Μ年夂月％曰修『更.)正替換頁 1304569 九、發明說明： 【發明所屬之技術領域】 本發明係涉及一種音樂伴奏裝置之自動跟調方法以及 系統之創新設計者。 【先前技術】 按，一般人在有音樂伴奏（如伴唱機）的情形下阳 士 白'^ Β守，往往會因為伴奏音樂的音調過高或過低而失去音準， 如此當演唱者的歌聲音調跟不上伴奏音樂之音調時，便造 成歌聲與音樂聲不協調的情形，進而大幅影響歌唱的致= 聲音調而改 點，係必須 由運算獲取 音調」與一 配訊號，進 奏之自動跟 段時間週期 )，是以， )已然造成 比對所需的 知音樂伴奏 變伴奏音樂 變伴奏音樂音 測量演唱者歌 此時間週期之 匹配伴奏音樂 以轉變伴奏音 调方法，其對 之音调加以計 其每次獲得一 比對上產生明 時間，則延遲 之自動跟調方 音调以符合演Μ 夂 曰 曰 ) ) ) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 [Prior Art] Press, the average person in the case of music accompaniment (such as a phonograph), Yang Shibai '^ Β , , 往往 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , When you can't keep up with the pitch of the accompaniment music, it will cause the song and the music to be inconsistent, which will greatly affect the singing. The voice will be changed and the tone must be obtained by the operation. Segment time period), is, ) already caused the required music accompaniment to change the accompaniment music accompaniment music to measure the accompaniment song to match the accompaniment music for this time period to change the accompaniment pitch method, which is to calculate the pitch Each time a clear time is obtained on a comparison, the delay automatically matches the tone to match the performance.
夠讓音樂伴奏裝置隨著演唱者歌 調之設備；惟，其採用之技術特 聲之音調一預定時間週期，復藉 平均音調」，再透過該「平均 之芬考音調加以比較以提供_失 樂之音調；然，此種習知音樂伴 於演唱者歌聲音調之計算是取一 算而獲得一平均值（即平均音調 個平均值的時間點間隔（如5秒 顯延遲的情形，若再加上計算、 的情形將更加明顯，因此，該習 法於實際應用經驗中發現，其= 5 1304569 唱者音調的過程並無ϋϋ之 演唱者唱完-句歌詞之後，於=即時^往往必須等待 音樂音調之轉變；且上揭習 進行時才會達成 比對，實難以獲得較高準確性之=為兩個固定點數值的 應使用者最期盼的需求，實有再多调值’如此顯然不足因 。 再加以改善突破之空間存在 疋X針對上述習知音樂伴奏 ^ 、 用上所存在之問題點，如何研發出二跟调方法使 性之創新構造，實有待相=具理想實用 方向者。 力以心索大破之目標及 有鑑於此，發明人本於多年從 與設計經驗，針對上述之目標 ^品之製造開發 終得-確具實用性之本發明^…以與審慎評估後， 【發明内容】 本發明之主要目的，将尤担 置之 ，即 準的 提供 計算 器中 ，並 與主 須移 之音 人 ’、柃仏一種音樂伴奏裝 ::调方法，令人們唱歌時’能根據演唱音 且自動地調整伴奏歌曲的音調，以減 ：： :’亚適合自己的歌唱的音域；其技術特點包括曰 :调估測器’以將使用者的聲音每隔—段預 土 本頻率（Fundamental FreqUenCy);在一立 將使用者聲音的基本頻率轉為使用者；== 錄Γ藉一音階匹配器以比對使用者音階參= 律曰階序列之是異性；復利用移調判斷器決^ 及何時移調’若需移調’則自動調整音樂：：：: 1304569 參數’最後音樂人成哭、# t _.......... 卞口成益係將以數位記錄之樂譜合成實際的 二屮、本、’、、'、後與使用*歌聲一起經混音器（nuxer)"刺口八 =者4此創新獨特設計，使用者之歌聲音調計算方式 ^用：一小段時間計算-次之方式，復藉該音階序列記 4將使用者聲音的基本頻率轉為使用者音階參數序列； 亦即’本發明係比對主旋律音階序列及使用者歌聲音階序 列，所述音階匹配器係比對一段音階序列的匹配程度，屬 於動態比對音階曲線之模式，並輸出其最佳匹配時的音階 差，因為該音階匹配器是動態比對一段時間的音階序列曲 線，故所得之移調值具有較高之準確性，能夠獲得最佳之 音調調整效果，更符合使用者之所需。 又，本發明之技術特點係直接取得已記錄歌曲之主旋 律，並無需複雜的運算過程，具有較低之系統運算量，佔 用較低的系統資源，故整體具有較高之運作效率和即時性 ，可大幅改善習知系統延遲之問題點，達到更加實用好用 之進步性者。 【實施方式】 請參閱第1 、2圖所示，係本發明音樂伴奏裝置之自 動跟調方法之較佳實施例，惟此等實施例僅供說明之用，· 在專利申請上並不受此結構之限制；所述自動跟調方法係 包括： 如第1圖所示，首先將使用者的聲音每隔一小段時間 (約〇· 1秒），經由一音調估測器11計算其基本頻率（Fund 7 1304569 ?7年t月Μ曰絛(更)正替換貝 酬加F卿崎)’該音調估測器i i乃是計算此段聲音之基 本週期或頻率，它可由計算最大值之自相關函妻丈(aut〇c〇rr elation iunctlon)獲得或在頻譜上由峯值（祕％㈣之相關 位置及間距而得到，基本週期與.頻率的關係是： 基本頻率=取樣頻率/基本週期 該取樣頻率是聲音每秒取樣的樣本點數，接著在一立 階序列記錄器，將—連串使用者輸人聲音的基本料 轉為使用者音序列121，並加以記錄，音階與頻率的關 係如下： 12音階A4的頻率是44〇 Hz，每升高半個音階頻率升高 Λ^倍，每降低半個音階頻率降低%倍， 個音階頻率共升高2倍，每降低12個音階頻率亦降低=倍 ，然後，透過一音階匹配器13計算比對使用者音階序列工 21與主旋律音階序列14之差異性，在此，該主旋律音階序 列14是預先儲存於音樂的樂譜15中，例如：在(咖以。 