TWI894591B - Sound ray configuration method, system and computer program product - Google Patents

Abstract

A voice configuration method implemented by a voice configuration system includes: (A) obtaining a plurality of character data and a plurality of voice groups, each voice group including a plurality of voice data corresponding to a plurality of specific voices, wherein the voice data in each voice group includes an initial voice data and one or more derived voice data generated by at least performing an audio range adjustment on the initial voice data. (B) for each character data, selecting a matching voice group from the voice groups that corresponds to the same character type as the character data, and then setting one of the voice data in the matching voice group as a matching voice data corresponding to the character data.

Description

Sound ray configuration method, system and computer program product

The present invention relates to a voice configuration method, and more particularly to a voice configuration method suitable for assigning appropriate voices to multiple characters. The present invention also relates to a voice configuration system suitable for assigning appropriate voices to multiple characters, and a computer program product.

In dubbing work, different characters' lines are usually interpreted using different voices. In addition to real-life dubbing, using computer-synthesized voices (i.e. virtual human voices) for dubbing has become another feasible approach.

Existing computer-generated speech synthesis (CSTS) software often offers a variety of virtual human voices to choose from. However, to achieve a more natural dubbing experience, it's important to avoid using voices with accents or overly unnatural timbre. This can limit the selection of suitable voices CSTS software can offer.

Therefore, given the large number of characters requiring dubbing but limited choice of suitable voices, how to better utilize computer-generated speech to complete dubbing has become a topic worth exploring.

Therefore, one of the objects of the present invention is to provide a voice configuration method that helps to improve the dubbing effect of multiple characters.

The present invention provides a voice ray configuration method implemented by a voice ray configuration system. The method comprises: (A) obtaining voice ray requirement information and a voice ray data set. The voice ray requirement information includes a plurality of character data corresponding to a plurality of characters, each character data corresponding to one of a plurality of character types. The voice ray data set includes a plurality of voice ray groups, each voice ray group corresponding to one of the character types and including a plurality of voice ray data. The voice ray data in each voice ray group corresponds to a plurality of specific voices and includes an initial voice ray data and one or more derived voice ray data generated by at least adjusting the audio range of the initial voice ray data. (B) For each character data, a matching voice group corresponding to the same character type as the character data is selected from the voice groups, and one of the voice data in the matching voice group is set as a matching voice data corresponding to the character data, thereby allocating the specific voice corresponding to the matching voice data to the character corresponding to the character data.

In some embodiments of the voice configuration method of the present invention, in step (A), one or more of the character data each includes a voice feature tag associated with an audio frequency range and is each designated as a key character data item. In step (B), if the character data item is key character data, the voice configuration system selects the matching voice ray group from the voice ray groups and then sets one of the voice ray data items in the matching voice ray group that matches the voice feature tag of the key character data item as the paired voice ray data item corresponding to the key character data item.

In some embodiments of the voice configuration method of the present invention, in step (B), for multiple important character data corresponding to the same character type in the character data, if multiple voice groupings in the voice groupings correspond to the same character type as the important character data, the voice configuration system selects the matching voice data corresponding to the important character data from the voice groupings in turn.

In some embodiments of the voice configuration method of the present invention, in step (A), at least one of the character data is selected as a main character data, and the voice group used in the voice groups to determine the matching voice data corresponding to the main character data is selected as a protagonist voice group. In step (B), for multiple important character data corresponding to the same character type in the character data, if multiple voice groups in the voice groups correspond to the same character type as the important character data, but the protagonist voice group exists in the voice groups, the voice configuration system does not select the matching voice data corresponding to the important character data from the protagonist voice group.

In some embodiments of the voice ray configuration method of the present invention, in step (A), one or more pieces of character data that do not include a voice ray feature label are each treated as a secondary character data. In step (B), for multiple pieces of secondary character data corresponding to the same character type in the character data, if multiple voice ray groups in the voice ray groups correspond to the same character type as the secondary character data, the voice ray configuration system selects the matching voice ray data corresponding to each of the secondary character data in turn from the voice ray groups.

In some embodiments of the voice configuration method of the present invention, the voice configuration method further includes, after step (B): (C) for one of the examined character data in the character data and the paired voice data currently corresponding to the examined character data, judging whether a voice conflict condition is met based on text data corresponding to the voice requirement information, wherein the text data includes a plurality of lines, each line part corresponds to one of the character data in the character data and the paired voice data currently corresponding to the one of the character data, and the examined character data corresponds to Each dialogue portion is considered a checked dialogue portion, and the voice conflict condition includes: there is a neighboring dialogue portion among the dialogue portions that meets a proximity condition and corresponds to a character data other than the checked character data, and the matching voice data corresponding to the neighboring dialogue portion is the same as the matching voice data corresponding to the checked character data. For each dialogue portion other than the checked dialogue portion(s), the proximity condition indicates that the number of characters between the dialogue portion and any of the checked dialogue portions in the text data is less than or equal to a preset character count threshold. (D) If the voice conflict condition is determined to be met, a voice conflict prompt indicating the checked character data is output.

