A method, system and storage medium for collaborative editing of grouped timed aligned rich format subtitles

Abstract

The application relates to audio and video subtitle processing technology and discloses a collaborative editing method and system for group time sequence alignment rich-format subtitles and a storage medium. The method establishes a stable identity identifier for a subtitle object, constructs a dynamic subtitle group according to a time adjacency relationship, a display association relationship or a semantic association relationship, displays and receives a predetermined editing event in groups in a subtitle list and a time axis, generates a semantic incremental patch according to original group content, a time boundary and a target input, executes local re-identification, re-translation or re-planning on a selected area, synchronizes the semantic incremental patch to a collaborative terminal for playback under version checking, and can adopt a concurrent slice identification and concurrent slice rendering / publishing mechanism to improve the efficiency of subtitle first draft generation, local processing and final publishing, so that the editing consistency of the rich-format subtitles, the local processing efficiency and the multi-person collaboration stability are improved.

Description

Technical Field

[0001] This application relates to audio and video subtitle processing technology, and more particularly to a method, system, and storage medium for collaborative editing, timeline adjustment, local reprocessing, and synchronization control of grouped time-aligned rich format subtitles. Background Technology

[0002] Currently, complex subtitle projects typically involve rich format information such as bilingual correspondence, word-level highlighting, character association, style inheritance, and fine-grained time boundaries. Existing technologies for this type of subtitle project mainly include the following processing solutions: 1. Line-level subtitle editing solutions: Traditional subtitle editing tools typically process single-line subtitles as the basic object. While they offer capabilities such as waveform / spectrum display, karaoke timing, and word-level timecode adjustment, they struggle to split, merge, and reorganize multiple subtitles with demonstrable and semantic relationships as a whole. 2. Online collaborative editing solutions: Existing collaborative subtitle editing solutions largely rely on full file overwriting, region-locked editing, master document synchronization, or coarse-grained version control. They struggle to support consistent linkage between group-level editing results in the subtitle list, timeline, and video preview, and also fail to reduce overwriting conflicts when multiple people are editing simultaneously. 3. Automatic transcription and translation solutions: Automatic speech recognition, speaker separation, and machine translation technologies can generate initial subtitle drafts, but the results are usually linear text fragments, lacking a closed-loop connection with dynamic subtitle group data models, group-level timeline adjustments, local reprocessing, and versioned patch playback. The above solutions present at least the following technical challenges: First, how to edit dynamic subtitle groups without disrupting standard text editing habits; second, how to maintain consistency in group-level structure and time boundaries across the subtitle list, timeline, and video preview; and third, how to synchronize and replay group-level changes with low conflict costs in multi-terminal collaboration. Therefore, a collaborative editing method, system, and storage medium for grouped, time-aligned, rich-format subtitles is needed to at least partially address these technical challenges. Summary of the Invention

[0003] In view of this, embodiments of this application provide a collaborative editing method, system, and storage medium for grouped time-aligned rich format subtitles, in order to at least solve one of the problems in the prior art.

[0004] In a first aspect, embodiments of this application provide a collaborative editing method for grouped, time-aligned, rich-format subtitles. The method first acquires the target media content and corresponding subtitle engineering data, generating an initial subtitle object with a stable identity identifier; then, it constructs dynamic subtitle groups based on temporal adjacency, display association, or semantic association, and displays them in the subtitle list and timeline as groups; next, upon receiving predetermined editing events such as Enter key, Backspace, deletion, target text input, drag-and-drop timeline adjustment, or selection area reprocessing, it determines group-level structural change operations based on the original group content, time boundaries, and target input, and generates a semantic incremental patch; then, based on the semantic incremental patch, it synchronously updates the subtitle list, timeline display, and video preview results, and synchronizes the semantic incremental patch to the collaborative terminal for playback under version control; if necessary, it can also perform local re-identification, re-translation, role recognition, or re-planning on the target selection area.

