TW587241B - Automatic accord identification and generation method - Google Patents

Automatic accord identification and generation method Download PDF

Info

Publication number
TW587241B
TW587241B TW90132861A TW90132861A TW587241B TW 587241 B TW587241 B TW 587241B TW 90132861 A TW90132861 A TW 90132861A TW 90132861 A TW90132861 A TW 90132861A TW 587241 B TW587241 B TW 587241B
Authority
TW
Taiwan
Prior art keywords
chords
song
chord
main
melody
Prior art date
Application number
TW90132861A
Other languages
Chinese (zh)
Inventor
Jr-Shing Jang
Hung-Ru Li
Wen-Ni Jeng
Original Assignee
Cweb Technology Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Cweb Technology Inc filed Critical Cweb Technology Inc
Priority to TW90132861A priority Critical patent/TW587241B/en
Application granted granted Critical
Publication of TW587241B publication Critical patent/TW587241B/en

Links

Landscapes

  • Auxiliary Devices For Music (AREA)

Abstract

The automatic accord identification and generation method implemented by the invention enables user to input major melody data of song by virtue of electronic peripherals such as microphone, computer keyboard and so on and convert that into specific format with digital processing to further utilize major melody of song in song database for pattern creation in effort to acquire major melody accord intentionally searched by user.

