The present disclosure relates to content delivery and consumption systems and, more particularly, providing alternate language support for media content.
In the modern world of global data traffic, media content produced in one country is often made available in other countries. People of each country often have their own language and may not be able to understand, or may have difficulty understanding, the language in which the media content was produced. Many content producers include some alternate language data, but often only one or two languages are included, and a user may desire to view and/or hear at least some portions of the media content in a language that is not included.
Systems and methods are described herein for replaying content dialogue in an alternate language in response to a user command. While the content is playing on a media device, a first language in which the content dialogue is spoken is identified. Upon receiving a command to repeat a portion of the dialogue, the language in which the command was spoken is identified. The portion of the content dialogue to repeat is identified and translated from the first language to the second language. The translated portion of the content dialogue is then output. In this way, the user can simply ask in their native language for the dialogue to be repeated and the repeated portion of the dialogue is presented in the user's native language.
To translate the identified portion of the dialogue, the media device accesses subtitle text associated with the portion of the content dialogue. The media device translates the subtitle text to the second language and outputs for display the translated subtitles. Alternatively or additionally, the media device accesses audio data associated with the portion of the dialogue. The media device transcribes the audio into text and translates the text. The media device then uses a speech synthesizer to output audio of the translated dialogue.
The media device may determine the portion of the dialogue to repeat in several ways. In one example, the media device identifies a time at which the command was received and a duration of the command. The media device then calculates the start time of the command and identifies a portion of the dialogue in progress at the time start time of the command.
In a second example, the media device identifies a plurality of portions of the content dialogue within a predefined amount of time prior to the time at which the command was received. The media device determines an accent with which each portion of dialogue was spoken, compares all of the determined accents and determines, based on the comparison, whether any portion is spoken with an accent relative to the other portions. If a portion of the dialogue is spoken with an accent relative to the other portions, the media device selects that portion of the dialogue for translation.
The media device may, in a third example, determine the portion of the dialogue to repeat by determining a word count and duration of each portion of dialogue within the predefined time prior to the time the command was received. The media device then selects the portion of the dialogue having both the highest word count and the shortest duration for translation.
The media device detects the first language by transcribing a portion of the dialogue and comparing at least a first word of the transcribed text with a database of languages. If more than one language is identified as using the first word, additional words are compared until only one language is identified.
The above and other objects and advantages of the disclosure will be apparent upon consideration of the following detailed description, taken in conjunction with the accompanying drawings, in which like reference characters refer to like parts throughout, and in which:
Upon receiving the voice command 106, media device 100 identifies the language in which the voice command 106 was spoken. The media device 100 identifies the portion of the content dialogue 104 to repeat and translates the identified portion 104 from the first language to the second language. The media device 100 then outputs the translated portion of the content dialogue 108. Audio of the translated portion of the content dialogue 108 may be output using speakers integral to media device 100, external speakers physically connected to media device 100, or an audio output device to which media device 100 is connected via a wireless connection (e.g., Bluetooth speakers or an IoT device).
Control circuitry 400 comprises media access and playback circuitry 402. Media access and playback circuitry 402 receives content 102 from a content provider. The content provider may be a service (e.g., Netflix), a television network (e.g., NBC), or a media company (e.g., NBCUniversal). Content 102 may be received via a physical RF channel over a cable television connection or terrestrial broadcast, or may be received over an Internet connection from an over-the-top (OTT) service using a wired connection (e.g., Ethernet) or wireless connection (e.g., 802.11a/b/g/n (WiFi), WiMax, GSM, UTMS, CDMA, TDMA, 3G, 4G, 4G LTE, or any other suitable type of wireless data transmission). Media access and playback circuitry processes content 102 and outputs a corresponding video signal 102a and audio signal 102b using video output circuitry 404 and audio output circuitry 406, respectively.