In some embodiments of the voice ray configuration method of the present invention, in step (D), when it is determined that the voice ray conflict condition is met, the voice ray configuration system first updates the voice ray data corresponding to the inspected character data. Only after the voice ray data corresponding to the inspected character data is updated and it is again determined that the voice ray conflict condition is met, does the voice ray conflict prompt be output.

Another object of the present invention is to provide a voice configuration system that helps improve the dubbing effect of multiple characters.

The voice configuration system of the present invention includes a processing unit and a storage unit electrically connected to the processing unit, wherein the processing unit is used to: obtain voice demand information and a voice data combination, the voice demand information includes a plurality of character data corresponding to a plurality of characters, and each character data corresponds to one of a plurality of character types; the voice data combination includes a plurality of voice groups, each voice group corresponds to one of the character types and includes a plurality of voice data, and the voice data of each voice group corresponds to a plurality of A specific voice line is provided, and includes an initial voice line data and one or more derived voice line data generated by at least adjusting the audio range of the initial voice line data; for each character data, a matching voice line group corresponding to the same character type as the character data is selected from the voice line groups, and one of the voice line data in the matching voice line group is set as a matching voice line data corresponding to the character data, thereby allocating the specific voice line corresponding to the matching voice line data to the character corresponding to the character data.

In some embodiments of the voice configuration system of the present invention, one or more of the character data each includes a voice feature tag associated with an audio frequency range and is each considered as a key character data item. If the character data item is a key character data item, the processing unit selects the matching voice ray group from the voice ray groups and then sets one of the voice ray data items in the matching voice ray group that matches the voice feature tag of the key character data item as the paired voice ray data item corresponding to the key character data item.

In some embodiments of the voice configuration system of the present invention, for multiple important character data corresponding to the same character type in the character data, if multiple voice groups in the voice groups correspond to the same character type as the important character data, the processing unit selects the matching voice data corresponding to the important character data from the voice groups in turn.

In some embodiments of the voice configuration system of the present invention, at least one of the character data is designated as a main character data, and the voice group used in the voice groups to determine the matching voice data corresponding to the main character data is designated as a protagonist voice group. For multiple pieces of important character data corresponding to the same character type in the character data, if multiple voice groups in the voice groups correspond to the same character type as the important character data, but the protagonist voice group is present in the voice groups, the processing unit does not select the matching voice data corresponding to the important character data from the protagonist voice group.

In some embodiments of the voice configuration system of the present invention, one or more pieces of character data that do not include a voice feature tag are each treated as secondary character data. For multiple pieces of secondary character data corresponding to the same character type in the character data, if multiple voice groups in the voice groups correspond to the same character type as the secondary character data, the processing unit sequentially selects the matching voice data corresponding to the secondary character data from the voice groups.

In some embodiments of the voice configuration system of the present invention, the processing unit is further used to: after setting the matching voice data corresponding to each character data, for one of the inspected character data in the character data and the matching voice data currently corresponding to the inspected character data, determine whether a voice conflict condition is met based on text data corresponding to the voice requirement information, wherein the text data includes multiple dialogue parts, each dialogue part corresponds to one of the character data in the character data and the matching voice data currently corresponding to one of the character data, and each dialogue part corresponding to the inspected character data is used as a received Check the dialogue parts, and the voice conflict condition includes: there is a neighboring dialogue part among the dialogue parts that meets a neighboring condition and the corresponding character data is not the neighboring dialogue part of the checked character data, and the matching voice data corresponding to the neighboring dialogue part is the same as the matching voice data corresponding to the checked character data, wherein, for each dialogue part other than the (etc.) checked dialogue part(s), the neighboring condition represents that the number of characters between the dialogue part and any checked dialogue part in the text data is less than or equal to a preset character number threshold; when it is determined that the voice conflict condition is met, a voice conflict prompt indicating the checked character data is output.

In some embodiments of the sound ray configuration system of the present invention, when it is determined that the sound ray conflict condition is met, the processing unit first updates the sound ray data corresponding to the inspected character data, and only outputs the sound ray conflict prompt after the sound ray data corresponding to the inspected character data is updated and it is again determined that the sound ray conflict condition is met.

Another object of the present invention is to provide a computer program product that helps improve the dubbing effect of multiple characters.

The computer program product of the present invention includes an application program, wherein when the application program is loaded and executed by an electronic device, the electronic device can implement the sound ray configuration method as described in any of the aforementioned embodiments.

The effectiveness of the present invention lies in that, when the number of voices with suitable accents and timbres is limited, the voice configuration system can use a suitable specific voice (i.e., the specific voice corresponding to the initial voice data) as a basis to expand more specific voices with suitable accents and timbres (i.e., the specific voice corresponding to the derived voice data), thereby increasing the number of suitable voices to choose from. This helps achieve better dubbing effects when the number of suitable voices is limited.