[0005] Secondly, embodiments of this application provide a collaborative editing system for grouped, time-aligned, rich-format subtitles. The system includes an initial subtitle generation module, a dynamic subtitle group construction module, an event planning module, a subtitle list linkage module, a timeline linkage module, a local reprocessing module, a patch synchronization module, a concurrency identification module, and a concurrent rendering and publishing module. These modules work collaboratively to achieve the aforementioned collaborative editing method.

[0006] Thirdly, embodiments of this application provide a storage medium storing a computer program, which, when executed by a processor, implements the aforementioned collaborative editing method.

[0007] According to the technical solution of this application, dynamic subtitle groups are used as the core editing object, semantic incremental patches are used as the collaborative carrier, and the interface is based on the linkage refresh between the subtitle list, timeline and video preview. This can reduce the reloading of full subtitles, reduce conflict windows in multi-person collaboration, reduce the number of operations required for splitting / merging dynamic subtitles, and improve the execution efficiency and result consistency of local reprocessing.

[0008] Additional advantages, objectives, and features of this application will be set forth in part in the description which follows, and will in part become apparent to those skilled in the art upon review of the following description, or may be learned by practice of the application. The objectives and other advantages of this application can be realized and obtained by means of the structures specifically pointed out in the specification and drawings.

[0009] Those skilled in the art will understand that the purposes and advantages that can be achieved with this application are not limited to those specifically described above, and that the above and other purposes that this application can achieve will be more clearly understood from the following detailed description. Attached Figure Description

[0010] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, do not constitute a limitation thereof. The components in the drawings are not drawn to scale but are merely for illustrating the principles of this application. For ease of illustration and description of certain parts of this application, corresponding portions in the drawings may be enlarged, i.e., may appear larger relative to other components in an exemplary device actually manufactured according to this application. In the drawings: Figure 1 This is a flowchart of a collaborative editing method according to an embodiment of this application; Figure 2 This is a schematic diagram of the dynamic subtitle group data structure according to an embodiment of this application; Figure 3 This is a schematic diagram illustrating the display and group-level editing of dynamic subtitle groups in a subtitle list according to an embodiment of this application; Figure 4 This is a schematic diagram illustrating the group-based display and group-level operation of a timeline according to an embodiment of this application; Figure 5 This is a schematic diagram illustrating partial reprocessing and automatic role recognition according to an embodiment of this application; Figure 6 This is a schematic diagram illustrating semantic incremental patching and multi-user collaborative synchronization according to an embodiment of this application; Figure 7 This is a schematic diagram illustrating multi-GPU container concurrency identification and sharding caching according to an embodiment of this application; Figure 8 This is a schematic diagram illustrating multi-server, multi-container concurrent rendering, shard caching, and automatic full deployment according to an embodiment of this application. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and their descriptions are used to explain this application, but are not intended to limit it.

[0012] In this application, unless otherwise specified, "initial subtitle object" refers to a data object that contains at least text content, time boundaries, and stable identity identifiers; "dynamic subtitle group" refers to a group-level editing unit composed of multiple subtitle objects based on time adjacency, display association, or semantic association; and "semantic incremental patch" refers to replayable incremental data used to characterize changes in group-level structure, text, and time boundaries.

[0013] The reference numerals in the drawings of this application shall follow the following rules: Figure 1 For the method flowchart, the step labels are represented by step numbers with the letter "S", such as steps S110 to S150; Figures 2 to 8For structure diagrams, interface diagrams, data structure diagrams, or system flowcharts, the figure labels should use numeric labels without the letter "S," and preferably be set according to the hundreds digit of the corresponding figure number. Figure 2 Using 2xx tags Figure 3 Using 3xx tags Figure 4 Using 4xx tags Figure 5 Using 5xx tags Figure 6 Using 6xx tags Figure 7 Using 7xx tags Figure 8 Use 8xx tagging.

