KR101301148B1 - Song selection method using voice recognition - Google Patents

Abstract

PURPOSE: A song selection method using a voice recognition is provided to improve a rate of a voice recognition in a karaoke machine by using a structured control command according to a state of the karaoke machine. CONSTITUTION: A song selection method using a voice recognition comprises a song item display step (115), a voice data and control command comparing step (S117,S119), and a song item selecting step (S121). The song item selecting step aims to display multiple song items corresponding to multiple searched songs. The voice data and control command comparing step aims to receive an audio signal to select the song item, and to compare a voice data of the received audio signal with multiple designated control commands. The song item selecting step aims to correspond with the control command and to select the song item. [Reference numerals] (S100) Start; (S101) Receive a voice signal to select a song by voice recognition; (S103) Do the voice data of the voice signal correspond to the designated control command?; (S105) Receive the voice signal for searching for songs; (S107) Search for the songs corresponding to song search DB by using voice data of the voice signal; (S109) Fail to search songs or to recognize the voice data?; (S111) Order the searched songs; (S113) Generate a song item for the searched songs made in order; (S115) Display the generated song item; (S117) Receive the voice signal to control the displayed song item; (S119) Control command for the voice data of the voice signal to select the song item?; (S121) Select the song item corresponding to the determined control command; (S123) Perform other relevant control command; (S200) End

Description

Song selection method using speech recognition {SONG SELECTION METHOD USING VOICE RECOGNITION}

The present invention relates to a song selection method using voice recognition, specifically, a song selection method using voice recognition, which enables the user to search for a song and conveniently selects a song in response to a voice signal received through a microphone. It is about.

As the performance of the processor is improved, a speech recognition technology capable of converting analog speech into digital speech data and recognizing corresponding words or sentences therefrom may be practically used in many technical fields. For example, such a voice recognition technology is mounted on a smart phone or the like to extract words or sentences from the user's voice.

However, the speech recognition technology has a limitation in finding words that have low word recognition rate or exact correspondence, and thus, there is a limitation in applying them to commercial products requiring reliability or accuracy.

For example, a song accompaniment device is an environment in which not only a user's voice but also a music signal is output through a speaker. In such an environment, it is necessary to consider not only a user's voice signal but also a music signal, so that the recognition rate by the voice recognition technology is lowered.

In addition, when a song accompaniment device searches for a song using a specific word, there may be a plurality of songs corresponding to or similar to the word, thus navigating the results of the found songs, but in a karaoke room Due to the deterioration of the recognition rate according to the environment, the voice recognition technology is applied to the voice search results again by using the remote controller.

Therefore, there is a need for a song selection method using voice recognition, which can provide reliability and accuracy on a commercially available product such as a song accompaniment device, and further facilitate search and selection of songs.

The present invention has been made to solve the above-described problems, and provides a song selection method using speech recognition to increase the speech recognition rate in the song accompaniment apparatus using a control command structured according to the state of the song accompaniment apparatus. Its purpose is to.

Another object of the present invention is to provide a song selection method using voice recognition, which enables a song accompaniment device to directly select desired songs through voice recognition.

Another object of the present invention is to provide a song selection method using voice recognition, which enables a user to search and select a song according to voice recognition through an intuitive and convenient user interface.

Another object of the present invention is to provide a song selection method using speech recognition, which can provide a higher speech recognition rate by analyzing a form of a word to be used for speech recognition as a control command and applying the same.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention, unless further departing from the spirit and scope of the invention as defined by the appended claims. It will be possible.

In order to achieve the above object, a song selection method using voice recognition includes: (a) displaying a plurality of song items corresponding to a plurality of retrieved songs, and (b) one of the plurality of song items. Receiving a voice signal to select a, comparing the voice data of the received voice signal with a plurality of specified control commands, and (c) corresponding to the control command determined according to the result of the comparison and being one of the plurality of song items. Selecting a song item, wherein each of the plurality of control commands corresponds to each of the plurality of songs.