Control circuitry 400 also comprises input circuitry 412. Input circuitry 412 receives 410 voice command 106 from microphone 408. Upon receiving voice command 106, input circuitry transfers 414 the audio data to language processing circuitry 416, which comprises language identification circuitry 418 and language translation circuitry 420. The audio data is processed using language identification circuitry 418 to determine the language in which the voice command 106 was spoken. For example, language identification circuitry 418 transcribes the audio into text and compares at least one word of the transcribed text to a language database, which may be stored locally in a memory of media device 100, or at a remote server. The language database may return an identifier of a language or languages in which the word appears. Language circuitry 418 may continue comparing words to the database until only a single language identifier is returned.
The input circuitry 412 also sends a command 422 to dialogue identification circuitry 424 to identify a portion of the content dialogue to translate. Dialogue identification circuitry 424 accesses audio data corresponding to the content 106 and identifies discrete portions of dialogue. For example, dialogue identification circuitry 106 may user frequency analysis to determine when each character in content 106 is speaking. Alternatively or additionally, dialogue identification circuitry 424 may access subtitle data of content 106 and identify discrete portions of the subtitle data corresponding to discrete portions of dialogue. Dialogue identification circuitry determines, based on a time the voice command 106 was received, a particular portion of the content dialogue to be translated.
Media access and playback circuitry 402 receives 426, from language identification circuitry 418, the identified language in which the voice command 106 was spoken. Media access and playback circuitry 402 also receives 428, from dialogue identification circuitry 424, an indication of the portion of dialogue to be translated. Media access and playback circuitry 402 may use the identified language to determine if a translation of the dialogue in the identified language is already included in the content 102. For example, content 102 may comprise a secondary audio stream with dialogue in the identified language. Media access and playback circuitry 402 extracts the identified portion of dialogue and any preexisting translation and transfers them 430 to language translation circuitry 420. Language translation circuitry 420 translates the extracted portion of the dialogue into the identified language. For example, language translation circuitry 420 transcribes the dialogue into text and processes the text through a translation algorithm to generate translated text corresponding to the dialogue. Language translation circuitry 420 generates audio corresponding to the translated text using a speech synthesizer, and transfers 432 the audio to audio output circuitry 406 for output. Audio output circuitry 406 may be connected to speakers that are integral to media device 100, a physical audio connector (e.g., a headphone jack) for use with external speakers, or a wireless audio connection (e.g., Bluetooth or WiFi) for use with an IoT audio device. In cases where a preexisting translation is available, language translation circuitry passes the translated audio directly to audio output circuitry 406 without further processing. In cases where subtitles are available in the identified language, language translation circuitry performs speech synthesis using the preexisting subtitle text.
At 502, control circuitry 400 detects a first language in which the content dialogue is spoken. For example, control circuitry 400 (e.g., using media access and playback circuitry 402) extracts a segment of audio data from the content 102. Control circuitry processes the extracted audio (e.g., using language processing circuitry 416) by transcribing the audio into corresponding text of the first language. For example, language processing circuitry 416 analyzes the audio of the content dialogue to identify various phonemes and compares the phonemes to a database of known words to determine which words were spoken in the content dialogue. Control circuitry 400 identifies the first language based on the transcribed text. This may be accomplished using methods described below in reference to
At 504, control circuitry 400 (e.g., using input circuitry 412) receives a command to repeat a portion of the content dialogue. The command may be an analog or digital signal received from a user input device, such as a remote control or the user's mobile device. The command may be a voice command received from the user input device or using microphone 408.
