# US 2020/0394271 A1 · Machine-assisted translation for subtitle localization > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2020/0394271 A1 - Document ID: US20200394271A1 - Status: Published application - Kind: A1 - Application number: 16/442,403 - Filed: 2019-06-14 - Published: 2020-12-17 - Earliest claimed: 2019-06-14 - Assignees: Netflix, Inc. - Applicants: NETFLIX, Inc. - Inventors: Ballav Bihani; Matthew James Rickard; Marianna Semeniakin; Ranjith Kumar Shetty; Allison Filemyr Smith; Patrick Brendon Pearson - Web page: https://patentgrep.com/patent/US20200394271A1 ## CPC - G06F 40/47: Machine-assisted translation, e.g. using translation memory - G06F 40/58: Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation - G06F 40/45: Example-based machine translation; Alignment - G10L 15/26: Speech to text systems (G10L15/08 takes precedence) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING ## Abstract One embodiment of the present disclosure sets forth a technique for generating translation suggestions. The technique includes receiving a sequence of source-language subtitle events associated with a content item, where each source-language subtitle event includes a different textual string representing a corresponding portion of the content item, generating a unit of translatable text based on a textual string included in at least one source-language subtitle event from the sequence, translating, via software executing on a machine, the unit of translatable text into target-language text, generating, based on the target-language text, at least one target-language subtitle event associated with a portion of the content item corresponding to the at least one source-language subtitle event, and generating, for display, a subtitle presentation template that includes the at least one target-language subtitle event. ## Claims ### Claim 1 A computer-implemented method for generating translation suggestions for subtitle events, the method comprising: receiving a sequence of source-language subtitle events associated with a content item, wherein each source-language subtitle event includes a different textual string representing a corresponding portion of the content item; generating a unit of translatable text based on a textual string included in at least one source-language subtitle event from the sequence; translating, via software executing on a machine, the unit of translatable text into target-language text; generating, based on the target-language text, at least one target-language subtitle event associated with a portion of the content item corresponding to the at least one source-language subtitle event; and generating, for display, a subtitle presentation template that includes the at least one target-language subtitle event. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, wherein generating the unit of translatable text comprises combining a textual string included in each of multiple subtitle events into a single unit of translatable text. ### Claim 3 (depends on 1) The computer-implemented method of claim 1, wherein generating the unit of translatable text comprises separating the textual string into multiple units of translatable text. ### Claim 4 (depends on 3) The computer-implemented method of claim 3, wherein the at least one source-language subtitle event includes no more than a single textual string. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, wherein translating the unit of translatable text into target-language text comprises: translating a first unit of translatable text based on a first textual string included in a first source-language subtitle event from the sequence into a first unit of target-language text; and translating a second unit of translatable text based on a second textual string included in a second source-language subtitle event from the sequence into a second unit of target-language text. ### Claim 6 (depends on 5) The computer-implemented method of claim 5, wherein generating the at least one target-language subtitle event comprises combining the first unit of target-language text and the second unit of target-language text into a single target-language subtitle event. ### Claim 7 (depends on 1) The computer-implemented method of claim 1, wherein generating the at least one target-language subtitle event comprises separating the target-language text into multiple target-language subtitle events. ### Claim 8 (depends on 1) The computer-implemented method of claim 1, wherein generating the subtitle presentation template comprises temporally aligning the target-language text with the unit of translatable text. ### Claim 9 (depends on 8) The computer-implemented method of claim 8, wherein temporally aligning the target-language text with the unit of translatable text comprises positioning in the subtitle presentation template a first target-language subtitle event that includes a portion of the target-language text to begin when a first source-language subtitle event that includes a portion of the unit of translatable text begins. ### Claim 10 (depends on 8) The computer-implemented method of claim 8, wherein temporally aligning the target-language text with the unit of translatable text comprises positioning in the subtitle presentation template a last target-language subtitle event that includes a portion of the target-language text to end when a last source-language subtitle event that includes a portion of the unit of translatable text ends. ### Claim 11 A non-transitory computer-readable storage medium includes instructions that, when executed by one or more processors, configure the one or more processors to perform the steps of: receiving a sequence of source-language subtitle events associated with a content item, wherein each source-language subtitle event includes a different textual string representing a corresponding portion of the content item; generating a unit of translatable text based on a textual string included in at least one source-language subtitle event from the sequence; causing the unit of translatable text to be translated into machine translated text in a target language; generating, based on the machine translated text, at least one target-language subtitle event associated with a portion of the content item corresponding to the at least one source-language subtitle event; and generating, for display, a subtitle presentation template that includes the at least one target-language subtitle event. ### Claim 12 (depends on 11) The non-transitory computer-readable storage medium of claim 11, wherein causing the unit of translatable text to be translated comprises: causing the unit of translatable text to be translated into first target-language text with a machine translation application; and causing the unit of translatable text to be translated into second target-language text with a translation memory application. ### Claim 13 (depends on 12) The non-transitory computer-readable storage medium of claim 12, further comprising instructions that, when executed by one or more processors, configure the one or more processors to perform the step of ranking a translation quality of the first target-language text and the second target-language text. ### Claim 14 (depends on 13) The non-transitory computer-readable storage medium of claim 13, wherein ranking the translation quality of the first target-language text and the second target-language text comprises: determining a first translation quality of the first target-language text based on one or more heuristic parameters associated with the unit of translatable text; determining a second translation quality of the second target-language text based on the one or more heuristic parameters; and based on the first translation quality and the second translation quality, selecting one of the first target-language text or the second target-language text as a highest quality translation. ### Claim 15 (depends on 13) The non-transitory computer-readable storage medium of claim 13, wherein ranking the translation quality of the first target-language text and the second target-language text comprises: determining a first translation quality of the first target-language text with a machine-learning model; determining a second translation quality of the second target-language text with the machine-learning model; and based on the first translation quality and the second translation quality, selecting one of the first target-language text and the second target-language text as a highest quality translation. ### Claim 16 (depends on 11) The non-transitory computer-readable storage medium of claim 11, wherein generating the unit of translatable text comprises combining a textual string from each of multiple source-language subtitle events from the sequence into a single unit of translatable text. ### Claim 17 (depends on 11) The non-transitory computer-readable storage medium of claim 11, wherein generating the unit of translatable text comprises separating the textual string into multiple units of translatable text. ### Claim 18 A system, comprising: a memory that stores instructions; and a processor that is coupled to the memory and, when executing the instructions, is configured to: receive a sequence of source-language subtitle events associated with a content item, wherein each source-language subtitle event includes a different textual string representing a corresponding portion of the content item; generate a unit of translatable text based on a textual string included in at least one source-language subtitle event from the sequence; translate, via software, the unit of translatable text into target-language text; generate, based on the target-language text, at least one target-language subtitle event associated with a portion of the content item corresponding to the at least one source-language subtitle event; and generate, for display, a subtitle presentation template that includes the at least one target-language subtitle event. ### Claim 19 (depends on 18) The system of claim 18, wherein the processor translates the unit of translatable text by: causing the unit of translatable text to be translated into first target-language text with a machine translation application; causing the unit of translatable text to be translated into second target-language text with a translation memory application; determining a first translation quality of the first target-language text with a machine-learning model; determining a second translation quality of the second target-language text with the machine-learning model; and based on the first translation quality and the second translation quality, selecting one of the first target-language text and the second target-language text as a highest quality translation. ### Claim 20 (depends on 18) The system of claim 18, wherein the subtitle template temporally arranges the sequence of source-language subtitle events with respect to a series of time codes that each indicate a specific playback time in the content item.