[0014] In the accompanying drawings, the same reference numerals represent the same or similar parts, or the same or similar steps. First, refer to... Figure 1 A collaborative editing method 100 according to an embodiment of this application is described. For example... Figure 1 As shown, the collaborative editing method 100 may include steps S110 to S150, as follows: In step S110, the target media content and the corresponding subtitle engineering data are acquired to generate multiple initial subtitle objects. The subtitle engineering data can originate from imported existing subtitle files, or from automatic speech recognition, forced alignment, translation pipelines, or manual draft input. Each initial subtitle object includes at least text content, start time, end time, and a stable identity identifier; in optional embodiments, it may also include word-level start and end times, style identifiers, bilingual correspondence information, dynamic highlighting information, and role association information.

[0015] In step S120, a dynamic subtitle group is constructed based on the temporal adjacency, display association, or semantic association among the multiple initial subtitle objects. In one embodiment, such as... Figure 2 As shown, the group building module can combine the time intervals, style continuity, bilingual mapping relationships, role consistency, and word-by-word highlighting dependencies between adjacent subtitle objects to aggregate multiple subtitle objects into the same dynamic subtitle group. Each dynamic subtitle group can correspond to a group identifier, group boundary, set of objects within the group, word-level time intervals within the group, role identifier, and stable order index, and can be further used for group-level editing in the subtitle list, group-level timeline adjustment in the timeline, and local reprocessing.

[0016] in, Figure 2 In the attached diagram, 210 indicates the subtitle interface displayed by group, 220 indicates the internal structure of the dynamic subtitle group, 221 indicates the group identifier and group boundary field, 222 indicates the set of subtitle objects of group members, 223 indicates the word-level time interval and role / style related fields within the group, and 224 indicates the stable sequence index and group-level purpose field.

[0017] In step S130, in response to a predetermined editing event for a target dynamic subtitle group or subtitle unit within a group, a group-level structure change operation is determined based on the event type, original group content, time boundary, and target input. The predetermined editing event may include Enter, Backspace, Delete, target sentence input, dragging to adjust the timeline, and selecting a timeline area. Taking text editing events as an example, when the user triggers an Enter event, the system generates a group-level split operation based on the current cursor position and the original text and time boundary within the group; when the user triggers a Backspace event at the beginning of a group, the system generates an operation to merge with the previous group or sentence; when the user triggers a Delete event at the end of a group, the system generates an operation to merge with the next group or sentence.

[0018] In a preferred embodiment, when a user modifies the target text in a target dynamic subtitle group, the system does not immediately treat the modification as a simple string replacement. Instead, it first parses the correspondence between the target text and the original word sequence, and then selects the target operation sequence from candidate operations such as word replacement, word merging, word splitting, group-level splitting, group-level merging, and whole-group rewriting to obtain a new dynamic subtitle group structure and time boundary. This processing method can reduce the cost of subsequent manual timing readjustment while maintaining the consistency of editing habits.

[0019] In one embodiment, such as Figure 3 As shown, the dynamic subtitle groups in the subtitle list are displayed as group editing objects. The left side of each group editing object displays the start time, end time, and duration of the entire group, distinguishing it from the traditional list method that only displays the start and end times of individual subtitle lines. The right side is divided into upper and lower sections. The upper section displays the target language text and serves as an editable input area, while the lower section displays the original text after merging, removing tags, and de-highlighting from the original dynamic subtitle groups. The upper section also displays the current character name and / or character identifier; clicking the character name brings up a character list for the user to change characters.

[0020] in, Figure 3 In the attached diagram, 310 indicates the subtitle list interface; 311 indicates the dynamic subtitle groups displayed by group, as well as the start time, end time, and duration of the entire group on the left; 312 indicates the target language editing area and the original text display area; 313 indicates the character name, character list, and editing status; 314 indicates line break splitting, backspace / deletion merging, and automatic space and punctuation standardization; 315 indicates the whole sentence semantic planning method triggered by directly inputting the target sentence; and 316 indicates that after splitting or merging, the cursor is automatically positioned to the next target group and kept in an editable state.