Description

經濟部智慧財產局員工消費合作社印製 587241 A7 B7 五、發明說明(1 ) [發明之技術領域] 本發明係有關於一種和弦(Chord )自動辨識及產生之方 法,特別是指,本發明係令使用者能藉由輸入歌曲之主旋律資 訊,經由數位處理並轉換成特定格式,進而利用歌曲資料庫中 之歌曲主旋律加以訓練建立模型,以獲得與主旋律適當對應和 弦之方法。 [發明背景與習知技術] 在習知技術中,可藉由歌曲主旋律配伴奏及配樂之電腦系 統,如數位鋼琴、電子琴,係必須藉由使用者以右手彈奏主旋 律、左手配合彈奏和弦,才得以配出適當的伴奏,而在藉由此 系統配出適當的伴奏之前,使用者必須先獲知歌曲之主旋律和 弦才能配出適當的伴奏,所以當使用者不具備依主旋律自行辨 識和弦之能力時,亦無法利用此電腦系統來為主旋律配出伴奏 及其配樂。 然而,除了要找出旋律之和弦外,尚需找出主旋律之調性 (Music Style),因此,基於習知技術之不便,本發明提供一 主旋律自動配伴奏之方法,係令使用者能藉由輸入的歌曲主旋 律而獲得適當的和弦伴奏及調性,且於使用者不具備任何和弦 樂理常識及彈奏樂器的能力時,系統亦能依據使用者所輸入之 主旋律,並配合資料庫的訓練以找出最適當的和弦,進而達到 配伴奏的目的。 [發明概述] 本發明之主要目的係提供一種和弦自動辨識及產生之方 法,讓使用者能藉由輸入歌曲主旋律的方式,並經由資料庫的 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) --------^--------- (請先閱讀背面之注意事項再填寫本頁) 587241 A7 B7 五、發明說明() 訓練而獲得匹配所輸入歌曲主旋律之和弦及調性。 (請先閱讀背面之注意事項再填寫本頁) 本發明之次要目係提供一種和弦自動辨識及產生之方 法,讓使用者在未具備任何樂理及彈奏樂器能力之前提下,仍 能藉由輸入歌曲之主旋律的方式於資料庫中獲得適當之和弦 及調性。 為達成上述目的,本發明所提供之一種和弦自動辨識及產 生之方法,令使用者能藉由輸入一段歌曲之主旋律,進而將匹 配主旋律調性之和弦伴奏從資料庫中擷取出來之方法,前述方 法包括:主旋律輸入及處理程序,係令使用者藉由電子週邊設 備輸入歌曲之主旋律,例如麥克風、键盤等,並儲存成一定的 電子資料格式;資料庫處理程序,係將資料庫中之中介格式歌 曲,將其主旋律及和弦資料取出並存成特定之資料格式;資料 訓練程序,係將資料庫處理程序所得之資料加以訓練並建立模 型,以作為和弦產生之處理資料;和弦產生程序,係根據主旋 律輸入及處理程序所得之主旋律資料及資料訓練程序所得之 和弦模型,找出主旋律之和弦;結果輸出程序,係依據和弦產 生程序所得之和弦輸出並與主旋律配合伴奏。 [圖式標號說明] 101 — 主旋律輸入及處理程序 經濟部智慧財產局員工消費合作社印製 102 --- 資料庫處理程序 103 -― 資料訓練程序 104 --- 和弦產生程序 105 --- 結果輸出程序 [發明之詳細說明] 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 經濟部智慧財產局員工消費合作社印 587241 五、發明說明( 首先明參α考圖—與圖二,係顯示本發 產生方法之架構圖及方法流程圖。本發;動辨 輸入一段歌曲之主旋律,進而將使用者欲ΐ: “使用者藉由 料庫中擷取出來之方法,此方法包含:顿之和弦伴奏從i 权序101 ·主旋律輪入及處理 係令使用者能籍由電子週邊設備 包含使用者藉由麥克風哼唱—段歌曲\曲的主旋律,其中 歌曲的主旋律或電子設備預先編辑的 _1腦鍵盤輪入 電腦鍵盤輪人歌曲主旋律編辑㈣以^,例如,利用 式’或經由錄音駿置事先將歌曲 (的電子資料格 子資料格式,錢人電子㈣之財財心同之電 格式可為—種數位之聲音格式,如_E 前述之電子資料 取之電子數位資料格式進行基頻 =、AU..·等。且將操Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 587241 A7 B7 V. Description of the Invention (1) [Technical Field of the Invention] The present invention relates to a method for automatically identifying and generating a chord, and in particular, the present invention The user can input the main melody information of the song, digitally process and convert it into a specific format, and then use the song main melody in the song database to train and build a model to obtain a method that appropriately corresponds to the main melody. [Background of the Invention and Known Techniques] In the known techniques, computer systems such as digital pianos and electronic pianos can be accompanied by accompaniment and music to the main melody of the song. The user must play the main melody with the right hand and the chord with the left hand. To be able to match the appropriate accompaniment. Before the system can match the appropriate accompaniment, the user must first know the main melody and chord of the song to match the appropriate accompaniment. Therefore, when the user does not have the ability to identify the chord by the main melody. When you are able, you cannot use this computer system to accompany the main melody and its soundtrack. However, in addition to finding the chords of the melody, it is necessary to find the music style of the main melody. Therefore, based on the inconvenience of the known technology, the present invention provides a method for automatically accompaniment of the main melody. The appropriate chord accompaniment and tonality are obtained from the inputted song's main melody. When the user does not have any common knowledge of chord theory and the ability to play musical instruments, the system can also follow the main melody entered by the user and cooperate with the training of the database. In order to find the most appropriate chord, and then achieve the purpose of accompaniment. [Summary of the Invention] The main purpose of the present invention is to provide a method for automatic identification and generation of chords, so that the user can input the theme of the song and apply the Chinese National Standard (CNS) A4 specification through the paper size of the database ( 210 X 297 mm) -------- ^ --------- (Please read the notes on the back before filling out this page) 587241 A7 B7 V. Description of the invention () Training and matching Chords and tonality of the main melody of the entered song. (Please read the notes on the back before filling this page) The secondary purpose of the present invention is to provide a method for automatic identification and generation of chords, which allows users to borrow before they have any theory and ability to play musical instruments. The appropriate chords and tonality are obtained from the database by entering the main melody of the song. In order to achieve the above object, a method for automatic identification and generation of chords provided by the present invention enables a user to extract a chord accompaniment that matches the tonality of the main melody from a database by inputting the main melody of a song. The aforementioned method includes: a main melody input and processing program, which enables a user to input the main melody of a song through an electronic peripheral device, such as a microphone, a keyboard, and the like, and stores it into a certain electronic data format; a database processing program, which stores data in a database Songs in intermediary format, taking out their main melody and chord data and storing them into a specific data format; data training procedures, training and building models based on data obtained from database processing procedures, as processing data for chord generation; chord generation procedures, It is based on the main melody input and processing program and the chord model obtained from the data training program to find the chord of the main melody; the result output program is based on the chord output obtained by the chord generation program and accompanied with the main melody. [Explanation of Graphical Symbols] 101 — Main melody input and processing program Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs Employee Consumer Cooperatives 102 --- Database processing program 103 --- Data training program 104 --- Chord generation program 105 --- Result output Procedure [Detailed description of invention] This paper size is applicable to China National Standard (CNS) A4 (210 X 297 mm). Printed by the Consumers ’Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs. Second, it shows the architecture diagram and method flow chart of the method of generating this hair. This hair; identify the main melody of a song, and then make the user want: "The method that the user retrieves from the database, this method Contains: Tonson ’s chord accompaniment from i. Order 101. The main melody rotation and processing system enables users to use electronic peripherals to include the user's humming through a microphone—the main melody of a song / song, where the main melody or electronic device of the song The pre-edited _1 brain keyboard turns into the computer keyboard and turns the song into the main melody editor. ^, For example, using the formula 'or through the recording to set the song (the Electronic data grid data format. The electronic format of Qianren Electronic's financial wealth can be a digital audio format, such as _E, the electronic digital data format taken from the electronic data above, the base frequency =, AU .., etc. ... and will fuck

