201025301 六、發明說明： 【發明所屬之技術領域】 本發明係有關於一種語音點唱系統，為一種藉由聲音 接爻單接收語音訊號，將此語音訊號與資料庫進行比 辨識出相符的歌曲，並判斷語音訊號之音頻高低獲 侍一參考音調值，及根據該語音訊號搜尋所欲演唱之歌 曲，再根據參考音調值之音調播放所欲演唱之歌曲。 【先前技術】201025301 VI. Description of the Invention: [Technical Field] The present invention relates to a voice point-singing system, which is a voice receiving signal received by a sound interface, and the voice signal is compared with a database to identify a matching song. And determining the audio level of the voice signal is given a reference tone value, and searching for the song to be sung according to the voice signal, and then playing the song to be sung according to the pitch of the reference tone value. [Prior Art]
按，目前之伴唱機具有下列之缺點： 1.手動點歌，速度慢。 2·若不知道曲名尋找歌曲不容易。 3·如果只憑部分歌詞’無法找到所要點播的歌曲。 也升調（升key)或降調（降key)需要手動操作 者也常因聲調不對而出糗。 、目别有中華民國專利公報公告第287347號「音效控制 ◎ 2、控制裝置及利用該等之伴唱裝置」，其主要方法為， 身訊，並且分別預先預定將對於與上述音響資訊 1土=: 一種不同的音響資訊所實施的音效的種類，將 二二音r制用資料，並附加與上述音響資 之際，亦回1 存放起來’當進行讀出上述音響資訊 用資料，'對皮附加與之關連性的各個上述音效控制 設定的音效、34至少—個不同的音響資訊分別實施預先 其目的在於提供， 對於一般的音樂資訊可容易且柔軟 3 201025301 , * 地設定所期的音效内容，以自 方法、控制裝置及利用該等之伴效操作之音效控制 置」，：二國= 據及音節波形數據組合對曰應:奏 =夂:依據旋律數 重疊於伴唱演奏裝置之伴唱演奏。歌唱音， 合’輸出開關即可隨時輸出 唱音組 ❹ 曲發生唱損部分時，因能以組合=中之歌 故可不變壓力享受伴唱。 ㈠線補其°曰知部分， 上述習用之伴唱裝置及I方 多’但對於前述之缺點仍點唱機的專利相當 【發明内容】 道曲名用讀决昔日伴唱機手動點歌速度慢、不知 ❹要手動：：曲及升調（升key)或降調（降key)需 要手動操作之問題者。 馬 因此本發明係提供—種語音點m =A.對聲音接受單元輸人-語音訊號；B.將上述語= D判It!料庫進行比對;C.找出符合語音訊號之歌曲： 春曰訊號之音頻高低，獲得一參考音調值；e.根據 參考θ調值之音調播放步驟c之歌曲。 上述之歌曲資料庫係儲存於一儲存單元。 上述之聲音接收單元係為麥克風。 201025301 η士ίΓ?驟β之比對’當聲音接收單元接收語音訊號後， =比對單元連結歌曲f料庫搜尋符合語音訊號之 =及經由—音調韻單元_該語音减之音頻高低 者。 料*上述之步驟D後將該參考音調值記錄下來，並將之 對應於步驟A之語音訊號，以後當步驟a再次接受到符合 的扣曰訊號後’直接輸出該音調值以配合所點播之歌曲。 ❹ 本發明之具體特點與功效在於： + 1.利用語音辨識功能進行點歌，可以歌曲名、歌曲之 段落、演唱者、…等進行語音輸入，點歌速度快。 2.可辨識歌唱者之音調’自動調整成符合歌唱者音 調’演唱時可避免走音出糗。 音調3.可，己憶點歌唱者的音調’點歌後自動調整最佳伴唱 4.唱一段歌詞也可點歌’以免忘記歌名而有遺珠之憾。 ❹ 5.點歌方式人性化，人人均可操作。 6·具經濟利用價值。 【實施方式】 請參閱第-圖，係本發明之第—實施例，為一種語音 點唱系統’包括下列步驟： Α.對聲音接受單元輸入_語音訊號：該聲音接受單元 係為-麥克風，以麥克風接收歌唱者一小段歌詞。 Β.將上述語音訊號與|曲資料庫進行㈣：以歌唱者 所演唱之-小段歌詞，進入儲存單元之歌曲資料庫中搜尋 201025301 比對。 c.找出符合語音訊號之歌曲：比對出符合該段歌詞之 歌曲。 D. 判斷語音訊號之音頻高低，獲得一參考音調值：並 針對歌唱者所演唱歌詞之音調進行調整。 E. 根據參考音調值之音調播放步驟C之歌曲：使該歌 曲符合歌唱者之音調。 上述步驟B之比對，係藉由一聲音接收單元接收語音 ®訊號，經由一語詞比對單元連結歌曲資料庫搜尋符合語音 訊號之歌曲，及，經由一音調辨識單元判斷該語音訊號之 .音頻高低。 上述之步驟D後，將該參考音調值記錄下來，並將之 對應於步驟A之語音訊號，以後當步驟A再次接受到符合 的語音訊號後，直接輸出該音調值以配合所點播之歌曲。 請再參閱第一圖，係本發明之第二實施例，為一種語 0音點唱系統，包括下列步驟： A. 對聲音接受單元輸入一語音訊號：該聲音接受單元 係為一麥克風，以麥克風接收歌唱者所講之歌曲名、演唱 者、作曲者、…等。 B. 將上述語音訊號與歌曲資料庫進行比對：將辨識比 對後的歌曲進行播放；若比對後產生多筆可能之結果，則 顯示於螢幕進行選擇再播放。 C. 找出符合語音訊號之歌曲：接收演唱之訊號，比對 出符合輸入之歌曲。 201025301 D. 判斷語音訊號之音頻高低，獲得一參考音調值：並 針對歌唱者所演唱歌詞之音調進行調整，使該歌曲符合歌 唱者之音調。 E. 根據參考音調值之音調播放步驟C之歌曲：使該歌 曲符合歌唱者之音調。 上述步驟B之比對，係藉由一聲音接收單元接收語音 訊號，經由一語詞比對單元連結歌曲資料庫搜尋符合語音 訊號之歌曲。 ® 綜合上述，本發明係針對利用語音點唱系統之應用技 術，使歌唱者經由聲音接收單元預先演唱一小段足可供進 . 行比對之歌詞，經由語詞比對單元連結歌曲資料庫搜尋符 合語音訊號之歌曲，再經由音調辨識單元判斷該語音訊號 之音頻高低，而進行調整，並可記憶歌唱者之音調，之後 即可辨識因應調整，也可用語音輸入歌曲名等進行辨識比 對後得到所要演唱的歌曲，使點歌方式更為人性化。 0【圖式簡單說明】 第一圖：本發明之步驟圖。 