al mstniment digital interface)檔案中即可同時記錄這些樂 譜的資訊，該音階匹配器13則是利用動態時間扭曲（Dyn· ic Tune Wannng，簡稱DTW)校準的方式，比對使用者音階序 列與主旋律音階序列14的差異性，其具體作法如下： 假設使用者音階序列121為nl，n2,…，nj，代表每隔固 疋一小段時間（如〇· 1秒）所連續計算出的使用者（或歌 者）的音階（或音調），並假設主旋律音階序列14為ml，m 2, ···，mj，代表每隔固定一小段時間（如〇. 1秒）的主旋律 曰^ ’在此音階是以數字1〜255表示，音階C3的代表數值為 6〇 ’音階D3的代表數值為61，音階β3的代表數值為％，以 8 1304569 p--— 乃年月κ曰修(更)正替換頁 • 此類推’因為在唱歌時，歌唱者的聲音拍點位置不見得會 與月景音樂播放的拍點位置相同，所以在音階比對時必須 做動悲日寸間校準，以產生較正確的比對結果，如下示意圖 nl n2 n3 n4 n5 n6A device that allows the music accompaniment device to sing along with the singer; however, the technical tone of the technique is used for a predetermined period of time, and the average pitch is re-bored, and then the average fen test tone is compared to provide _ loss The tone of music; of course, the calculation of this kind of music with the singer's voice is calculated by taking an average value (that is, the average time interval of the average pitch) (such as the case of 5 seconds of explicit delay, if Plus the calculation, the situation will be more obvious, therefore, the practice in the practical application experience found that its = 5 1304569 singer tone process is not singer singer after singing - sentence lyrics, in = instant ^ often must Waiting for the change of the tone of the music; and the comparison will be reached when the review is carried out, it is difficult to obtain higher accuracy = the demand for the two fixed point values should be the most demanded by the user, there are more adjustments. This is obviously an inconvenience. The space for improving the breakthrough exists. 疋X is aimed at the above-mentioned conventional music accompaniment ^, the problems existing in the use, how to develop a two-tune method to make the innovative structure of sex, it remains to be treated = The person with the ideal and practical direction. The goal of the heart and the heart is broken. In view of this, the inventor has been working with the design experience for the purpose of the above-mentioned target product development - the practical invention of the invention ^... After careful evaluation, [Summary of the Invention] The main purpose of the present invention is to provide a special burden, that is, to provide a calculator in the quasi-sense, and to play with the main voice, a music accompaniment:: When people sing, they can automatically adjust the pitch of the accompaniment song according to the singing sound to reduce the :: : ' sub-sound range suitable for their own singing; its technical features include: 调 调 测 以 to the user's voice The frequency of the pre-soil frequency (Fundamental FreqUenCy); the basic frequency of the user's voice is converted to the user in one set; the == is recorded by the first-order matcher to compare the user's scale parameter = the sequence of the law It is the opposite sex; the complex use of the transposition judger ^ and when to transpose 'if you need to transpose' will automatically adjust the music:::: 1304569 Parameter 'The last musician becomes crying, # t _.......... Chengyi Department will be recorded in digital Synthesize the actual two, 本, 、, ', 、, and after using the singer together with the sound mixer (nuxer) " 刺口八=者4 This innovative unique design, the user's song voice tone calculation method ^: In a small time calculation-secondary manner, the scale sequence 4 is used to convert the basic frequency of the user's voice into a