At 506, control circuitry 400 identifies a second language used by the user. If the command is a text command, control circuitry 400 identifies the language of the text. This may be accomplished using methods described below in reference to
At 508, control circuitry 400 determines whether the second language is different from the first language. Control circuitry 400 compares the language identifier of the first language with the language identifier of the second language. If the identifier of the first language does not match the identifier of the second language, then, at step 510, control circuitry 400 determines the portion of the content dialogue to repeat. This may be accomplished using methods described below in reference to
At 512, control circuitry 400 translates the portion of the content dialogue from the first language to the second language. Control circuitry 400 (e.g., using media access and playback circuitry 402) retrieves the audio data corresponding to the portion of the content dialogue. Control circuitry 400 then, using language processing circuitry 416, transcribes the portion of the content dialogue into corresponding text of the first language using methods described above, and translates the transcribed text into text of the second language. Control circuitry 400 may transmit the transcribed text to a third party translation service, such as Google Translate, or may perform the translation locally using any known methods.
At 514, control circuitry 400 outputs the translated content dialogue. This may be accomplished using methods described below in reference to
The actions or descriptions of
At 602, control circuitry 400 (e.g., using media access and playback circuitry 402) accesses metadata of the content 102. For example, content 102 may be an MPEG-2 stream, which contains metadata such as program specific information (e.g., using the PSIP protocol). Content 102 may be streaming content received from a server over a network (e.g., the Internet) and may include metadata in packet headers.
At 604, control circuitry 400 (e.g., using media access and playback circuitry 402) determines whether the content 102 includes subtitles in the second language. For example, media access and playback circuitry 402 retrieves, extracts, or otherwise analyzes the metadata of the content 102 for a field, flag, or other indicator of languages in which subtitles are available. If subtitles are available in the second language, processing proceeds to step 612. If subtitles are not available in the second language, then, at 606, control circuitry 400 (e.g., using media access and playback circuitry 402) determines whether any subtitles are available, for example, in the first language. If subtitles are available, then, at 610, control circuitry 400 translates the subtitle text from the first language to the second language using methods described above in reference to
At 612, control circuitry 400 (e.g., using video output circuitry 404) outputs for display the subtitles in the second language.
The actions or descriptions of
At 702, control circuitry 400 (e.g., using media access and playback circuitry 402) accesses metadata of the content 102. This may be accomplished using methods described above in reference to
At 704, control circuitry 400 (e.g., using media access and playback circuitry 402) determines if audio data is available in the second language. For example, media access and playback circuitry 402 retrieves, extracts, or otherwise analyzes the metadata of the content 102 for a field, flag, or other indicator of languages in which audio data is available. For example, the content 102 may contain multiple audio tracks, each comprising dialogue in a different language. Such alternate audio tracks may be included using the Secondary Audio Protocol or may be separate audio streams to which pointers are embedded in the content 102. For example, content 102 may be received as part of an MPEG-2 transport stream comprising a video stream and a plurality of audio streams corresponding to the content 102. The MPEG-2 transport stream also contains metadata, such as program-specific information (e.g., using the PSIP protocol) which identifies each video and audio stream contained within the MPEG-2 transport stream. Such metadata may identify a language of each audio stream corresponding to the content 102. If audio data is available in the second language, processing proceeds to step 714. If audio data is not available in the second language, then, at 706, control circuitry 400 (e.g., using media access and playback circuitry 402) accesses audio data associated with the portion of the content dialogue. For example, media access and playback circuitry 402 extracts, retrieves, or otherwise copies the audio data of the content 102 from a playback time at which the portion of the content dialogue begins to a second playback time at which the portion of the content dialogue ends. In some cases, where multiple characters speak at the same time, control circuitry 400 uses frequency filtering or other audio processing techniques to capture only the portion of the content dialogue to be translated.
At 708, control circuitry 400 (e.g., using language processing circuitry 416) transcribes the audio data associated with the portion of the content dialogue into text of the first language. This may be accomplished using methods described above in reference to
At 710, control circuitry 400 (e.g., using language processing circuitry 416) translates the text of the first language into text of the second language. This may be accomplished using methods described above in reference to
At 712, control circuitry 400 (e.g., using language processing circuitry 416) generates synthesized speech from the text of the second language. This may be accomplished using concatenation synthesis, such as unit selection synthesis, diphone synthesis, or domain-specific synthesis, formant synthesis, articulatory synthesis, hidden Markov models-based (HMM) synthesis, sinewave synthesis, or any other suitable speech synthesis technique.