Before the present invention is described in detail, it should be noted that, unless otherwise specified, the term "electrically connected" as used in this patent specification is used to describe the "coupled" relationship between computer hardware (e.g., electronic systems, devices, apparatuses, units, components), and generally refers to "wired electrical connections" achieved by physically connecting multiple computer hardware components to each other through conductive/semiconductor materials, as well as "radio connections" achieved by wireless data transmission using wireless communication technologies (such as, but not limited to, wireless networks, Bluetooth, and electromagnetic induction). On the other hand, unless otherwise specified, the term "electrical connection" as used in this patent specification broadly refers to "direct electrical connection" achieved by directly coupling multiple computer hardware components to each other, as well as "indirect electrical connection" achieved by indirectly coupling multiple computer hardware components through other computer hardware components. Furthermore, the term "voice" as used in this patent specification refers to the overall auditory quality of speech, including acoustic characteristics of speech such as the frequency range (also known as pitch), timbre, accent (also known as intonation), and speech rate. "Timbre" is a sensory property of sound that allows listeners to distinguish between two sounds of the same loudness and pitch.

Referring to FIG. 1 , one embodiment of the sound ray configuration system 1 of the present invention is implemented as a computer device (e.g., a desktop computer, laptop computer, or server). Furthermore, the sound ray configuration system 1 includes a processing unit 11 and a storage unit 12 electrically connected to the processing unit 11. More specifically, in this embodiment, the processing unit 11 is a processor implemented as an integrated circuit and having data processing and instruction transmission and reception functions, while the storage unit 12 is a data storage device for storing digital data (e.g., a hard drive or other type of computer-readable recording medium). However, in similar implementations, the processing unit 11 may also be a collection of multiple processors, or a processing circuit including a processor, and the storage unit 12 may also be a collection of multiple storage devices of the same or different types. Furthermore, in different implementations, the sound ray configuration system 1 may also be implemented as a different type of electronic device, such as a tablet computer or smartphone. Based on the above, the actual implementation of the sound ray configuration system 1 in terms of computer hardware is not limited to this embodiment.

In this embodiment, the storage unit 12 stores a virtual voice database DB, for example. The virtual voice database DB is created by the processing unit 11 running computer speech synthesis software, and includes multiple voice profiles corresponding to various virtual voices. Specifically, each voice profile can be used by the processing unit 11 to control a speaker (not shown) to produce the virtual voice corresponding to the voice profile. Therefore, if a text string is paired with the voice profile, the processing unit 11 can control the speaker to play the text string using the virtual voice corresponding to the voice profile, thereby achieving computer-synthesized speech output.

It should be noted that the computer speech synthesis software and the voice profiles can be implemented using existing technologies, so their details are not described in detail here.

With reference to FIG. 2 , the following exemplarily describes in detail how the sound ray configuration system 1 of this embodiment implements a sound ray configuration method.

First, in step S1, the processing unit 11 obtains a text data and a sound ray demand information D1 corresponding to the text data (shown in FIG3 ).

In this embodiment, the text data is, for example, a text file of a novel, and is equivalent to a target to be dubbed in this embodiment. The text data indicates multiple characters and includes multiple dialogue sections, where each dialogue section is a dialogue or monologue of one of the characters and includes one or more sentences presented in natural language.

As shown in FIG3 , the voice requirement information D1 includes a plurality of character data 10 , and the character data 10 respectively correspond to the characters indicated by the text data.

Each character data 10 includes a character tag 101, and the character tag 101 indicates a character type corresponding to the character data 10, such as "adult male", "adult female", "child male", and "child female" shown in FIG3, but is not limited thereto. It should be noted that, depending on the content of the text data, multiple pieces of character data 10 may correspond to the same character type. For example, the character data 10A, character data 10D, and character data 10E shown in FIG3 all correspond to the character type of "adult male".

In this embodiment, multiple pieces of character data 10 (e.g., character data 10A through 10E shown in FIG3 ) further include a voice characteristic tag 102. For each piece of character data 10 including a voice characteristic tag 102, the voice characteristic tag 102 indicates which pitch of voice within the corresponding character type is suitable for dubbing the character corresponding to the character data 10. For example, character data 10B in FIG3 corresponds to the "adult female" character type, and its voice characteristic tag 102 indicates "high pitch," indicating that the character corresponding to the character data 10B is suitable for dubbing with a voice of the "high pitch" pitch within the "adult female" character type.

In this embodiment, one of the character data 10 (e.g., character data 10A in FIG. 3 ) is designated as a main character data 10*. In other words, the character corresponding to the main character data 10* corresponds to a protagonist (e.g., a male protagonist, but not limited thereto) in the text data. On the other hand, for each of the character data 10 other than the main character data 10*, each character data 10 including the voice feature tag 102 (e.g., character data 10B through character data 10E in FIG. 3 ) is designated as an important character data 10′. In other words, the characters corresponding to the important character data 10′ correspond to multiple important supporting roles in the text data. On the other hand, each character data 10 that does not include the voice feature label 102 (such as character data 10F to character data 10J in Figure 3) is regarded as a secondary character data 10". In other words, the characters corresponding to the secondary character data 10" are equivalent to multiple minor supporting roles in the text data.

It should be noted that, depending on the actual content of the text data, in different implementations of this embodiment, only a single piece of character data 10 may include a voice feature label 102. Alternatively, more than one piece of character data 10 may be designated as primary character data 10*. It should be understood that FIG3 merely illustrates an exemplary embodiment of the voice requirement information D1. Therefore, the actual embodiment of the voice requirement information D1 is certainly not limited to this embodiment.