user scale parameter sequence; that is, the present invention compares the main melody scale sequence with the user vocal sequence sequence. The scale matching device compares the degree of matching of a scale sequence, belongs to the mode of the dynamic comparison scale curve, and outputs the scale difference when it is optimally matched, because the scale matcher dynamically compares the scale sequence curve for a period of time, so The resulting transposition value has higher accuracy and can achieve the best pitch adjustment effect, which is more suitable for users. Moreover, the technical feature of the present invention directly obtains the main melody of the recorded song, does not require a complicated operation process, has a low system operation amount, and occupies low system resources, so the overall operation efficiency and immediacy are high. It can greatly improve the problem of the delay of the conventional system and achieve more practical and useful progress. [Embodiment] Please refer to Figures 1 and 2, which are preferred embodiments of the automatic tracking method of the musical accompaniment apparatus of the present invention, but the embodiments are for illustrative purposes only, and are not applicable to patent applications. The limitation of the structure; the automatic tracking method includes: as shown in FIG. 1, first, the user's voice is calculated by a tone estimator 11 every short period of time (about 1 sec.). Frequency (Fund 7 1304569 ?7年月月Μ曰绦(more) is replacing Beijiao plus F Qingqi)' The tone estimator ii is the basic period or frequency for calculating the sound of this segment, which can be calculated by the maximum value. The autocorrelation of the abbreviation (aut〇c〇rr elation iunctlon) is obtained or spectrally obtained from the correlation position and spacing of the peak (%). The relationship between the fundamental period and the frequency is: Basic frequency = sampling frequency / basic period The sampling frequency is the number of sample points sampled by the sound per second, and then, in a first-order sequence recorder, the basic material of the series of user input sounds is converted into the user sound sequence 121, and recorded, scale and frequency. The relationship is as follows: Frequency of 12-scale A4 44〇Hz, every half of the scale frequency increases by Λ^ times, every half of the scale frequency is reduced by a factor of doubling, the scale frequency is increased by a factor of 2, and the frequency of each scale is reduced by a factor of 2, then The first scale matcher 13 calculates the difference between the user scale sequencer 21 and the main melody scale sequence 14, where the main melody scale sequence 14 is pre-stored in the music score 15 of the music, for example: in (alm. al mstniment) The digital interface) can record the information of the scores at the same time, and the scale matching unit 13 uses the dynamic time warping (Dyn·ic Tune Wannng, DTW for short) calibration method to compare the user scale sequence with the main melody scale sequence 14 The difference is as follows: Assume that the user scale sequence 121 is nl, n2, ..., nj, which represents the user (or singer) continuously calculated every fixed period of time (such as 〇·1 second). The scale (or pitch), and assume that the main melody scale sequence 14 is ml, m 2, ···, mj, representing the main melody 每隔 ^ every other fixed time (such as 1. 