In addition, in order to achieve the above object, the song selection method using speech recognition, step (a) comprises the step of highlighting a specified song item of the plurality of song items, step (c), Highlighting the selected song item and unhighlighting the designated song item, wherein one or more of the control commands of the plurality of control commands are included in the song item of the corresponding song and displayed; Another song item among a plurality of song items can be directly selected through voice recognition.

The song selection method using voice recognition according to the present invention as described above has the effect of increasing the speech recognition rate in the song accompaniment apparatus by using a control command structured according to the state of the song accompaniment apparatus.

In addition, the song selection method using the voice recognition according to the present invention as described above, the song accompaniment apparatus has the effect that the desired song can be selected directly through the voice recognition through the music recognition.

In addition, the song selection method using the voice recognition according to the present invention as described above, there is an effect that the user can search and select the song according to the voice recognition through an intuitive and convenient user interface.

In addition, the song selection method using the voice recognition according to the present invention as described above, there is an effect to provide a higher speech recognition rate by analyzing the form of the words to be used for speech recognition as a control command and apply it.

The effects obtained by the present invention are not limited to the above-mentioned effects, and other effects not mentioned can be clearly understood by those skilled in the art from the following description will be.

1 is a block diagram illustrating a system for performing a song selection method according to the present invention.
Fig. 2 is a diagram showing an exemplary hardware block diagram of the song accompaniment apparatus.
FIG. 3 is a diagram illustrating a high level control flow of a song selection method using speech recognition performed in a song accompaniment apparatus. Referring to FIG.
FIG. 4 is a diagram showing a detailed control flow incorporating the higher level control flow of FIG. 3.
FIG. 5 is a diagram illustrating a result displayed through the display device in step S115 or step S121 of FIG. 4.

The above and other objects, features and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings, in which: It can be easily carried out. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram illustrating a system for performing a song selection method according to the present invention. According to FIG. 1, the system includes a song accompaniment device 100, a display device 200, and one or more microphones 300.

Referring briefly to each device of the present system, the song accompaniment apparatus 100 reproduces audio data of a song selected according to a song selection by a user using the song accompaniment apparatus 100, outputs it as an audio signal, and is selected. The lyrics or video corresponding to the song are output to the display device 200 or the like.

In addition, the song accompaniment apparatus 100 receives a voice signal from the microphone 300 and mounts it in software and / or hardware form on the song accompaniment apparatus 100 according to a requested control command according to voice data of the received voice signal. The voice search engine is driven to search for corresponding songs in the plurality of songs, or the voice command recognition engine is driven to perform a specified control command according to the state of the song accompaniment apparatus 100 or the state on the voice search.

Such a voice search engine and a voice command recognition engine are preferably mounted on the song accompaniment apparatus 100 and further implemented in software form.

A more detailed description of the song accompaniment apparatus 100 will be described with reference to FIGS. 2 to 5.

The display device 200 may be connected to the song accompaniment apparatus 100 by wire or wirelessly, for example, to receive a video signal output from the song accompaniment apparatus 100 and output the received video signal through a display panel. Device.

The display device 200 includes an LCD panel, a PDP panel, an LED panel, or the like so that a user who uses the song accompaniment apparatus 100 can view the lyrics, provide the user with a guide for speech recognition, or play a song. To display the background image.

The microphone 300 is connected to the song accompaniment apparatus 100 by wire or wirelessly, and converts the voice signal received from the user into an analog voice signal to the song accompaniment apparatus 100 or a digital voice signal converted from the analog voice signal. Output

The voice signal from the microphone 300 enables the song accompaniment apparatus 100 to select a song through voice recognition and / or output a user's voice to a speaker (not shown).

2 is a diagram illustrating an exemplary hardware block diagram of the song accompaniment apparatus 100.

According to FIG. 2, the song accompaniment apparatus 100 includes an input interface 110, an output interface 120, a memory 130, a hard disk 140, a microphone receiving terminal 150, and an analog-to-digital converter 160. And a processor 170 and a system bus / control bus 180.

This exemplary song accompaniment apparatus 100 may be configured to omit some of the blocks shown as needed, or may be configured to further include other blocks not shown.