In addition, in this embodiment, the processing unit 11 first obtains the text data, and then uses a language model to segment, break down sentences, and perform natural language understanding on the text data to summarize the roles indicated by the text data and the voice characteristics suitable for each role, thereby generating the voice requirement information D1. In other words, in this embodiment, the role data 10 and the role tags 101 and voice feature labels 102 included therein are determined by the processing unit 11 through semantic analysis of the text data. However, optionally, the voice requirement information D1 can also be generated by the processing unit 11 through semantic analysis of the text data and manual editing and adjustment by the user. In addition, in different embodiments, the processing unit 11 may also receive the text data and the sound ray demand information D1 from other electronic devices in advance and store them in the storage unit 12, and read the storage unit 12 when starting to execute the sound ray configuration method to obtain the text data and the sound ray demand information D1 from the storage unit 12. Therefore, the sound ray demand information D1 is not limited to being generated by the processing unit 11, and the way in which the processing unit 11 obtains the sound ray demand information D1 is not limited to this embodiment.

After the processing unit 11 obtains the text data and the sound ray demand information D1, the process proceeds to step S2.

In step S2, the processing unit 11 obtains a sound ray data set D2 (shown in FIG4 ) based on the virtual human voice database DB.

As shown in FIG4 , the voice ray data set D2 includes a plurality of voice ray groups 20 (six in this embodiment, but not limited thereto). Each voice ray group 20 corresponds to one of the character types, such as, but not limited to, the aforementioned "adult male," "adult female," "child male," and "child female." In this embodiment, the character types to which the voice ray groups 20 in the voice ray data set D2 correspond can be pre-determined. However, in other embodiments, the processing unit 11 can also determine these character types in real time based on the character tags 101 in the character data 10.

Each voice ray group 20 includes a plurality of voice ray data 201 corresponding to a plurality of specific voice lines suitable for interpreting the voices of the character type to which the voice ray group 20 corresponds. For example, in voice ray group 20A in FIG4 , the five voice ray data 201 included in voice ray group 20A correspond to five specific voice lines with different pitches suitable for interpreting the character "adult male." Each sound ray data 201 includes, for example, a plurality of voice setting parameters. Furthermore, these voice setting parameters are used to collectively define the voice acoustic characteristics, such as the frequency range, timbre, accent, and speaking speed, of the specific sound ray corresponding to the sound ray data 201. Thus, if a string of text is combined with the sound ray data 201, the processing unit 11 can control the speaker to play the string of text using the specific sound ray corresponding to the sound ray data 201.

In this embodiment, the specific voice lines corresponding to the voice line data 201 of the same voice line group 20 exhibit the same accent (e.g., the accent of a specific region), but differ at least in pitch. In other words, if the specific voice lines corresponding to the voice line data 201 of the same voice line group 20 are used to produce speech for the same sentence, the specific voice lines will exhibit the same pronunciation pattern for the sentence, but will exhibit different overall pitch.

More specifically, for the sound ray data 201 in the same sound ray group 20, one piece of sound ray data 201 is an initial sound ray data selected by the processing unit 11 from the virtual voice database DB. This initial sound ray data may, for example, be one of the voice profiles contained in the virtual voice database DB. Furthermore, within the sound ray data 201 in the same sound ray group 20, each piece of sound ray data 201 other than the initial sound ray data is a piece of derived sound ray data. Each piece of derived sound ray data is generated by, for example, the processing unit 11 by adjusting the audio frequency range of the fundamental pitch of the initial sound ray data. Therefore, the specific sound ray corresponding to each derived sound ray data is equivalent to another sound ray with a higher or lower pitch derived by the processing unit 11 based on the specific sound ray corresponding to the initial sound ray data and at least by increasing or decreasing the audio frequency range of its fundamental tone.

Taking Figure 4 as an example, within each sound ray group 20, the sound ray data 201 labeled "Mid-range" is, for example, the initial sound ray data selected by the processing unit 11 from the virtual human voice database DB within that sound ray group 20. Meanwhile, the sound ray data 201 labeled "Bass" or "Extreme Bass" is, for example, derived sound ray data generated by the processing unit 11 by lowering the fundamental frequency range of the initial sound ray data. Furthermore, the sound ray data 201 labeled "High-pitched" or "Extreme-high-pitched" is, for example, derived sound ray data generated by the processing unit 11 by raising the fundamental frequency range of the initial sound ray data. In this way, the present embodiment can adjust the fundamental frequency range of a single specific voice line, thereby expanding other specific voice lines with different pitches. In this way, even if the number of specific voice lines with suitable accents and timbres in the virtual human voice database DB is limited, the present embodiment can expand more specific voice lines with suitable accents and timbres based on the suitable specific voice line, thereby increasing the number of suitable specific voice lines.