1 second) ' ^ in this scale is digital 1 to 255 means that the representative value of the scale C3 is 6 〇 'the representative value of the scale D3 is 61, and the representative value of the scale β3 is %, and 8 1304569 p--- is the year 曰 曰 repair (more) replacement page • This type of push 'because when singing, the position of the singer's voice is not necessarily the same as the position of the beat of the moon music, so the scale must be calibrated during the scale comparison to produce a correct ratio. For the result, the following diagram is nl n2 n3 n4 n5 n6
ml m2 m3 m4 m5 m6Ml m2 m3 m4 m5 m6
上揭實施例中，就時間的角度而t，該n2、n3 (即使 用者音階序列）係隨著m2 (即主旋律音階序列）作校準的 狀悲’以使背景音樂播纟的拍點位置與歌唱者的聲音拍點 位置能夠在正確相對應之拍點位 直作匹配比對；而進行移 調時，則是令主旋律音階序列隨著 現者使用者音階序列作移調 的動作。 9 13 04569 %年月⑹3修(更)正替換頁 . 假5又dis"t(ni，mk)表示音階此献的誤差值，acu」]ist(ni ，mk)表示過去最佳路徑到音階ni，mk的累積誤差值，則上圖 匹配之每個節點的最小累積誤差為： acu_dist(ni，mk)二 dist(ni，mk) + min{ acu—dist(ni-l，rak)，acu _dist(ni，mk-1)，acu—dist(ni-1，mk-1)，···} 其中’ min{···}表示取最小值，在丨…丨中的範圍則由經 驗決定，通常取時間校準值—2〜+2的範圍内，因此最後 匹配的結果誤差為acu—dist(nj，mj)，j為此次比對的最後一 個時間點，其大小由實驗決定通常大於4〇(4秒）小於1〇〇( 10 #> )而其中表佳路徑是指累積誤差最小的路徑，實作 上並不需求出。 基於上述做法，我們可用以算出主旋律需移調多少， 如第2圖所示，首先設定主旋律音階移調值為s = ki ，s 表示音階升高—個半音’ 表示音符降低一個半 音，然後，利用上述做法（即前述動態時間扭曲（办⑽瓜沁 階序列及 結果的累 直到s=K2 ，其中K 调及何時 某一經，驗 音符移調 然，通常 長時做調In the above embodiment, in terms of the angle of time t, the n2, n3 (ie, the user scale sequence) is calibrated with m2 (ie, the main melody scale sequence) to make the background music broadcast the beat position. The position of the beat with the singer's voice can be directly matched at the correct corresponding beat position; and when the transposition is performed, the main melody scale sequence is transposed with the current user's scale sequence. 9 13 04569 %年月(6)3修(more) is replacing page. False 5 and dis"t(ni,mk) indicates the error value of the scale, acu"]ist(ni,mk) indicates the past best path to scale Ni, the cumulative error value of mk, the minimum cumulative error of each node matched by the above figure is: acu_dist(ni,mk) two dist(ni,mk) + min{ acu—dist(ni-l,rak), acu _dist(ni,mk-1), acu-dist(ni-1,mk-1),···} where 'min{···} means taking the minimum value, and the range in 丨...丨 is determined by experience Usually takes the time calibration value -2~+2, so the result of the final match is acu_dist(nj,mj), where j is the last time point of the comparison, and its size is usually determined by the experiment. 4〇(4 seconds) is less than 1〇〇(10#>) and the best path is the path with the smallest accumulated error, which is not required in practice. Based on the above, we can use to calculate how much the main melody needs to be transposed. As shown in Figure 2, first set the main melody scale shift value to s = ki, s to indicate that the scale is raised - a semitone ' indicates that the note is reduced by one semitone, and then, using the above Practice (that is, the aforementioned dynamic time warping (doing (10) the sequence of the melon order and the result of the accumulation until s = K2, where K is adjusted to a certain time, the accent is shifted, usually long time