[0021] Furthermore, after the user clicks on the upper part of the text, the system directly enters the editing state and allows the user to input the target text. The system can display the tagged text combination in the original multi-line time-aligned rich format subtitles as a group editing object, and automatically remove extra spaces and standardize the preset Chinese punctuation to English punctuation during the input process. When the user presses Enter in the target language input area, the system splits the dynamic subtitle group based on the current cursor position, the original group text, and the word-level time boundary, and automatically determines the corresponding split position of the other language text; when the user presses Backspace at the beginning of a line or Delete at the end of a line, the system merges with the previous group or sentence, and merges with the next group or sentence, respectively.

[0022] Furthermore, users can directly input the target sentence in the target language input area. The system utilizes a sentence-based semantic approach, determining the target operation sequence from candidate semantic operations such as splitting sentences, merging sentences, splitting words, merging words, and replacing words, based on the correspondence between the input target sentence and the original dynamic subtitle group. This automatically transforms the original multi-line dynamic subtitle into a target multi-line subtitle corresponding to the user's input, while maintaining the dynamic subtitle group format. After splitting or merging sentences, the system automatically positions the cursor to the next split or merged target group, keeping it editable. This ensures that editing complex multi-line tagged dynamic subtitle groups achieves the same user experience and efficiency as editing ordinary single-line subtitles.

[0023] In step S140, local reprocessing is performed based on the user's actions or selections on the timeline. In one embodiment, such as... Figure 4 As shown, the timeline is preferably displayed within a single overall display area, where the waveform illustration layer and the subtitle group display layer are superimposed in a layered manner, rather than placing the waveform and subtitle groups on separate tracks. The display and operation within the timeline can include displaying dynamic subtitle groups by group, adjusting the overall time interval, adjusting proportionally, adjusting only the start boundary, adjusting only the end boundary, and adjusting the time interval of words within a group. During dragging, the system can perform anti-overlap control and allow overlap according to preset rules when receiving shortcut keys or other unauthorized commands, thus balancing default safe timeline adjustment with advanced timeline adjustment needs.

[0024] in, Figure 4 In the attached diagram, 410 represents the overall timeline, 411 represents the waveform diagram layer displayed layer by layer with the subtitle group, 412 represents the dynamic subtitle group displayed by group and the upper and lower language areas, 413 represents the upper edge, lower edge and middle drag area, 414 represents the dynamic group proportional scaling and text adaptive hiding in scrolling and zooming states, and 420 represents the group-level timeline adjustment and local reprocessing operation triggered by timeline interaction.

[0025] Furthermore, in one embodiment, the display block of the target dynamic subtitle group on the timeline is divided into upper and lower parts, with the upper part used to display the first language content and the lower part used to display the second language content. The group edges and the time boundaries for word display within the group can be separated by solid lines and / or dashed lines, and draggable handles are provided at the edge positions and / or the middle positions. On a PC, the user can enter the drag mode by clicking with the mouse; on a mobile device, the user can enter the drag mode by long-pressing. The upper edge is used to perform proportional adjustments to the overall time interval of the entire dynamic subtitle group, the lower edge is used to adjust the display duration of the first word and / or the last word, and the middle area is used to drag the entire dynamic subtitle group to change the overall time interval corresponding to the dynamic subtitle group.

[0026] Furthermore, in one embodiment, the timeline supports scrolling and zooming operations. When zooming the timeline, the system can proportionally adjust the display size of the dynamic subtitle group within the timeline, and automatically hide text content exceeding the display threshold based on the current zoom ratio, display threshold, and available display width, in order to reduce display congestion while ensuring that group boundaries, key text, and word time boundaries are identifiable.