Tracking),將其轉成相對應之音二冉基頻追蹤,Pitch =係用於提供歌曲主旋律和弦產生之:據。::換 =追蹤健將特的數位衫格··位減_轉換,轉 換成在頻域上的分佈’例如:傅利葉轉換(F。咖Tracking), which is converted to the corresponding tone Erran base frequency tracking, Pitch = is used to provide the song's main melody chord generation: data. :: Transformation = Tracking of the special digital shirt grid · · Subtraction _ conversion, converted to the distribution in the frequency domain ’For example: Fourier transform (F. Coffee

TranSf〇rmati〇n)、快速傅利葉轉換(Fast F斷ier W_tion)、小波轉換(—七以㈣卿)等方式。然 後’再將其高頻部份與低頻部份分離,藉以將每一小段之訊號 的基頻找出。隨後將所得之主旋律基頻編碼··依照每一小節的 基頻組成作為編碼依據。 程序1〇2 ··資料庫處理 係從資料庫中,將所有中介格式資料做主旋律及和弦提出TranSfromation), Fast Fourier Transform (Fast Fierier W_tion), Wavelet Transform (-Qi Yiqing) and other methods. Then, the high frequency part is separated from the low frequency part, so as to find out the fundamental frequency of each small signal. Subsequently, the obtained main melody fundamental frequency is coded... According to the fundamental frequency composition of each measure as a coding basis. Program 102 ··· Database processing From the database, put all the intermediate format data as the main melody and chord