【主要元件符號說明】 無 7Press, the current phonograph has the following shortcomings: 1. Manual song, slow. 2. It is not easy to find a song without knowing the name of the song. 3. If you only rely on some of the lyrics, you can't find the song you ordered. It is also necessary to manually adjust the (up key) or down (key) to the manual operator. The purpose of the Republic of China Patent Gazette Notice No. 287347 "Sound Effect Control ◎ 2, control device and the use of these singer devices", the main method is, the body, and each predetermined advance will be for the above-mentioned audio information 1 soil = : The type of sound effect implemented by a different kind of audio information, and the data of the second and second sounds is added to the audio resource, and is also stored back to 1 'when reading the audio information, 'for the skin The sound effects of each of the above-mentioned sound effect control settings, 34 at least one different audio information are respectively implemented in advance for the purpose of providing, and the general music information can be easily and softly 3 201025301, * setting the desired sound effect content The self-method, the control device and the sound effect control using the associated effect operation:: two countries = according to the syllable waveform data combination pair: 夂 = 夂: according to the melody number overlapped with the vocal performance of the vocal performance . Singing sound, combined with the 'output switch, you can output the vocal group at any time. ❹ When the song is damaged, you can enjoy the vocal with constant pressure because of the combination of the song. (1) The line complements the 曰 曰 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , To manually:: the song and the rise (up key) or down (down key) need to manually operate the problem. Therefore, the present invention provides a speech point m = A. inputting a voice signal to a voice receiving unit; B. comparing the above language = D to the It! library; C. finding a song that conforms to the voice signal: The audio level of the spring signal is obtained, and a reference tone value is obtained; e. the song of step c is played according to the pitch of the reference θ value. The above song database is stored in a storage unit. The above sound receiving unit is a microphone. 201025301 η士 Γ 骤 之 之 ’ ’ 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当After the above step D, the reference tone value is recorded and corresponds to the voice signal of step A. After step a, the corresponding button signal is received again, and the tone value is directly output to match the on-demand broadcast. song.具体 The specific features and effects of the present invention are as follows: + 1. Using the voice recognition function to perform a song, the voice name can be input by the song name, the passage of the song, the singer, etc., and the song is fast. 2. The vocalist's tone can be recognized as 'automatically adjusted to match the singer's tone'. Tone 3. Yes, I have recalled the vocalist's tone. 'Automatically adjust the best vocals after the song. 4. Sing a lyric can also sing songs' so as not to forget the song name and have a regret. ❹ 5. The song mode is user-friendly and can be operated by everyone. 6.