Furthermore, in an advanced implementation of this embodiment, after the processing unit 11 adjusts the fundamental audio frequency range of one of the initial sound ray data to generate a corresponding derivative sound ray data, the processing unit 11 can, for example, use the derivative sound ray data as the initial sound ray data of another sound ray group 20 corresponding to a different character type. For example, assuming one of the initial voice ray data corresponds to the "adult female" character type, the processing unit 11, after increasing the fundamental pitch frequency range of the initial voice ray data to generate corresponding derivative voice ray data with a sharper voice, generates another voice ray group 20 corresponding to the "young male" character type based on the derived voice ray data, and uses the derived voice ray data as the initial voice ray data included in the voice ray group 20 corresponding to the "young male" character type. Therefore, by adjusting the fundamental pitch frequency range of the initial voice ray data to generate the derived voice ray data, this embodiment can utilize the same specific voice ray with a suitable accent/timbre to expand other specific voices suitable for characters of different ages/genders.

It should be noted that, in various implementations, one or more of the sound ray groups 20 may include only a single piece of derived sound ray data. Furthermore, optionally, within the same sound ray group 20, any piece of derived sound ray data may be generated by the processing unit 11 by simultaneously adjusting the audio frequency range of the fundamental pitch and the speech rate of the initial sound ray data.

In this embodiment, because certain character types have a relatively high demand for voices, multiple voice groups 20 within the voice groupings 20 correspond to the same character type. For example, voice groupings 20A and 20B in FIG4 both correspond to the "adult male" character type, while voice groupings 20C and 20D both correspond to the "adult female" character type. It should be noted that each voice grouping 20 is unique in at least one of its timbre and accent characteristics. Therefore, even if two voice groups 20 correspond to the same character type, the two specific voices corresponding to those two voice groups 20 will differ in at least one of their timbre and accent characteristics. In other words, for a specific voice ray corresponding to any voice ray data 201 of any voice ray group 20, the speech presented by the specific voice ray is unique in the combination of pitch, timbre and accent.

After the processing unit 11 obtains the sound ray data combination D2, the process proceeds to step S3.

In step S3, the processing unit 11 matches the character data 10 in the voice requirement information D1 with the voice data 201 in the voice data combination D2 to establish a correspondence between each character data 10 and one of the voice data 201. In this embodiment, the processing unit 11 prioritizes the voice data 201 corresponding to the primary character data 10*, then determines the voice data 201 corresponding to the important character data 10', and finally determines the voice data 201 corresponding to the secondary character data 10", but this is not a limitation.

More specifically, for each character data 10, the processing unit 11 first selects a matching voice group from the voice groups 20 that corresponds to the same character type as the character data 10, and then sets one of the voice data 201 in the matching voice group as a matching voice data corresponding to the character data 10, thereby allocating the specific voice corresponding to the matching voice data to the character corresponding to the character data 10. For example, if the character data 10B in FIG3 corresponds to the character type of "adult female", the processing unit 11 will select one of the voice groups 20 (such as voice group 20C or voice group 20D) in FIG4 that also corresponds to the character type of "adult female" as a matching voice group, and then select the matching voice data corresponding to the character data 10B from them.

For each character data 10, if the character data 10 is main character data 10* or important character data 10' including a voice feature tag 102, the processing unit 11 selects a corresponding matching voice group from the voice groups 20 and then sets one of the voice data 201 of the matching voice group that matches the voice feature tag 102 of the character data 10 as the matching voice data corresponding to the character data 10.

For example, taking character data 10A in FIG3 , character data 10A corresponds to the character type "adult male" and includes a voice feature tag 102 indicating "bass." Therefore, assuming that processing unit 11 selects voice group 20A in FIG4 as a matching voice group, processing unit 11 will further select voice data 201 matching voice feature tag 102 of "bass" from voice group 20A as the matching voice data corresponding to character data 10A. Furthermore, since character data 10A is the main character data 10*, in this example, voice group 20A will be used as a main voice group 20* among the voice groups 20 to determine the matching voice data corresponding to the main character data 10*.

Taking the character data 10B in FIG. 3 as another example, the character data 10B corresponds to the character type of "adult female" and includes a voice feature label 102 indicating a "high pitch". Therefore, assuming that the processing unit 11 selects the voice group 20C in FIG. 4 as the matching voice group, the processing unit 11 will further select the voice data 201 that matches the voice feature label 102 of "high pitch" from the voice group 20C as the matching voice data corresponding to the character data 10B.

In this embodiment, for multiple important character data 10' corresponding to the same character type in the character data 10, if multiple voice groups 20 in the voice groups 20 correspond to the same character type as the important character data 10', and there is no protagonist voice group 20* in the voice groups 20, then the processing unit 11 selects the matching voice data corresponding to the important character data 10' respectively from the voice groups 20 in turn. For example, the character data 10B and character data 10C in Figure 3 and the voice group 20C and voice group 20D in Figure 4 all correspond to the character type of "adult female", and the voice group 20C and voice group 20D do not belong to the protagonist voice group 20*. In this case, the processing unit 11 will follow the order in which the two characters corresponding to the character data 10B and the character data 10C first appear in the text data, and for example, first select the matching voice data corresponding to the character data 10B from the voice group 20C, and then select the matching voice data corresponding to the character data 10C from the voice group 20D, so that the matching voice data corresponding to the character data 10B and the character data 10C respectively come from different voice groups 20. In this way, even if the character data 10B and the character data 10C both correspond to the "adult female" character type, this embodiment can still configure two specific voices with different timbres or accents for the character data 10B and the character data 10C, which helps to avoid the voices of multiple important supporting characters being too similar.