[0027] In one alternative implementation, such as Figure 5 As shown, local reprocessing may also include role recognition processing. Specifically, the system extracts voiceprint features from the media segment corresponding to the target selection area and compares these voiceprint features with the single-manuscript voiceprint library and the platform-level voiceprint library corresponding to the current manuscript. At the same time, it combines the manuscript metadata and the role semantic matching results output by the semantic matching model to determine the role name or role candidate result. When the matching confidence is lower than a preset threshold, the corresponding subtitle object is marked as awaiting manual review.

[0028] in, Figure 5 In the attached diagram, 510 represents the target selection area, 511 represents the input object for local reprocessing, 520 represents the process of local re-identification, local re-translation, local re-generation, and local re-planning, 521 represents the local result, 530 represents voiceprint feature extraction, 531 represents the comparison between the single-manuscript voiceprint library and the platform-level voiceprint library, 532 represents the semantic matching result of the role between the manuscript metadata and the output of the semantic matching model, 540 represents the role candidate result, and 550 represents the subsequent patch generation entry after writing the local result back to the subtitle object.

[0029] In step S150, the processing result obtained in step S130 or step S140 is converted into a semantic incremental patch, and version verification, patch synchronization, and patch replay are performed. In one embodiment, as... Figure 6As shown, the semantic incremental patch may include at least one or more of the following: base version number, target object identifier, target group identifier, operation type, old text, new text, old group content, new group content, word index, time anchor, role identifier, and author identifier. The operation type may include at least one of the following: group-level splitting, group-level merging, group-level rewriting, word replacement, word splitting, word merging, role rebinding, line insertion, and line deletion.

[0030] in, Figure 6 In the attached diagram, 610 represents the collaborative terminal, 611 represents the local caption project, text line mapping, and local preview rendering, 620 represents the patch pre-inspection, patch submission, version verification, conflict resolution, and local patch application process, 621 represents the semantic incremental patch field, 630 represents the collaborative server, 631 represents the version repository and patch repository, 632 represents event stream push, SSE, or polling, 633 represents the fragmented fetch or full fetch after local playback failure, and 640 represents the consistency control effect brought by patch playback.

[0031] Specifically, see Figure 6 Before applying the semantic incremental patch locally, the patch synchronization module can verify the base version number against the current subtitle project version number. If the versions match, the semantic incremental patch is replayed to the local subtitle project, and the subtitle list, timeline display, and video preview results are updated synchronously. If the versions do not match, conflict resolution window data is retrieved, and the patch to be submitted is replayed, retried, or conflict handled. By introducing a stable identity identifier and a patch replay mechanism, the risk of overwrite conflicts in multi-user collaboration can be significantly reduced, and preview desynchronization caused by full refresh can be minimized.

[0032] In one implementation of this application, see [link to implementation details]. Figure 6 The patch synchronization module can also maintain a mapping from stable identity identifiers to local sequential indexes, enabling rapid location of the target dynamic subtitle group and subtitle objects within the group when a remote patch is received. Additionally, it can prevent duplicate playback through patch fingerprints or idempotent identifiers. For shared re-identification states or remote local reprocessing results, the system can use either streaming event push or polling to restore the local state.

[0033] See Figure 7This application embodiment also provides a multi-GPU container concurrent recognition and segmented caching mechanism. Specifically, the system divides the target media content to be processed into multiple recognition segments and schedules them to multiple GPU containers to perform concurrent recognition tasks; an recognition segment cache is established for the intermediate results corresponding to each recognition segment; when a duplicate processing request, associated processing request, or partial rerun request is detected, the unexpired recognition segment cache is reused first; after all recognition segments are completed, they are merged into a unified subtitle draft in chronological order. This mechanism helps to improve recognition throughput and reduce duplicate recognition, corresponding to the concurrent segmented recognition method described in claim 8.

