthesizing, using a second machine-learning-algorithm trained with (i) second audio data comprising the first accent and (ii) third audio data comprising a second accent, fourth audio data representative of the received speech content having the second accent, wherein synthesizing the fourth audio data comprises mapping at least a first non-text linguistic representation of a first phoneme of the set of phonemes associated with the first pronunciation of the received speech content to a second non-text linguistic representation of a second phoneme of an updated set of phonemes associated with a second pronunciation of the received speech content that is different from the first pronunciation of the received speech content, wherein the first and second phonemes are different phonemes;
• and
• converting the synthesized fourth audio data into a synthesized version of the received speech content having the second accent, wherein the synthesized version of the received speech content having the second accent comprises the updated set of phonemes associated with the second pronunciation of the received speech content.