2, the input interface 110 is an interface for receiving an input from a user using the song accompaniment apparatus 100. The input interface 110 includes, for example, one or more remote control receiving modules and one or more buttons. This input interface 110, for example, allows a user to select a song, play a song, or cancel playback of a song.

The output interface 120 includes a wired or wireless audio output terminal and a video output terminal. The output interface 120 converts an audio signal converted from MIDI data or compressed audio data of a song being played by the song accompaniment apparatus 100 into an audio output terminal. The video signal is output to the display device 200 through the video output terminal through the video output terminal.

The memory 130 includes a volatile memory and / or a nonvolatile memory. The memory 130 may temporarily store the MIDI data, the compressed audio data, or the voice data of the voice signal used for voice recognition under the control of the processor 170, and furthermore, the initial data necessary for booting the processor 170. You can save the program.

The hard disk 140 provides a large storage space to store various data and various programs. Such data and programs may be used by the processor 170.

For example, the hard disk 140 includes a plurality of video files and stores a plurality of audio files. Such a moving picture file may be used as a background image that may be used at the time of playback of a song, and an audio file such as a compressed audio file and / or a MIDI file may be used at the time of playback of a song and output as an audio signal.

In addition, the hard disk 140 including a plurality of programs may be driven independently or under control of a control engine, a control engine for controlling the song accompaniment apparatus 100 according to an input received through the input interface 110. A voice engine, a playback engine for decoding an audio file or a video file according to a specified format under the control of a control engine, and an interface engine for controlling a hardware block included in the song accompaniment apparatus 100 may be included in the form of a software program. Can be.

The hard disk 140 may include a song search DB for searching for a song. This song search DB can output the result according to the input of the search parameters (e.g., song number, song name, singer, composer, lyrics, etc.) by the control engine or by the voice engine. This search result may include one or more parameters.

In addition, the hard disk 140 may include a popularity list for songs used in the song accompaniment apparatus 100. Such a popularity list may include, for example, a list of popularity rankings at a given reference point for all songs, or a list of popularity rankings by genre (pop, dance, ballad, pop), and the like. For example, the title of the song or the song number. This popularity list can then be used to order the search results.

Here, the one or more parameters included in the search result may include, for example, a song number, and further include a search parameter and other parameters associated with the search parameter or required to recognize the song to the user.

Here, the voice engine may also convert voice data from the voice signal received from the microphone 300 into input data of a parameter to be used in the song search DB and use the converted input data to select one or more corresponding songs in the song search DB. And a voice command recognition engine for recognizing a word used as a control command in voice recognition from the received voice signal.

Although each voice search engine and voice command recognition engine have been described in a separate form, the present invention is not limited thereto and may be configured as one engine or two or more separate engines.

And the voice command recognition engine is configured to limit the control commands to be recognized in each state or each control according to the state of the song accompaniment apparatus 100 or according to the control of the control engine or the state such as the step in the voice recognition. Can be. The speech recognition rate can be increased by limiting the control commands to be recognized in each of these states or controls.

Each of these engines may be loaded and driven by the memory 170 by the processor 170.

A song selection method using voice recognition included in the song accompaniment apparatus 100 and performed by a speech engine (in conjunction with a control engine) will be described in more detail with reference to FIGS. 3 to 5.

The microphone receiving terminal 150 receives a voice signal from the microphone 300 connected by wire or wirelessly. The microphone receiving terminal 150 includes a connection terminal for connecting to a wired cable of the microphone 300 or an antenna for receiving wireless voice data.

The analog-to-digital converter 160 converts an analog audio signal from the microphone 300 connected by wire into a series of continuous digital audio data and outputs the same to the processor 170. The analog-to-digital converter 160 converts the digital data of a specified size (for example, 8 bits, 16 bits, etc.) according to a designated sampling period and according to a promised internal data communication format (for example, I2S or SPDIF). This digital data is output to the processor 170.

If the microphone receiving terminal 150 wirelessly receives the voice data of the microphone, the analog-to-digital converter 160 may be omitted, in which case the processor 170 receives the wireless data packet to receive the wireless data packet. Digital audio data can be extracted from the data.

The processor 170 includes one or more execution units that can load and execute instructions of a program to control each hardware block according to execution of the instructions of the program.