In another case, for multiple important character data 10' corresponding to the same character type in the character data 10, if multiple voice groups 20 in the voice groups 20 correspond to the same character type as the important character data 10', but there is a protagonist voice group 20* in the voice groups 20 that has been used to determine the matching voice data corresponding to the main character data 10*, then the processing unit 11 does not select the matching voice data to be corresponded to the important character data 10' from the protagonist voice group 20*. For example, the character data 10D and the character data 10E in FIG3 and the voice group 20A and the voice group 20B in FIG4 all correspond to the character type of "adult male", but the voice group 20A is the protagonist voice group 20*. In this case, the processing unit 11 will only select the matching voice data corresponding to the character data 10D and the character data 10E from the voice group 20B, and will not select the matching voice data corresponding to the character data 10D and the character data 10E from the voice group 20B. The matching voice data corresponding to the character data 10D and the character data 10E are selected from the main character voice group 20* (in this case, the voice group 20A). This ensures that the matching voice data corresponding to the character data 10D and the character data 10E, respectively, and the matching voice data corresponding to the main character data 10* are from different voice groups 20, thereby preventing the voices of important supporting characters from being too similar to those of the main character.

For the multiple secondary character data 10" corresponding to the same character type in the character data 10, if the multiple voice line groups 20 in the voice line groups 20 correspond to the same character type as the secondary character data 10", the voice line configuration system 1 selects the matching voice line data corresponding to the secondary character data 10" from the voice line groups 20 in turn. For example, the character data 10F, the character data 10G and the character data 10H in FIG3 and the voice line group 20C and the voice line group 20D in FIG4 all correspond to the character type of "adult female", and the character data 10F, the character data 10G and the character data 10H are all In the case of the secondary character data 10", the processing unit 11 will select the matching voice data corresponding to the character data 10F, the character data 10G and the character data 10H in the order in which the three secondary supporting characters first appear in the text data. For example, the processing unit 11 will first select the matching voice data corresponding to the character data 10F from the voice group 20C, then select the matching voice data corresponding to the character data 10G from the voice group 20D, and then select the matching voice data corresponding to the character data 10H from the voice group 20C, thereby ensuring that the matching voice data corresponding to the character data 10F to the character data 10H respectively do not all come from the same voice group 20.

It should be noted that for each secondary character data 10", after the processing unit 11 selects a matching voice line group corresponding to the same character type as the secondary character data 10", it may, for example, set the voice line data 201 in the matching voice line group that has not yet been matched with any other character data 10, or the voice line data 201 with the least number of corresponding other character data 10, as the matching voice line data corresponding to the secondary character data 10", but the present invention is not limited to this.

After the processing unit 11 sets the matching voice ray data corresponding to each character data 10, the process proceeds to step S4.

In step S4, for the secondary character data 10", the processing unit 11 treats each secondary character data 10" as a checked character data, and determines whether a voice conflict condition is met based on the text data for the checked character data and the paired voice data currently corresponding to the checked character data.

First, it should be noted that after step S3 is completed, the processing unit 11 has set the corresponding matching voice data for each character data 10. In this case, each line portion contained in the text data not only corresponds to one of the character data 10, but also corresponds to the matching voice data currently corresponding to one of the character data 10.

Furthermore, for the convenience of description, each of the dialogue parts corresponding to the inspected character data is referred to as an inspected dialogue part.

For the inspected character data and the currently corresponding paired voice data of the inspected character data, the voice conflict condition represents: there exists a character data 10 among the dialogue portions that meets a proximity condition and is not a neighboring dialogue portion of the inspected character data, and the paired voice data corresponding to the neighboring dialogue portion is the same as the paired voice data corresponding to the inspected character data. For each dialogue portion other than the inspected dialogue portion(s), the proximity condition represents that the number of characters between the dialogue portion and any of the inspected dialogue portions in the text data is less than or equal to a preset character count threshold (e.g., 600 characters, but not limited thereto).

For example, assuming that the character data 10F in Figure 3 is taken as the inspected character data, if the processing unit 11 determines that the voice conflict condition is met for the character data 10F and its corresponding paired voice data, it means that the number of characters between at least one dialogue portion corresponding to the character data 10F and the dialogue portion corresponding to another character data 10 (for example, the character data 10B in Figure 3) in the text data is less than or equal to the preset character number threshold, and the character data 10F and the character data 10B correspond to the same paired voice data, that is, they are configured to the same specific voice. That is, if the actual dubbing is performed using the paired voice data currently corresponding to the character data 10F and the character data 10B, the two characters corresponding to the character data 10F and the character data 10B will speak in exactly the same voice one after another in a short period of time, which may easily lead to a poor listening experience.

If the processing unit 11 determines that the sound ray conflict condition is met for any of the examined character data and its corresponding paired sound ray data, the process proceeds to step S5. On the other hand, if the processing unit 11 does not determine that the sound ray conflict condition is met for any of the examined character data and its corresponding paired sound ray data, the process proceeds to step S9.