[0034] in, Figure 7 In the attached diagram, reference numeral 710 indicates the target media content and the identified segments; 711 indicates the initial identification request, repeated processing request, associated processing request, and partial rerun request; 720 indicates the concurrent identification tasks of the identification scheduler and multiple GPU containers; 721 indicates the identification segment cache and the unified subtitle draft merged in chronological order; 730 indicates the segment cache field; 731 indicates the cache reuse strategy; 732 indicates the cache invalidation condition; 733 indicates the subtitle draft after merging the identification results; and 740 indicates the technical effect brought about by concurrent identification and cache reuse.

[0035] See Figure 8 This application embodiment also provides a multi-server, multi-container concurrent rendering, chunk caching, and automatic full-release mechanism. Specifically, the system divides the rendering time interval into multiple rendering chunks, and the rendering scheduler distributes the rendering chunks to multiple rendering servers and multiple rendering containers for concurrent execution; a rendering chunk cache is established for the output results corresponding to each rendering chunk; when a rendering request is triggered by a partial rendering request, an automatic full-release request, or a change in associated materials, the rendering chunk cache that has not expired is reused first; after all target chunks are completed, they are concatenated into a final release file according to the output order, and after passing verification, a complete release result is automatically generated and written to the target release location. This mechanism corresponds to the concurrent chunk rendering method described in claim 9.

[0036] in, Figure 8 In the attached diagram, 810 represents the target subtitle project, output specifications, and rendering segments; 811 represents partial rendering requests, automatic full release requests, and rendering triggered by changes in related materials; 820 represents concurrent rendering tasks of the rendering scheduler, multiple rendering servers, and multiple rendering containers; 821 represents rendering segment caching, splicing output, and release files; 830 represents the rendering segment caching field; 831 represents the rendering cache reuse strategy; 832 represents the splicing output after all target segments are completed; 833 represents the automatic full release process; and 840 represents the technical effects brought about by concurrent rendering, cache reuse, and automatic release.

[0037] refer to Figures 2 to 8 This application also provides a collaborative editing system 200 for implementing the collaborative editing method 100. The collaborative editing system may include an initial subtitle generation module, a dynamic subtitle group construction module, an event planning module, a subtitle list linkage module, a timeline linkage module, a local reprocessing module, a patch synchronization module, a concurrency identification module, and a concurrent rendering and publishing module. The dynamic subtitle group construction module may correspond to... Figure 2 The data structure organization shown; the subtitle list linkage module can correspond to Figure 3 The group-level editing interface shown; the timeline linkage module can correspond to... Figure 4 The group-level axis adjustment mechanism shown; the local reprocessing module can correspond to... Figure 5 The local reprocessing and automatic role recognition mechanism shown; the patch synchronization module can correspond to... Figure 6 The semantic incremental patching and multi-user collaborative synchronization mechanism shown; the concurrency identification module can correspond to... Figure 7 The multi-GPU container concurrency identification and sharding caching mechanism shown; the concurrent rendering and publishing modules can correspond to... Figure 8 The diagram illustrates a multi-server, multi-container concurrent rendering, sharded caching, and automatic full deployment mechanism.

[0038] Depending on the needs, the collaborative editing system 200 can be deployed on a single terminal device or distributed among the subtitle editing terminal, collaborative server, and rendering server. Those skilled in the art will understand that the aforementioned modules can be implemented either in software or through a combination of software and hardware.

[0039] Furthermore, this application embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the corresponding steps of the aforementioned collaborative editing method 100. The storage medium may include a hard disk, a solid-state drive, a read-only memory, flash memory, or other computer-readable storage media.

[0040] Although exemplary embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above exemplary embodiments are merely illustrative and are not intended to limit the scope of this application. Various changes and modifications can be made therein by those skilled in the art without departing from the scope and spirit of this application. All such changes and modifications are intended to be included within the scope of this application as claimed in the appended claims.

[0041] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0042] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed.

[0043] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features but not others included in other embodiments, combinations of features from different embodiments are intended to be within the scope of this application and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.