The processor 170 may, for example, assign a program such as a control engine and a speech engine to time division or one or more execution units so that the control engine and the speech engine may be executed simultaneously or may be recognized by the programmer or the user as being simultaneously executed. Allows data to be sent and received between engines according to the program's promised protocols.

A song selection method using voice recognition loaded and executed by the processor 170 will be described in detail with reference to FIGS. 3 to 5.

The system bus / control bus 180 may transmit data or control signals generated in each block to other blocks. This system bus / control bus 180 may be a parallel bus or a serial bus.

FIG. 3 is a diagram illustrating a high level control flow of a song selection method using voice recognition performed by the song accompaniment apparatus 100.

This song selection method is preferably performed by a program driven in the processor 170, for example, the speech engine and the control engine may be performed in conjunction.

Referring to the high-level control flow according to Fig. 3, first, in step S1, the voice engine, specifically the voice command recognition engine, the designated control command for performing this song search by voice and the converted voice data of the received voice signal are If the voice data matches the specified control command or corresponds to more than a certain similarity according to the similarity, the process shifts to step S2.

The designated control command here may be a word for requesting to start searching for a song through voice, for example, a word such as "voice search", "speech recognition", "gold karaoke", and the like.

Thus, in the entry step (step S1), only the designated control command, i.e., the designated word, needs to be recognized, so that the recognition rate according to the speech recognition can be increased by comparison with the specified word, and further, such a voice search can be performed only in a specific state. So that unnecessary speech recognition can be omitted.

Here, the control engine for controlling the song accompaniment apparatus 100 determines a state capable of allowing a voice search, and if the voice search is allowed, drives the voice engine, specifically the voice command recognition engine, or operates the voice engine. You can ask to start speech recognition.

For example, the song accompaniment apparatus 100 may include a state in which a song or song is being played, a standby state in which the song or song is stopped, and waiting to receive user input. There may be a state performing a specific function.

Such a state may be determined by the control engine to control the speech engine to allow such a voice search in one or more of the specified states.

In step S2 (input step), the voice engine receives voice data of a voice signal of a parameter (for example, a song name) that the user wants to search. This step S2 is, for example, the voice data of the word or sentence that the user wants to search by searching for the end of the received voice signal (for example, using the presence or absence of the voice data for a certain period) in the continuously received voice signal. It is configured to separate.

 Subsequently, in step S3 (recognition step), the corresponding words or sentences are recognized through analysis of the speech data separated by the speech data (by the voice engine, specifically, the voice search engine), and the corresponding parameters of the song search DB are recognized. Search for similar or matching songs by comparing with.

If in step S3 the corresponding words or sentences are not recognized through analysis of the voice data and / or one or more corresponding songs are not found, the process may transition to step S2.

One or more songs are searched through this search, and the songs searched in step S4 (result display step) are generated as a result and displayed on the display device 200. The generation of these results may be by a speech engine or by a control engine that receives the results from the speech engine.

Here, in the result display step S4, a control command may be received according to another voice recognition, and the display of the result may be changed according to the received control command. The words of the control command used in the result display step S4 may be configured to be different from the words of the control command used in the entry step S1.

Accordingly, in speech recognition, only words of different control commands are recognized according to each step (or state), thereby improving the speech recognition rate.

Hereinafter, the high level control flow of FIG. 3 will be described in more detail with reference to FIG. 4.

FIG. 4 is a diagram showing a detailed control flow incorporating the higher level control flow of FIG. 3. This control flow may be performed through the control of a hardware block by a voice engine or a control engine driven on the processor 170 of the song accompaniment apparatus 100 and by the driving of a program on the processor 170.

Hereinafter, although an exemplary voice search engine and a voice command recognition engine are described, various modification configurations exist.

First, in step S100, the present control flow starts automatically after booting in response to power-on, or according to control by the control engine. If starting by the control engine, the control engine may start by driving the voice engine according to the state of the song accompaniment apparatus 100 recognized by the control engine or by transmitting a signal to a voice engine that is already driven. .