In step S5, once the processing unit 11 determines that a voice ray conflict condition is met for any of the examined character data and its currently corresponding paired voice ray data, the processing unit 11 automatically updates the voice ray data 201 corresponding to the examined character data that causes the voice ray conflict condition to be met, that is, the paired voice ray data corresponding to the examined character data is replaced from the original voice ray data 201 to another different voice ray data 201, in an attempt to eliminate the situation where two different characters speak with the same voice in a short period of time. For example, assuming that the paired voice ray data corresponding to the examined character data is originally the voice ray data 201 marked "bass" in the voice ray group 20C in Figure 4, the processing unit 11 may, for example, replace the paired voice ray data corresponding to the examined character data with the next voice ray data 201 in the voice ray group 20C (for example, the voice ray data 201 marked "middle" in the voice ray group 20C), or replace it with any voice ray data 201 in another voice ray group 20 corresponding to the same character type (for example, the voice ray group 20D). Specifically, the processing unit 11 may, for example, replace the paired sound ray data corresponding to the examined character data according to the order of the sound ray data 201, or may, for example, replace the paired sound ray data corresponding to the examined character data in a random manner.

After the processing unit 11 updates the sound ray data 201 corresponding to the examined character data, the process proceeds to step S6.

In step S6, the processing unit 11 again determines whether the sound ray conflict condition is met for each piece of examined character data and the paired sound ray data corresponding to the examined character data. If the processing unit 11 again determines that the sound ray conflict condition is met, the process proceeds to step S7. On the other hand, if the processing unit 11 again determines that the sound ray conflict condition is not met, the process proceeds to step S9.

In step S7, the processing unit 11 determines whether the cumulative number of times it has determined that the sound ray conflict condition has been met has reached a preset conflict threshold (e.g., three times, but not limited thereto). If the determination is yes, the process proceeds to step S8. On the other hand, if the determination is no, the process restarts from step S5, for example.

In step S8, once the processing unit 11 determines that the cumulative number of times the sound ray conflict condition is met has reached the preset conflict threshold, the processing unit 11 generates a sound ray conflict prompt and outputs the sound ray conflict prompt. The sound ray conflict prompt, for example, indicates the detected character data that caused the sound ray conflict condition to be met, and the processing unit 11 outputs the sound ray conflict prompt by, for example, controlling a display device to display the sound ray conflict prompt, but the present invention is not limited thereto. It should be noted that the sound ray conflict prompt is used, for example, to prompt the user to manually determine the matching sound ray data corresponding to the detected character data, but the present invention is not limited thereto.

In step S9, the processing unit 11 generates a voice ray configuration result and outputs the voice ray configuration result. The voice ray configuration result, for example, indicates all of the character data 10 and the corresponding paired voice ray data. The processing unit 11 may output the voice ray configuration result by, for example, controlling the display device to display the voice ray configuration result and/or outputting the voice ray configuration result to the storage unit 12 for storage, but the present invention is not limited thereto. It should be noted that the voice ray configuration result can be used by the processing unit 11 to perform an actual dubbing process on the text data, thereby generating a dubbing result corresponding to the text data and capable of being played by a speaker.

The above is an example of how the sound ray configuration system 1 of this embodiment implements the sound ray configuration method.

It should be noted that steps S1 through S8 of this embodiment and the flowchart in FIG2 are merely exemplary of one possible implementation of the sound ray configuration method of the present invention. It should be understood that even if steps S1 through S8 are combined, separated, or their order is adjusted, as long as the resulting process achieves substantially the same effect as the present embodiment in a substantially identical manner, it will still fall within the scope of the sound ray configuration method of the present invention. Therefore, steps S1 through S8 of this embodiment and the flowchart in FIG2 are not intended to limit the scope of implementation of the present invention.

The present invention also provides an embodiment of a computer program product, wherein the computer program product includes an application that can be stored in a computer-readable recording medium and can be loaded and run by an electronic device (for example, but not limited to a desktop computer, laptop computer, tablet computer, smart phone or server), and when the electronic device loads and runs the application of the computer program product, the application enables the electronic device to be used as the sound ray configuration system 1 and implement the aforementioned sound ray configuration method.

In summary, by implementing this voice configuration method, the voice configuration system 1 can, when the number of voices with suitable accents and timbre is limited, use a suitable specific voice (i.e., the specific voice corresponding to the initial voice data) as a basis to expand the number of specific voices with suitable accents and timbre (i.e., the specific voice corresponding to the derived voice data), thereby increasing the number of suitable voices to choose from. Furthermore, the voice configuration system 1 helps prevent the voices of important supporting characters from being too similar to the main character, and prevents the voices of multiple important supporting characters from being too similar to each other. Furthermore, after setting the matching voice data corresponding to each character's data, the voice configuration system 1 can proactively detect whether different characters speak with the same voice in a short period of time, and attempt to eliminate these situations. Therefore, the voice ray configuration system 1 helps to achieve better dubbing effects when the number of suitable voice rays is limited, and can indeed achieve the purpose of this invention.