本紙張尺度適用t國國家標準(CNS)A4規格(210 X 297公爱"J --------訂---------^Awi. Γ凊先閱讀背面之注意事項再填寫本頁) A7 A7 經濟部智慧財產局員工消費合作社印製 B7 五、發明說明(4 ) 2:。:中介格式是以音高、音長方式儲存,所以必須將音高、 曰長及和弦資訊提取出來並加以編碼儲存。 程序103 :資料訓練程序 圖―’係顯不本發明和弦自動辨識及產生之方法之 口弦決朿樹之基本架翻。如圖三所示,經由資料庫處理程序 ,中’丨格式之音鬲、音長及和弦資訊提取出來並加以編 再依序對每-首歌的資料輪人以建立和弦決策樹。而 依據則述資料庫處理程序102所得之編碼資料,可藉由下列兩 種方式以建立和弦之決策樹: (1)以統计方式為基準: 首先’先將資料庫中每一首歌的每—小節乂内的音符做數 里、.克计亚依照音階作排序得&。例如:^丨⑽仙心 DV〇1 S〇1卜處理後可得卜[D〇 S〇1 La];第二步,建立 決,樹7; ’ 1 = 12,方式為分別以十二個組成半音作為第一個 層節點仏(樹根),但是必須符合下列性質 ㈣第7層節點内的“個向量元素,7心、向量元素的個數) (l)'uezu 為第(樹,V/ ,其中,“/ + 1,K,U,众2 ⑶中的每-個元素皆小於、的每—個元素。 請繼續參閱圖三,如圖三所示,第三步,係利用"整決 策樹(,也就是依照其向量元素'依序調整決策樹厂立中 ㈣。例如:pD〇 ”2=Sol ”3=La ’因為樹第—層(片) 且^ 1二Do,所以找第/;顆樹;樹第二層(广2 )且 % ;樹第三層(户3 )且二二 --------------------訂--------- (請先閱讀背面之注咅?事項再填寫本頁) A7 五、發明說明( 因為4的向量元素個數為3,也 閱 四,係顯示圖四本發明和弦自動辨識:厂停止。請參閲圖 和弦標料四所示’於第四步時,將資料庫中的 重複上述至::庫 行過-遍後停止。 / 斤有的歌曲之所有小節都執 (2)以規則方式為基準: 之和:=型=發::::_-生一 弦之標準答案::=藉由人為的樂理_^ 則1_〇1]1—_1^1^1、¥11^1^小 第三步,依序走至w a、a,㊉;步,依序找出決策樹7卜 將c、〜埴入w "5'2'^〜1。’·第四步’依據樂理判斷依序 . 5,28气3,丨。中。重稷上述四個步驟,直至資料庫 中所有的歌曲之所有小節都執行過一遍後停止。付庫 此外,亦可利用樂理之簡易觀念找出可 經濟部智慧財產局員工消費合作社印製 ==是尋找和弦的重要依據,所以可將各個和弦_ 做排列組合’並利用小節中1^弦、、且成曰 :,得_可_首:^ 和弦為其組成音為心二 〇^。1} ’ {Dc),Mi,sol}的其中組合時,則可This paper size applies to the national standard (CNS) A4 specification (210 X 297 Public Love " J -------- Order --------- ^ Awi. Γ 凊 Please read the note on the back first Please fill in this page again for matters) A7 A7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs B7 V. Description of Invention (4) 2 :. : The intermediate format is stored in the form of pitch and pitch, so the pitch, length, and chord information must be extracted and encoded for storage. Program 103: Data training program Figure ― 'shows the basic frame of the chord deciduous tree which shows the method for automatic identification and generation of chords of the present invention. As shown in Fig. 3, through the database processing program, the sound information, sound length, and chord information in the Chinese format are extracted and compiled, and then the data of each song is turned in turn to build a chord decision tree. According to the coded data obtained by the database processing program 102, the chord decision tree can be established in the following two ways: (1) Based on statistical methods: First, 'the first of each song in the database The notes in each bar measure are counted in miles, and Kejiya is sorted according to the scale to get &. For example: ^ 丨 ⑽ 仙 心 DV〇1 S〇1 can be obtained after processing [D〇S〇1 La]; the second step is to establish a decision, the tree 7; '1 = 12, the way is to twelve The semitone is composed as the first layer node 仏 (root of the tree), but must meet the following properties: 个 vector elements, 7 hearts, and the number of vector elements in the 7th layer node (l) 'uezu is the (tree, V /, where "/ + 1, K, U, Zhong 2 (3) each element is less than every element of. Please continue to refer to Figure 3, as shown in Figure 3. The third step is to use " Entire decision tree (that is, adjust the decision tree in accordance with its vector elements 'sequentially. For example: pD〇 ”2 = Sol” 3 = La' Because the first layer (layer) of the tree and ^ 1 two Do , So find the first /; tree; the second layer of the tree (wide 2) and%; the third layer of the tree (house 3) and two two ------ -Order --------- (Please read the note on the back? Matters before filling out this page) A7 V. Description of the invention (because the number of vector elements of 4 is 3, also read 4) The chords of the present invention are automatically identified: the factory stops. Please refer to the figure of the chords on the fourth mark 'on the fourth In the database, repeat the above to the :: to the library line and stop after it. / All the verses of some songs are performed (2) Based on the regular method: Sum: = type = hair ::::: _-The standard answer to a string :: = by artificial music theory ^^ 1_〇1] 1—_1 ^ 1 ^ 1, ¥ 11 ^ 1 ^ Small third step, go to wa, a in order,步; step, in order to find the decision tree 7 bu will c, ~ 埴 into w " 5'2 '^ ~ 1.' · The fourth step 'according to music theory to judge in order. 5,28 qi 3, 丨. In . Repeat the above four steps until all the sections of all songs in the database have been executed once and stop. Pay the library. In addition, you can also use the simple concept of music theory to find out that the Intellectual Property Bureau of the Ministry of Economic Affairs can print the consumer cooperatives = = Is an important basis for finding chords, so each chord _ can be arranged and combined 'and use the 1 ^ string in the measure, and then say :, get _ 可 _ 首: ^ The chord is the composition of the heart. 1} '{Dc), Mi, sol}