Thereafter, in step S101, a voice signal for song selection is received through the microphone receiving terminal 150 through voice recognition by the voice engine. When the received voice signal is received through the wire, the voice signal is converted into digital voice data through the analog-to-digital converter 160, and the converted series of digital voice data is transferred to the processor 170. Is transmitted to the voice command recognition engine.

The series of voice data then identifies the end of the voice data for comparison with a specified word that is a control command. For example, the voice command recognition engine of the speech engine may identify this end, and the digital magnitude value of the voice data may be equal to or less than a certain threshold by using the digital magnitude value of the voice data. If it persists, the end can be identified. This end identification can be made in the frequency domain or in the time domain.

Thereafter, in step S103, it is compared whether the voice data received and identified through the analog-to-digital converter 160 corresponds to the designated control command.

For example, such designated control command may be a designated word for starting or starting a voice search engine, and such word may be a word such as "speech search", "speech recognition", "gold karaoke", or the like.

The comparison with the designated word does not need to compare with other words, thereby increasing the speech recognition rate for entering the voice search.

Such a step S103 may be performed using a known speech recognition technique, for example, a feature parameter (or a feature vector) at a frequency value converted by converting a feature parameter or identified speech data extracted by linear prediction into a frequency domain. And determine the words by linear prediction or comparison of feature parameters on the frequency domain or by determining words from the linear prediction or feature parameters on the frequency domain according to the identified speech data. It may also consist of a comparison of a word with a word of a specified control command.

In step S103, if the voice data and the designated control command have the same or similar words (for example, through word determination) or have similarities (for example, through frequency comparison or linear prediction, etc.) within a predetermined threshold range, the process goes to step S105. Otherwise, the process proceeds to step S101.

Subsequently, in step S105, the voice search engine is driven to receive a voice signal for song search through the microphone receiving terminal 150, convert the received voice signal into digital voice data, and deliver the voice signal to the voice search engine. And the voice search engine identifies the end for distinguishing a word or sentence from the received series of voice data.

Subsequently, in step S107, the voice search engine searches for a corresponding song using the voice data identified in the song search DB stored in the hard disk 140.

The retrieved song may include one or more songs and generally may include a plurality of songs. According to such a search, the voice search engine may obtain a plurality of parameters (song name, song number, singer, composer, lyrics, release date, etc.) for each of the retrieved songs from the song search DB.

As in step S103, in step S107 the feature parameters of the identified voice data and the feature parameters of the parameters for retrieving the song search DB (e.g., Linear Prediction Cepstral Coefficient (LPCC) or Mel Frequency Cepstral Coefficient (MFCC), etc.) It can be made by comparison.

Songs searched through this comparison have similarities within a certain range with voice data for song search.

Then, in step S109, the voice search engine determines whether the music search has failed or the word recognition from the voice data, etc. has failed, and if it fails, the process shifts to step S105 and, if successful, to step S111.

Then, in step S111, the voice search engine orders the retrieved songs.

For example, the voice search engine compares the searched songs with a stored popularity list and if a high ranking song exists in the searched songs according to the ranking in the popularity list, the high ranking song has a higher rank on the searched songs. Can be ordered. This popularity list is a list representing the popularity ranking for all songs and / or the popularity ranking by genre. And it can be configured to give priority to the popularity ranking of the entire song rather than the popularity ranking by genre.

Alternatively, the voice search engine may display on the display device 200 such that the songs most similar to the identified voice data are listed first according to the similarity with the identified voice data, and then the similar songs are listed and displayed second. You can also order them to be.

For the similarity determination here, the highest weight is placed on the first character of the speech data and the next weight is placed on the next character between the word of the identified speech data and the word of the specified parameter of the retrieved songs. Similarity can be determined by summing the similarity between them according to this weight.

Alternatively, the voice search engine may order the searched songs in a combination according to the identified voice data and the popularity ranking by the popularity list. For example, songs that are in the popularity list are ranked according to popularity rankings, and songs that are not in the popularity list are ordered in a lower rank than songs that are in the popularity list with similarity to the identified voice data. can do.