At 714, control circuitry 400 (e.g., using audio output circuitry 406) outputs the audio of the content dialogue in the second language.
It is contemplated that the actions or descriptions of
At 802, control circuitry 400 (e.g., using input circuitry 412) receives a voice command from a user to repeat a portion of the content dialogue. For example, microphone 408 may capture a vocal utterance of the user and transfer audio data corresponding to the captured vocal utterance to input circuitry 412. Input circuitry 412 processes the captured vocal utterance and determines that the user wishes to repeat a portion of the content dialogue.
At 804, control circuitry 400 identifies a first time at which the voice command was received. In response to receiving the voice command, control circuitry 400 may access a real-time clock and store the time at which the command was received. Alternatively, when transferring the captured vocal utterance to input circuitry 412, microphone 408 may include a header or other metadata in which microphone 408 includes the time at which microphone 408 captured the vocal utterance.
At 806, control circuitry 400 determines a duration of the voice command. Control circuitry 400 processes the audio data corresponding to the vocal utterance and determine a duration of the captured vocal utterance. Alternatively, as above, microphone 408 may include in a header an indication of the duration of the vocal utterance.
At 808, control circuitry 400 calculates a second time at which the command began based on the first time and the duration. Control circuitry 400 converts the first time into an integer. For example, control circuitry 400 converts the first time into a UTC integer representing a number of seconds which have elapsed since Jan. 1, 1970. Control circuitry 400 subtracts the duration of the voice command from the integer representing the first time to calculate an integer representing the second time. Control circuitry 400 may convert the integer representing the second time to another time format, such as HH:MM:SS, and/or may convert the second time to a relative playback time of the content 102.
At 810, control circuitry 400 determines if any portion of the content dialogue is in progress at the second time. Control circuitry 400 (e.g., using media access and playback circuitry 412) may access audio data of the content 102 to determine if any character is speaking during the second time. Alternatively or additionally, media access playback circuitry 412 may access video data of the content 102 and analyze character facial movements to determine if any character is speaking at the second time. Then, at 812, control circuitry 400 identifies a portion of the content dialogue in progress at the second time.
It is contemplated that the actions or descriptions of
At 902, control circuitry 400 receives a voice command to repeat a portion of the content dialogue. This may be accomplished using methods described above in reference to
At 904, control circuitry 400 identifies a time Tcmd at which the voice command was received. This may be accomplished using methods described above in reference to
At 906, control circuitry 400 identifies a portion P of the content 102 from a predefined amount of time prior to Tcmd through Tcmd and, at 908, identifies a plurality of portions of the content dialogue that are in spoken during period P. This may be accomplished using methods described above in reference to
At 910, control circuitry 400 captures a first portion of dialogue D spoken during period P. This may be accomplished using methods described above in reference to
At 912, control circuitry 400 identifies an accent in which the dialogue of portion D is spoken. This may be accomplished using methods described in U.S. Pat. No. 9,854,324, which is owned by Applicant and is incorporated herein by reference.
At 914, control circuitry 400 determines whether there are any additional portions of the content dialogue in the plurality of portions of the content dialogue. If so, then, at 916, control circuitry 400 captures the next portion and processing returns to step 912, at which control circuitry 400 identifies an accent with which the portion was spoken.
If there are no additional portions, then, at 918, control circuitry 400 begins again at the first portion of the plurality of portions and, at 920, determines whether the accent with which the portion was spoken is different that the accents with which a majority of the remainder of the portions were spoken. If the accent is not different than that of the majority of the remainder of the portions, then, at 922, control circuitry moves to the next portion in the plurality of portions and processing returns to step 920.
If the accent of the portion is different than that of the majority of the remainder of the portions, then, at 924, control circuitry 400 selects the current portion D for translation.