However, the above description is merely an example of the present invention and should not be used to limit the scope of the present invention. All simple equivalent changes and modifications made within the scope of the patent application and the contents of the patent specification of the present invention are still within the scope of the present patent.

1: Voice Configuration System 11: Processing Unit 12: Storage Unit DB: Virtual Voice Database D1: Voice Requirement Information 10, 10A-10J: Character Data 101: Character Tag 102: Voice Feature Label 10*: Primary Character Data 10': Important Character Data 10”: Secondary Character Data D2: Voice Data Group 20, 20A-20D: Voice Groups 20*: Main Character Voice Group 201: Voice Data S1-S9: Steps

Other features and functions of the present invention are clearly illustrated in the accompanying drawings, wherein: Figure 1 is a block diagram illustrating an exemplary embodiment of the sound ray configuration system of the present invention; Figure 2 is a flow chart illustrating how the embodiment implements a sound ray configuration method; Figure 3 is a diagram illustrating exemplary sound ray requirement information utilized in executing the sound ray configuration method; and Figure 4 is a diagram illustrating exemplary sound ray data combinations utilized in executing the sound ray configuration method.

S1~S9: Steps

Claims (7)

A voice configuration method is implemented by a voice configuration system comprising a processing unit and a storage unit electrically connected to the processing unit. The method comprises: (A) obtaining voice requirement information and a voice data set, wherein: The voice requirement information comprises a plurality of character data corresponding to a plurality of characters, each character data set corresponding to one of a plurality of character types; one or more of the character data sets each comprises a voice feature label associated with an audio frequency range and each is considered as an important character data set; The voice data set includes a plurality of voice groups, each voice group corresponding to one of the character types and including a plurality of voice data items. The voice data items in each voice group correspond to a plurality of specific voices, and include an initial voice data item and one or more derived voice data items generated by at least adjusting the audio range of the initial voice data item. (B) For each character data, selecting a matching voice group from the voice groups that corresponds to the same character type as the character data, and then setting one of the voice data in the matching voice group as a matching voice data corresponding to the character data, thereby assigning the specific voice corresponding to the matching voice data to the character corresponding to the character data. If the character data belongs to the important character data, after selecting the matching voice group from the voice groups, the voice configuration system sets one of the voice data in the matching voice group that matches the voice feature label of the important character data as the matching voice data corresponding to the important character data; (C) for one of the examined character data and the paired voice data currently corresponding to the examined character data, determining whether a voice conflict condition is met based on text data corresponding to the voice requirement information, wherein the text data includes a plurality of dialogue parts, each dialogue part corresponds to one of the character data and the paired voice data currently corresponding to one of the character data, and each dialogue part corresponding to the examined character data is used as an examined dialogue part, and the The voice conflict condition includes: there is a neighboring voice portion among the dialogue portions that meets a proximity condition and corresponds to character data other than the examined character data, and the matching voice data corresponding to the neighboring voice portion is identical to the matching voice data corresponding to the examined character data; wherein, for each dialogue portion other than the examined dialogue portion(s), the proximity condition indicates that the number of characters between the dialogue portion and any examined dialogue portion in the text data is less than or equal to a preset character count threshold; and (D) if the voice conflict condition is determined to be met, outputting a voice conflict prompt indicating the examined character data. As described in claim 1, the voice configuration method, wherein in step (B), for multiple important character data corresponding to the same character type in the character data, if multiple voice groupings in the voice groupings correspond to the same character type as the important character data, then the voice configuration system selects the matching voice data corresponding to the important character data respectively from the voice groupings in turn. The voice configuration method of claim 1, wherein: In step (A), at least one of the character data is designated as a main character data, and the voice group in the voice groups used to determine the matching voice data corresponding to the main character data is designated as a protagonist voice group; and In step (B), for multiple important character data in the character data corresponding to the same character type, if multiple voice groups in the voice groups correspond to the same character type as the important character data, but the protagonist voice group exists in the voice groups, the voice configuration system does not select the matching voice data from the protagonist voice group to correspond to the important character data. The voice configuration method of claim 1, wherein: In step (A), one or more pieces of character data in the character data that do not include a voice feature label are each treated as a secondary character data; In step (B), for multiple pieces of secondary character data in the character data corresponding to the same character type, if multiple voice groups in the voice groups correspond to the same character type as the secondary character data, the voice configuration system selects the paired voice data corresponding to the secondary character data from the voice groups in turn. The sound ray configuration method as described in claim 1, wherein, in step (D), when it is determined that the sound ray conflict condition is met, the sound ray configuration system first updates the sound ray data corresponding to the inspected character data, and only outputs the sound ray conflict prompt after the sound ray data corresponding to the inspected character data is updated and it is again determined that the sound ray conflict condition is met. A sound ray configuration system includes a processing unit and a storage unit electrically connected to the processing unit, wherein the sound ray configuration system is configured to implement the sound ray configuration method as described in any one of claims 1 to 5. A computer program product includes an application program, wherein when the application program is loaded and executed by an electronic device, the electronic device can implement the sound ray configuration method as described in any one of claims 1 to 5.