Time Waging，簡稱DTW)校準方式）比對使用者音 移調s後的主旋律音階序列，並記錄其最後匹配 積誤差值=Dls(s)，接著令ps+1，再次計算Dis(s) ，最後找出移調值s = Smn使得Dis (Smin)是最小值 1<= s <=K2。通常設 K1= 一 6，K2 二 6。 接著，再藉由一移調判斷器16決定是否須移 移=，該移調判斷器16是根據誤差值Dis㈤小: 值常數D時才做移調處理，移調處理是將主旋律 s個半音，在調整時，並應考慮音樂的和諧及自 每隔-段時間調整一次’且通常在主旋律音長較 10 1304569 整。 乃年少月’乙曰修(更)正替換頁 其中，該音樂合 貝際的音樂波形，最 18(mixer)輪出，若需 階參數，具體做法係 高或降低幾個音階數 半曰，但並不限定， 剛好相差兩倍，頻率 高於6個半音以上可 則可用升調來處理。 成器17 後與使 移調時 可將樂 ，此所 因為相 差兩倍 以用降 是將以數位 用者的歌聲 ，則調整音 譜15中所有 述音階數通 差12個半音 對音感而言 調來處理， 記錄的樂譜 一起經由一 樂合成器17 音符音階一 常小於或等 (8度音）等 似乎是相同 低於6個半 15合成 混音器 中之音 起做升 於6個 於頻率 的’而 音以上 以下茲提出另一實作範例： 、當播放背景音樂時，即開始進行錄音，並設定聲音格 ^單音16 bltS ’取樣頻率為侧Hz，每次錄製的長度 二。·"少’接著在步驟音調估測器u中，計算歌唱者的聲 音基本頻率，其方法如下，假設所錄製的聲音為：Time Waging (DTW) calibration method) Align the main melody scale sequence after the user's tone shift s, and record the last matching product error value = Dls(s), then let ps+1, calculate Dis(s) again, and finally Find the transposition value s = Smn such that Dis (Smin) is the minimum value 1 <= s <= K2. Usually K1 = one 6, K2 two 6. Then, by a transposition determiner 16, it is determined whether or not to shift =, the transposition determiner 16 is small according to the error value Dis (five): the transposition processing is performed when the value constant D, and the transposition processing is to s semitones of the main melody. And should consider the harmony of music and adjust it once every other time - and usually the length of the main melody is 10 1304569. It is a young moon, 'Bai Xiu repair (more) is replacing the page, where the music is in the middle of the music waveform, the most 18 (mixer) rounds, if you need the order parameters, the specific method is high or reduce the number of scales a few times, However, it is not limited. It is exactly twice the difference. If the frequency is higher than 6 semitones, it can be processed by the rising tone. After the player 17 and the transposition, the music can be played, because the difference is twice as long as the drop is the voice of the user, and then all the scales in the sound spectrum 15 are adjusted by 12 semitones. Processing, the recorded scores together via a music synthesizer 17 note scales are often less than or equal (8 degrees), etc. seem to be the same as the sounds in the 6-and-a-half-15 synthetic mixers up to 6 frequencies 'And the sound below is another example of the implementation: When playing background music, it starts recording, and sets the sound box ^ single tone 16 bltS 'sampling frequency is side Hz, the length of each recording is two. "less' then in the step tone estimator u, the vocalist's fundamental frequency of the sound is calculated as follows, assuming that the recorded sound is:
义似 料410，則 ι· 叶异自相關函數nd)，其中： n(k) = Ση x(n)x(n—k)，，…，N _ 卜 k=22, 23, 24,…，674 或々值的範圍代表所欲偵測的頻率範圍· 44100/22〜44100/674 = 2004.54 〜65· 43Hz 2,尋找，“代表當八⑻有最大值時的The meaning is 410, then the ι· leaf autocorrelation function nd), where: n(k) = Ση x(n)x(n—k),,...,N _ bk=22, 23, 24,... , the range of 674 or 々 represents the range of frequencies to be detected · 44100/22~44100/674 = 2004.54 ~ 65· 43Hz 2, looking for, "representing when eight (8) has the maximum value