It is contemplated that the actions or descriptions of
At 1002, control circuitry 400 (e.g., using input circuitry 412) receives a voice command to repeat a portion of the content dialogue. This may be accomplished using methods described above in reference to
At 1004, control circuitry 400 identifies a time Tcmd at which the voice command was received. This may be accomplished using methods described above in reference to
At 1006, control circuitry 400 identifies a portion P of the content 102 from a predefined amount of time prior to Tcmd through Tcmd and, at 1008, identifies a plurality of portions of the content dialogue that are in spoken during period P. This may be accomplished using methods described above in reference to
At 1010, control circuitry 400 captures a first portion of dialogue D spoken during period P. This may be accomplished using methods described above in reference to
At 1012, control circuitry 400 (e.g., using language processing circuitry 416) determines a word count for portion D. Language processing circuitry 416 may transcribe portion D to obtain a word count. Alternatively, language processing circuitry 416 may perform audio analysis on the dialogue of portion D to obtain a word count. For example, language processing circuitry 416 may track the frequency of a speaking voice to identify the speaker's phrasing, intonation, and cadence. Language processing circuitry 416 then determines specific points within the audio data of portion D at which the identified phrasing, intonation, and cadence indicate that the speaker has spoken a complete word.
At 1014, control circuitry 400 determines a duration of portion D. Control circuitry 400 (e.g., using media access and playback circuitry 402) determines a playback time at which portion D begins and a playback time at which portion D ends. Control circuitry 400 may convert the beginning and end times to respective integers. Control circuitry 400 subtracts the end time from the beginning time to calculate a duration of portion D.
At 1016, control circuitry 400 determines whether additional portions of dialogue are identified during period P. If additional portions are identified, then, at 1018, control circuitry 400 captures the next portion, and processing returns to step 1012, at which control circuitry 400 determines a word count for the next portion.
If no additional portions are identified during period P, then, at 1020, control circuitry 400 determines a particular portion having the highest word count and the shortest duration. Control circuitry 400 sorts the identified portions, first by word count in descending order (i.e., highest count first), and then by duration in ascending order (i.e., shortest duration first). Control circuitry 400 then determines the portion having the highest number of words in the shortest amount of time.
It is contemplated that the actions or descriptions of
At 1102, control circuitry 400 (e.g., using language processing circuitry 416) transcribes a random portion of the content dialogue into text of the first language. This may be accomplished using methods described above in reference to
At 1104, control circuitry 400 (e.g., using language processing circuitry 416) compares at least a first word of the transcribed text with a language database. The language database returns an identifier of each language in which the first word appears. At 1106, control circuitry 400 determines whether the word is found in more than one language. For example, the database returns an identifier of each language in which the word is found. If the word is found in more than one language, then, at 1108 control circuitry 400 compares a second word of the transcribed text with the language database, and processing returns to step 1106, at which control circuitry 400 determines whether the word is found in more than one language. If the word is only found in one language, then, at 1110, control circuitry 400 selects the language in which the word is found as the first language.
It is contemplated that the actions or descriptions of
The processes described above are intended to be illustrative and not limiting. One skilled in the art would appreciate that the steps of the processes discussed herein may be omitted, modified, combined, and/or rearranged, and any additional steps may be performed without departing from the scope of the invention. More generally, the above disclosure is meant to be exemplary and not limiting. Only the claims that follow are meant to set bounds as to what the present invention includes. Furthermore, it should be noted that the features and limitations described in any one embodiment may be applied to any other embodiment herein, and flowcharts or examples relating to one embodiment may be combined with any other embodiment in a suitable manner, done in different orders, or done in parallel. In addition, the systems and methods described herein may be performed in real time. It should also be noted that the systems and/or methods described above may be applied to, or used in accordance with, other systems and/or methods.
Number | Date | Country | |
---|---|---|---|
Parent | 16161883 | Oct 2018 | US |
Child | 17742657 | US |