本紙張尺度_ +關家鮮(CNS)A4規格⑽ x 297公釐) A7 B7 五、發明說明( 二理:广广音高為㈣’ {Mi},叫,。川 J 5¾仁该小即亦可能包舍c和 :’軸㈣小節的音高同時被兩個以二== 時’則必須將所有包舍此種組合的和弦全部加以 N〇te=[D〇],建議和弦有C、J、F。 . 在將::r[DoMi],建議和弦有二。依此類推。 4=+:弦的可能組合都建立之後,即可依據小節的立 u)每小即的第一個非零音高· 嫩峨和弦。 -=:=rr;Ar前述…模” 弦。 方法可用以找出更客觀的建議和 弦亚非早獨存在的,於前後的進行都有相 ;明=亦^ 建的一,下來 ⑴以小節4單位去預測判斷下-小節的和弦. 埴入、規則方式為基準的方法-樣,不同在於,所 填入之和弦答案乃下一小 ^ 向趨勢的決策樹。例如· U此方法建立和弦走Size of this paper _ + Guan Jiaxian (CNS) A4 size ⑽ x 297 mm) A7 B7 V. Description of the invention (Second reason: Guangguang Pitch is ㈣ '{Mi}, called, Sichuan J 5¾ should be small It is also possible to enclose c and: 'When the pitch of the axis bar is simultaneously divided by two ==', then all chords of this combination must be added to Note = [D〇]. It is recommended that the chord has C , J, F.. After putting :: r [DoMi], it is suggested that there are two chords. And so on. 4 = +: After the possible combinations of chords have been established, you can follow the bar standing. Non-zero pitch · Nene chords. -=: = rr; Ar mentioned above ... modulo "chords. The method can be used to find more objective suggestions. Chords of Asia and Africa existed early alone, and they have similarities in the progress of before and after; Ming = Yi ^ The first one, down to the next section 4 units to predict and judge the chord of the next bar. The method of entering and regular methods is the same-the difference is that the filled chord answer is the decision tree of the next small ^ trend. For example, this method creates a chord. go

Fa u];第二步,依序找出決 本紙張尺度適用中國國iii7CNS)A4規 A7 A7 五、發明說明( 策樹7; B7 + ;第三步’依序走至〜8、《咖四步,依據料 ㈣理規則判斷將養弦版填 :目充: 是係用目前的小節音符來判斷下-小節的和弦。 (2)結合方式d)及前述之兩種決策樹: 以此調整預測小節和弦的統計值,將兩種決策樹建議相同 弦的值提馬’相對地降低其他和弦的統計值。例如. ^ La];第-決策樹依序建議< :[c、£j、up : 弟二決策樹建議 < 後應為[K];就A/2而言,m[Z)、 化,藉此提高込的統計值。 ⑺ 程序104 :和弦產生 係將程序101 t編碼資料輪入程序1〇3 @決策樹取得建蟻 2弦,藉由程序103彳以得12顆決策樹,並且每一層的每一個 即點都會有數個和弦答案,在此採用下列兩種方式取得最佳 弦建議: (1)以動恶規劃法(办^am/c外观沿所所邮)做出最後決策· 由程序101的編碼資料輸入程序1〇3後,可能會產生多組 和弦答案。經由建立小節與小節間的關係,且藉資料庫以動鲅 規劃法調整出最佳的和弦組合。例如: <之最佳和弦組合有 [C、細]、仏之最佳和弦組合有[如、w、<之最佳 和弦組合有[⑼、F]、从4之最佳和弦組合有。此時, 由从的(7開始,藉由資料庫統計得知^至从2的^^q的 統計值;再來同樣地計算出Μι的以至Μ的如、以、的統 計值;依序將从2的元素至Μ的所有元素之統計值算出,並以 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) ^---------. (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作杜印製 五、發明說明(8 ) 物,+, ’最後利用動態規劃法依其間的統計值 u ]組合。·前㈣子,最佳和弦組合切: 請 ⑵以規雜計料建立祕和_㈣勢:7 ( 三㈣為例’和其相關的三聲和弦就有大三和弦c、小 屬七和二弦心和弦。四聲和弦有大七和弦、 弦等。再叫弦的ζ=(Γ弦、小七減五和弦及減七和 多、更複m I 1〇n 〇f Chord)種類就更 音樂:本广瞭_就等於大致瞭解 用辦 料沒有調性的音樂時,即可以利 用d作依據。以藍調曲式功能為例’其原有曲式如下··Fa u]; The second step is to find the paper size of the final paper in order to apply China iii7CNS) A4 regulations A7 A7 V. Description of the invention (Strategy tree 7; B7 +; The third step is to go to ~ 8, "Ca Four steps, according to the rules of material theory, fill in the string-cultivating version: Eye filling: It is to use the current bar notes to determine the chord of the next-bar. (2) Combination mode d) and the two above-mentioned decision trees: Adjust the statistics of the predicted bar chords, and use the two decision trees to suggest the same chord value, and lower the statistics of other chords relatively. For example. ^ La]; the first-decision tree proposes in sequence <: [c, £ j, up: the second decision tree proposes < followed by [K]; for A / 2, m [Z), To increase the statistical value of radon. 104 Program 104: The chord generation system turns the program 101 t coded data into program 103 @Decision tree to get 2 ant strings. With program 103, 12 decision trees are obtained, and each point of each layer will have a number For chord answers, here are two ways to get the best chord suggestions: (1) Make a final decision by using evil planning method (post ^ am / c appearance along the post) · Enter the coded data of program 101 After 103, multiple sets of chord answers may be generated. The relationship between the measures is established, and the best chord combination is adjusted by the database using the dynamic planning method. For example: The best chord combination of < is [C, Thin], the best chord combination of 仏 is [eg, w, < The best chord combination of [⑼, F], and the best chord combination from 4 is . At this time, starting from (7, the statistical value of ^ to ^^ q from 2 is obtained from the database statistics; again, the statistical values such as Mm and M are calculated in sequence; Calculate from the statistical value of all elements from 2 to M, and apply the Chinese National Standard (CNS) A4 specification (210 X 297 mm) at this paper size ^ ---------. (Please first (Please read the notes on the back and fill in this page again.) Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs, Consumer Cooperation Du printed five, Invention Description (8), +, 'Finally, the dynamic programming method is used according to the statistical value u]. , The best chord combination cut: Please use the rules and materials to establish the secret harmony _ potential: 7 (three examples of the three chords' and its related three chords have a major chord c, a small seventh and second chord heart Chords. Four-chord chords have major seventh chords, chords, etc. Also called strings ζ = (Γ chord, minor seventh minus five chords and minus seven sums, more complex m I 1〇n 〇f Chord) The type is more musical: this Wider _ is equivalent to a general understanding of the use of materials without tone, you can use d as a basis. Take the blues tune function as an example 'The original tune is as follows ...