[0044] It should be noted that the above embodiments are illustrative of this application and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

[0045] The above description is merely a specific embodiment or illustration of the embodiments of this application. The scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. The scope of protection of this application shall be determined by the scope of the claims.

Claims

1. A collaborative editing method for grouped, time-aligned, rich-format subtitles, characterized in that, The method, applied to a subtitle collaborative processing system, includes: Acquire target media content and subtitle engineering data corresponding to the target media content, and generate multiple initial subtitle objects. Each initial subtitle object includes at least text content, start time, end time, and stable identity identifier. Dynamic subtitle groups are constructed based on the temporal adjacency, display association, or semantic association among multiple initial subtitle objects, and the dynamic subtitle groups are displayed as editable units in the subtitle list and timeline; In response to a pre-defined editing event targeting a target dynamic subtitle group or subtitle unit within a group, the group-level structure change operation is determined based on the event type, original group content, time boundary, and target input, and a corresponding semantic incremental patch is generated. The subtitle list, timeline display, and video preview results are updated synchronously based on the semantic incremental patch. The semantic incremental patch carrying version information is submitted to the collaborative synchronization terminal so that other terminals can perform patch replay and consistency verification.

2. The collaborative editing method according to claim 1, characterized in that, The initial subtitle object also includes one or more of the following information: word-level start and end times, line-level start and end times, style identifiers, bilingual correspondence information, dynamic highlighting information, role association information, group identifier information, and stable sequence index.

3. The collaborative editing method according to claim 1, characterized in that, The dynamic subtitle groups in the subtitle list are displayed as group editing objects. Each group editing object includes at least a left-side time information area and a right-side upper and lower partitioned display area. The left-side time information area displays the start time, end time, and duration of the entire group. The upper half of the right-side upper and lower partitioned display area displays the target language text and serves as an editable input area, while the lower half displays the original text after merging, removing tags, and removing highlights from the original dynamic subtitle group. The upper half also displays the current character name and / or character identifier; clicking the character name will bring up a character list for changing characters. The pre-selected editing... The editing events include one or more of the following: splitting the entire group triggered by pressing Enter on the group editing object; merging with the previous group or sentence triggered by pressing Backspace at the beginning of the group; merging with the next group or sentence triggered by pressing Deletion at the end of the group; and replanning the entire group triggered after inputting a target sentence in the target language within the group. During input, the system removes extra spaces and standardizes preset punctuation. During replanning, the system determines the target operation sequence at least among the candidate operations of splitting entire sentences, merging entire sentences, splitting words, merging words, and replacing words. After splitting or merging, the system automatically positions the current cursor to the next target group and maintains an editable state. Specifically, the replanning of the entire group is used at least to determine the target operation sequence among the candidate operations of word replacement, word merging, word splitting, group-level splitting, group-level merging, and entire group rewriting.

4. The collaborative editing method according to claim 1, characterized in that, The display and operation in the timeline include: displaying waveform illustration layers and subtitle group display layers in a layered manner within the same overall timeline area, and displaying the dynamic subtitle groups by group; performing overall time interval adjustment, proportional adjustment, adjustment of only the start boundary, adjustment of only the end boundary, or adjustment of the time interval of words within the group on the target dynamic subtitle group; wherein, the upper half of the target dynamic subtitle group in the timeline is used to display the first language content, the lower half is used to display the second language content, the group edge and the word display time boundary are separated by solid lines and / or dashed lines, and a draggable handle is provided; the upper edge is used for proportional adjustment of the overall time, the lower edge is used for adjusting the duration of the first word and / or the last word, and the middle area is used for dragging the entire group to change the overall time interval; performing anti-overlap control during drag adjustment, and allowing overlap according to preset rules when receiving an over-authorization command; selecting the timeline area to be batch deleted, re-identified, or re-translated by dragging; and proportionally adjusting the display size of the dynamic subtitle group when scrolling and / or zooming the timeline, and automatically hiding text content exceeding the display threshold according to the zoom ratio.