Jc值。 基本頻率 f〇= 441001 音階代號，假設若基本頻 然後，將該基本頻率轉換為 率二440 Hz則將之換為音階 1304569 π年^月彳曰修(更)正替換頁 Μ(中音La)，音階代號為69，每差半音，頻率差 气倍’音符代號差1，由音階序列記錄器12記錄之 ，並在主旋律音階序列14記錄主旋律音階代號，在音 階匹配器13中，首先設定K1二—6 ，K2二6 ，然後設 定其音階代號序列長度為4秒（j=40)，每錄音〇·丨秒 計算一次，4秒共40次，假設記錄主旋律音階序列14 為mi，1 = 0，1，2，···，39，使用者歌聲音階序列為ni， i = 〇，1，2，…，39，移調為s，並假設音階代號此 • nk 的差值為 dist (mi，nk)，令其中 dist (mi，nk) >= 〇， 且令當mi, nk相差一個八度音（12個半音）時，其計 算誤差值結果相等，即： dist (mi, nk) = dist (mi+12*N, nk), 其中N為整數’並設時間校準值—i〜刊的範圍内 ，則音階匹配器13之做法如下， 、 1. 設定s=Kl 2. 設定i=l ’且設定累積誤差值陣列 Φ acu—dist[0〜39][0〜39]初值為一個非常大的值 1000000 3. 計算 acu_dist [0][0]= dist (mO+s，n0) 4. 設定 j = i 一 1 5. 若 j 40則跳到步驟8 6· acu—dist [i][j]二 min{ acu一dist [i-l][j-1]， acu_dist [i-l][j]， acu一dist [i][j-l]} + dist (mi+s， nj) 7. j二：i +1若 j〈二i +1則回到步驟5 8. l二i +1 若 i < 40 則回到步驟4 12 1304569 -- ??年了月“日修(更)正替換頁 - 9. Dis(s) - dtwjist[39][39] 10· s = s+1 11. 若s <= K2 ，則回到步驟2 12. 結束。 驗值），且目前播放之主旋律立—㈣係為經 9付音長秒時，則將主 方疋律移調smin個半音並間隔4秒（4和 、 — · · 私係為經驗值）以上，才Jc value. The basic frequency f〇= 441001 scale code, assuming that the fundamental frequency is then converted to a rate of two 440 Hz, then it is converted to a scale of 1304569 π years ^ month 彳曰 repair (more) is replacing the page Μ (middle La ), the scale code is 69, each semitone, the frequency difference is 'the note number difference 1, is recorded by the scale sequence recorder 12, and the main melody scale code is recorded in the main melody scale sequence 14, in the scale matcher 13, first set K1 2-6, K2 2, then set the length of the scale code sequence to 4 seconds (j=40), calculate once per recording 〇·丨 seconds, 40 times in 4 seconds, assuming that the main melody scale sequence 14 is mi, 1 = 0,1,2,···,39, the sequence of the user's songs is ni, i = 〇,1,2,...,39, transposed to s, and the scale value of this nk is assumed to be dist ( Mi, nk), where dist (mi,nk) >= 〇, and when mi, nk differ by one octave (12 semitones), the calculated error values are equal, ie: dist (mi, nk ) = dist (mi+12*N, nk), where N is an integer 'and set the time calibration value—i within the range of the publication, then the scale The operation of the device 13 is as follows, 1. Set s=Kl 2. Set i=l ' and set the cumulative error value array Φ acu-dist[0~39][0~39] The initial value is a very large value 1000000 3 Calculate acu_dist [0][0]= dist (mO+s,n0) 4. Set j = i to 1 5. If j 40 then jump to step 8 6 · acu-dist [i][j] two min{ Acu-dist [il][j-1], acu_dist [il][j], acu-dist [i][jl]} + dist (mi+s, nj) 7. j two: i +1 if j< 2 i +1 then go back to step 5 8. l 2 i +1 If i < 40 then go back to step 4 12 1304569 -- ?? Years of the month "day repair (more) is replacing page - 9. Dis(s ) - dtwjist[39][39] 10· s = s+1 11. If s <= K2 , return to step 2 12. End. Test value), and the main melody currently playing is - (4) is the 9 When paying for a long second, the main law is transposed by smin semitones and separated by 4 seconds (4 and , — · · private is the empirical value).