·Economic use value. [Embodiment] Please refer to the first embodiment, which is a voice-singing system of the present invention, which comprises the following steps: Α inputting to the sound receiving unit _ voice signal: the sound receiving unit is a microphone, Receive a short lyric of the singer with a microphone. Β. Perform the above-mentioned voice signal and the music library (4): sing the lyrics sung by the singer, and enter the song database of the storage unit to search for 201025301. c. Find songs that match the voice signal: Compare the songs that match the lyrics. D. Determine the audio level of the voice signal, obtain a reference tone value: and adjust the tone of the lyrics sung by the singer. E. Play the song of step C according to the pitch of the reference pitch value: make the song match the vocalist's pitch. In the above step B, the voice signal is received by a voice receiving unit, the song database is connected to the song database via a word matching unit, and the voice signal is judged by a tone recognition unit. High and low. After the above step D, the reference tone value is recorded and corresponds to the voice signal of step A. After step A receives the corresponding voice signal again, the tone value is directly output to match the clicked song. Please refer to the first figure, which is a second embodiment of the present invention, which is a speech-sounding system including the following steps: A. inputting a voice signal to the sound receiving unit: the sound receiving unit is a microphone, The microphone receives the song name, singer, composer, ..., etc. spoken by the singer. B. Compare the above voice signal with the song database: play the identified songs; if there are multiple possible results after the comparison, display them on the screen to select and play. C. Find songs that match the voice signal: receive the signal of the singing, and compare the songs that match the input. 201025301 D. Determine the audio level of the voice signal, obtain a reference tone value: and adjust the tone of the lyrics sung by the singer to make the song match the vocalist's tone. E. Play the song of step C according to the pitch of the reference pitch value: make the song match the vocalist's pitch. In the above step B, the voice signal is received by a voice receiving unit, and the song database is connected to the song database via a word comparison unit to search for the voice signal-compliant song. ® In summary, the present invention is directed to the application technology of the voice-singing system, so that the singer pre-sings a small piece of lyrics that can be used for comparison and comparison through the voice receiving unit, and searches for the song database through the word matching unit. The song of the voice signal is then judged by the tone recognition unit to determine the audio level of the voice signal, and the tone of the singer can be memorized, and then the tone can be identified, and the voice name can be used for identification and comparison. The songs to be sung make the song-song method more human. 0 [Simple description of the drawing] First figure: The step diagram of the present invention. [Main component symbol description] None 7