Subsequently, in step S113, the control engine that receives the parameters for the ordered searched songs from the voice search engine or the voice search engine generates the song items to be displayed on the display device 200 for the ordered songs.

Each of the song items includes one or more parameter values included in the song search DB, and includes, for example, values of parameters such as a song title, a song number, a singer, and the like to be displayed on the display device 200.

Further, in step S113, the song items are generated such that a corresponding control command for selecting each song item is included in each song item and then displayed on the display device 200. This control command consists of, for example, words that enable voice recognition to directly select the corresponding song items.

Each control command, for example each word, may be used to randomly select one song item on an image displayed on the display device 200 and display the image device 200 as one image. When marked with, they can be distinguished from each other.

Accordingly, each of these control commands may be displayed in one-to-one mapping to the song titles or song numbers that are matched to the song title or song number and ordered according to the song order. Accordingly, the user of the song accompaniment device 100 The utterance of the matched control command through the microphone 300 allows the corresponding song item to be selected at random.

In operation S115, the voice search engine or the control engine displays the generated song items on the display device 200. These song items allow the user to identify the retrieved songs through voice recognition and further intuitively know the control commands corresponding to the retrieved songs.

The song items displayed here may be song items of some of the entire song items created by step S113. For example, 8 song items may be displayed on one screen of the display device 200 and 16 song items. In this case, the top eight song items having a high virtual similarity among the sixteen song items may be displayed.

FIG. 5 is a diagram illustrating a result displayed through the display device 200 in step S115 (or step S121) of FIG. 4.

As can be seen in FIG. 5, there are shown song items for words determined through speech recognition (step S107) (" forsythia " in FIG. 5) and songs retrieved according to similarity therefrom.

For example, eight song items are displayed on one screen, and each song item has a corresponding control command ("green", "red", "pink", "frequent", "purple", "blue", Words such as "black") so that the user can know the corresponding control commands visually and intuitively.

Furthermore, in addition to control commands for selecting a song item, other control commands may also be displayed. Control commands and other control commands corresponding to these song items are control commands that can be used on this displayed search result screen.

The control command corresponding to the song item may be, for example, a number such as "1", "2", "3", "4", or the like, "one", "two", etc., indicating the order. And such a control command preferably needs to be configured so that the control command for selecting each song item can be easily distinguished by speech recognition.

Accordingly, it is preferable to configure the words of the control command to be a combination of phonemes with a large discriminating power. In Korean, the length of the vowel sound is longer than the consonant. Increase recognition rates and increase rejections for words that don't need to be recognized.

For example, as shown in FIG. 5, combinations of vowels of the same size ("ㅕ TT", "TT", "ㅏㅏ", "TT", "ㅏ TT", "ㅗㅏ", etc.) are different from each other. In addition, the same vowel combination needs to be configured to be advantageous in speech recognition by different consonants of the first letter or the consonant of the first letter, or different combinations of the consonant or the consonant of the first letter.

In this way, the vowels are differentiated in two or more letters, and the control command is composed of consonants having different frequency characteristics even for the same vowel combination. It is a main part in the recognition of the vowel, and thus, by constituting two or more vowel parts, the vowel part can be easily identified, and further, the same vowel combination can provide a combination of different consonants to increase the speech recognition rate.

This word structure allows the phoneme to be composed of a phoneme having a large discrimination ability with each other, thereby increasing the recognition rate of the words. In addition, the speech recognition rate can be increased by limiting control commands (or words) that can be recognized in a specific display state.

The voice command recognition engine (or voice search engine) driven by the song accompaniment apparatus 100 may be, for example, a phoneme combination type isolated word recognition engine using a Hidden Markov Model (HMM), and the isolated word recognition engine may be a designated target word. Other than that, it can be configured not to recognize.

The phoneme combination isolated word recognition engine using the HMM extracts feature parameters or feature vectors (eg, LPCC, MFCC, or other variations) of the phonemes of the identified speech data, and extracts the feature parameters of the phonemes of the words of the designated control command. By comparing the parameters or feature vectors, one can determine the similarity between the words pronounced by the user and the words of one or more control commands that can be used in the current state, thereby determining or rejecting the words of the corresponding control command. Can be.