進仃下一次之移調；最後音樂人 斑吃入二— . 市口成杰17是將以數位記錄的 合成貫際的音樂波形，然後 哭w + 1更用者歌茸一起經混音 口口 18(mixer)、°刺卩八19輸出者。 【本發明之優點】 、本發明所揭音樂伴奏裝置之自 自動跟调方法，其使用者 之歌聲音調計算方式並非 、便用者 ，、平均值，而是採用每一 小段時間計算一次（如〇 立_ & , ·心）之方式，復藉由所述 曰階序列記錄器12將使用者舞立 去立时奴t 便用者聲音的基本頻率轉為使用 者音階芩數序列121 ； ^ gP , A ^ R9In the end, the next person will be transferred to the second; the last part of the music will be eaten into the second--. Shikou Kenji 17 is a synthetic music waveform that will be recorded in digital form, and then crying w + 1 is more used by the singer along with the sound mouth. 18 (mixer), ° Hedgehog eight 19 exporters. [Advantages of the Invention] The self-automatic tune method of the music accompaniment device disclosed in the present invention, the user's song tone calculation method is not, the user, the average value, but is calculated once every small time (such as In the manner of standing _ & , · heart), the basic frequency of the user's voice is converted to the user scale number sequence 121 by the step sequence recorder 12; ^ gP , A ^ R9
Rtb ^ j卩本發明係比對主旋律音 I1白序列14及使用者歌聲音 垧立1 + 0序歹】丨21 ，而非比對其平 均曰调，所述之音階匹g _的3疋比對一段音階序列的 匹配％度’此屬於一種動 動心比對音階曲線之模式，並 输出其攻佳匹配時的音階# ^ ^ ^ 差，口為該音階匹配器13是 動悲比對一段時間的音 “白相曲、線，而非僅-段時間 的曰调平均值，故所得 A n 亍之私凋值具有較高之準確性， 月匕夠獲件琅佳之音調調整 所需。 效果，進而更符合使用者之 13 13 04569 . 竹年r月/6日修(更)正替換頁j 本毛明之技術特點係直接取得已記錄歌曲之主旋律， 亚不需要一堆複雜的運算過程，具有較低之系統運算 里，佔用較低的系統資源，故整體具有較高之運作效 率和即時性，確可大幅改善習知系統延遲之問題點， 達到更加實用好甩之進步性者。Rtb ^ j卩 The present invention compares the main melody tone I1 white sequence 14 and the user's song voice to 1 + 0 sequence 丨 丨 21, rather than the average 曰, the scale of the scale g _ 3 疋Comparing the matching % degree of a scale sequence 'This is a mode of the dynamic kinetic comparison scale curve, and outputs the scale # ^ ^ ^ difference when the attack is well matched, the mouth is the scale matching device 13 is a sorrow comparison The sound of time "white phase curve, line, not just the average value of the time period, so the private value of the obtained A n 具有 has higher accuracy, and the moon is enough to obtain the better tone adjustment. , and more in line with the user's 13 13 04569. Bamboo Year r / 6 days repair (more) is replacing page j The technical characteristics of Ben Mao Ming directly obtain the main theme of the recorded song, Ya does not need a complicated process of calculation, With lower system operation and lower system resources, the overall operation efficiency and immediacy of the whole system can greatly improve the problem of the delay of the conventional system, and achieve more practical and good progress.
上述貫把例所揭示者係藉 雖透過特定的術語進行說明， 利範圍，熟悉此項技術領域之 神與原則後對其進行變更與修 等變更與修改’皆應涵蓋於如 定範_中。 以具體說明本發明，且文中 當不能以此限定本發明之專 人士當可在瞭解本發明之精 改而達到等效之目的，而此 后所述之申請專利範圍所界 14 1304569 (~----- 打年少^月仏日淑更^替換頁 【圖式簡單說明】 第1圖：係本發明音樂伴奏裝置之自動跟調方法之系統方 塊圖。 第2圖：係本發明之音階匹配器動作流程之方塊示意圖。 【主要元件符號說明】 音調估測器 11 音階序列記錄器 12 使用者音階序列 121 音階匹配器 13 主旋律音階序列 14 樂譜 15 移調判斷器 16 音樂合成器 17 混音器 18 口刺口八 19 15The above descriptions are disclosed by way of specific terminology, and the scope of the changes and modifications of the spirit and principles of the technical field should be covered in the definition. The invention is specifically described, and the person skilled in the art can not limit the invention to achieve the equivalent purpose, and the patent application scope is 14 1304569 (~- --- Playing the young ^ ^ 仏 仏 淑 ^ ^ ^ replacement page [schematic description] Figure 1 is a system block diagram of the automatic accompaniment method of the music accompaniment device of the present invention. Figure 2: the scale matching of the present invention Schematic diagram of the action flow of the device [Description of the main components] Tone estimator 11 Scale sequence recorder 12 User scale sequence 121 Scale matcher 13 Main melody scale sequence 14 Score 15 Transposition judger 16 Music synthesizer 17 Mixer 18 Mouth pierce eight 19 15