5. The collaborative editing method according to claim 1, characterized in that, The method further includes: determining the local reprocessing range based on at least one of the following: the time range corresponding to the target selection area, the range of the preceding and following context windows, the range of the group boundary, and the role continuity constraint; Only the media segments and / or subtitle objects corresponding to the local reprocessing range are re-identified, re-translated, regenerated, or re-planned, without performing full reprocessing on the remaining unselected subtitle project parts.

6. The collaborative editing method according to claim 5, characterized in that, The local reprocessing also includes role recognition processing, which includes: extracting voiceprint features from the media segment corresponding to the target selection area; comparing the voiceprint features with the single-manuscript voiceprint library and the platform-level voiceprint library corresponding to the current manuscript; and determining the role name or role candidate result by combining the manuscript metadata and the role semantic matching result output by the semantic matching model. When the confidence level of the character matching is lower than the preset threshold, the corresponding subtitle object will be marked as awaiting manual review.

7. The collaborative editing method according to claim 1, characterized in that, The semantic incremental patch includes at least one or more of the following: base version number, target object identifier, target group identifier, operation type, old text, new text, old group content, new group content, word index, time anchor, role identifier, and author identifier; The operation types include at least one of group-level splitting, group-level merging, group-level rewriting, word replacement, word splitting, word merging, role rebinding, line insertion, and line deletion. The consistency check includes verifying the base version number and the current subtitle project version number, and obtaining conflict resolution window data when the versions do not match, and performing replay, retry or conflict handling on the patch to be submitted.

8. The collaborative editing method according to claim 1, characterized in that, The initial subtitle object is generated using a concurrent segmented recognition method, which includes: dividing the target media content into multiple recognition segments; scheduling multiple recognition segments to multiple processing containers for concurrent recognition; establishing a segment cache for the intermediate results corresponding to the recognition segments; reusing the unexpired segment cache when repeated processing requests, associated processing requests, or partial rerun requests arrive; and merging the results corresponding to multiple recognition segments into a first draft subtitle in chronological order after all recognition segments are completed.

9. The collaborative editing method according to claim 1, characterized in that, The rendering and / or publishing of the video preview results adopts a concurrent slice rendering method, including: dividing the time interval to be rendered into multiple rendering slices; scheduling multiple rendering slices to multiple rendering servers and multiple rendering containers to perform concurrent rendering; establishing a rendering slice cache for the output results corresponding to the rendering slices; reusing the unexpired rendering slice cache when a partial rendering request, an automatic full publishing request, or a rendering request triggered by changes in related materials arrives; and after all target rendering slices are completed, concatenating the results corresponding to multiple rendering slices into a final publishing file, and automatically generating a complete publishing result after passing verification.

10. A collaborative editing system for grouped, time-aligned, rich-format subtitles, characterized in that, include: The module includes: initial subtitle generation module, dynamic subtitle group construction module, event planning module, subtitle list linkage module, timeline linkage module, local reprocessing module, patch synchronization module, concurrency recognition module, and concurrent rendering and publishing module. The initial subtitle generation module generates initial subtitle objects with stable identity identifiers; the dynamic subtitle group construction module constructs dynamic subtitle groups; the event planning module responds to predetermined editing events and generates semantic incremental patches; the subtitle list linkage module updates the display and editing status of group editing objects in the subtitle list; the timeline linkage module updates the timeline display and video preview results; the local reprocessing module performs local re-identification, re-translation, role identification, or re-planning; the patch synchronization module performs version verification, patch playback, and conflict repair; the concurrent identification module performs multi-container concurrent identification and identification fragment cache reuse; and the concurrent rendering and publishing module performs multi-server multi-container concurrent rendering, rendering fragment cache reuse, and automatic full publishing, to achieve the method described in any one of claims 1 to 9.

11. A storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1 to 9.

Images