Also, in the search result displayed in step S115, one designated song item (the first song item in the first display) may be highlighted (pink in FIG. 5), and other control such as "start" the highlighted song item in this way. By inputting a command by voice, playback of a song can be started.

Thereafter, in step S117, the voice signal is received through the microphone receiving terminal 150. The received voice signal is then converted into digital voice data through the analog-to-digital converter 160. The received voice data may be a control command for selecting one of the song items displayed in step S115. The voice data may be separated to identify the end of the voice data for word recognition.

Thereafter, in step S119, the voice command recognition engine compares voice data of the received voice signal with a plurality of control commands and other control commands corresponding to each song item, and if determined as a control command corresponding to the song item, Transition to S121, otherwise, shift to step S123.

The speech recognition method in step S119 can be made similarly to step S103 or step S107.

Subsequently, in step S121, the voice command recognition engine selects a song item corresponding to the determined control command, specifically, unhighlights the previously selected song item (“green” in FIG. 5) and highlights the selected song item with a specified color. do.

The control command corresponding to the displayed song item is displayed on the same line as the corresponding song number or song title, allowing intuitive selection of a specific song item, and sequential selection (up and down) through voice recognition or selection using the remote controller. Improved speech recognition allows for direct selection of other song items.

In step S123, it is determined whether the control command recognized from the voice data is the other control command, and in the case of the other control command, a function corresponding to the other control command can be performed by the voice command recognition engine or by requesting the control engine. have. After the execution of these other control commands, it may be terminated (S200) or transitioned to another step (for example, S105 or S117) depending on the type of the control command.

By such a control flow of FIG. 4, it is possible to increase the voice recognition rate and to randomly select a random song, thereby providing a more functionally improved and reliable song accompaniment apparatus 100.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. The present invention is not limited to the drawings.

100: song accompaniment device
110: input interface 120: output interface
130: memory 140: hard disk
150: microphone receiving terminal 160: analog-to-digital converter
170: processor 180: system bus / control bus
200: display device
300: microphone

Claims (6)

Song selection method using speech recognition,
(a) displaying a plurality of song items corresponding to each of the found plurality of songs;
(b) receiving a voice signal to select one of the plurality of song items, and comparing voice data of the received voice signal with a plurality of designated control commands; And
(c) selecting a song item corresponding to a control command determined according to a result of the comparison and being one of the plurality of song items;
Each of the plurality of control commands is displayed in step (a) as two or more Hangul words having the same number of characters in the song item of the corresponding song,
The first letter and the second letter of the Hangul word of each of the plurality of control commands are different from the second letter corresponding to the corresponding first letter of the Hangul word of the other control command, and the first letters of the first letter of the Hangul word. The combination of the initial and neutral is configured to be different from the combination of the first and neutral of the second letter,
How to choose a song. The method of claim 1,
The step (a) includes: highlighting a specified song item among the plurality of song items;
Step (c) may include highlighting a selected song item; And releasing highlights of the designated song item.
One or more control commands of the plurality of control commands may be directly selected through voice recognition from another selected song item to a different song item.
How to choose a song. The method of claim 2,
Prior to said step (a), receiving a voice signal and comparing the voice data of the received voice signal with a control command different from said plurality of control commands; Receiving a voice signal for song search when the voice data corresponds to the different control command according to the comparison result; And searching for a plurality of songs using the voice data of the voice signal received from the song search DB.
How to choose a song. The method of claim 3,
After retrieving the plurality of songs, ordering the retrieved plurality of songs according to a stored popularity list; And generating the plurality of song items including a control command corresponding to a song title of each of the ordered plurality of songs.
How to choose a song. 5. The method of claim 4,
In the generating of the plurality of song items, each of the plurality of control commands is matched with a song title according to the order of the plurality of songs,
How to choose a song. The method of claim 1,
The song selection method is performed in a song accompaniment device,
The Hangul words of each of the plurality of control commands are Hangul words representing different colors.
The background image of the Korean word displayed in the step (a) or the Korean word displayed and included in the song item in the step (a) is represented by a color corresponding to the Korean word.
How to choose a song.

Images