U I U I Τ I r I .. 經濟部智慧財產局員工消費合作社印製 以_和弦的七個和弦為例,其中最不穩定的Μ屬七和 tU的疋主音‘。從最不穩定的和弦到最穩定的和 ^疋種從緊張職鬆的感覺,這能使聽者放下沈重的、 所以解決三全音對調性確定而言是相當重要的。 曰、、h的方式去調签動態規劃法中和弦間的統計值,也就 疋當使用者期待的和弦組合是爵士藍調曲風時,則加重 及M4等的統計值,並且加重早決三全音的統計值。 程序1〇5結果輪出:U I U I Τ I r I .. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Take the seven chords of _ chord as an example. The most unstable M is the 疋 lead of seven and tU. From the most unstable chords to the most stable chords, this can make the listener put down a heavy, so solving the tritone is very important for tonal determination. The method of “,” and “h” is used to check the statistical values between chords in the dynamic programming method, that is, when the chord combination expected by the user is a jazz blues melody, the statistical values such as M4 are emphasized, and the decision is made early. Statistics of whole tones. The result of program 105 is rounded out:

本紙張尺㈣財 χ 297公釐) 587241 A7 B7 五、發明說明(9 ) 係依據和弦產生程序104所得之和弦藉由輸出裝置來輸出 並配合主旋律伴奏。 [發明功效] 根據本發明所實施之和弦自動辨識及產生之方法將具有 諸多優點與特徵,本發明係令使用者能藉由麥克風、電腦鍵盤 等電子週邊設備輸入歌曲主旋律,經由數位處理並轉換成特定 格式,進而利用歌曲資料庫中之歌曲主旋律加以建立模型,以 獲得適合主弦律之和弦及調性。 本發明可讓使用者藉由輸入歌曲主旋律的方式,並經由資 料庫的訓練而獲得欲配伴奏之和弦,並於使用者未具備任何樂 理及彈奏樂器能力之前提下,仍然能藉由輸入歌曲之主旋律的 方式於資料庫中獲得適當之和弦伴奏。 (請先閱讀背面之注音?事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) A7 五、發明說明( [圖式之簡單說明] Ξ:=明和弦自動辨識及產生方法之架構圖。 明和弦自動辨識及產生方法之流程圖。 基本架構圖。%明和弦自動辨識及產生方法之和弦决策樹之 模型明和弦自動辨識及產生之方法之和弦決策樹 圖五為本發明和弦自動辨識及產生之方法之和弦決策樹 模型示意圖之二。 經濟部智慧財產局員工消費合作社印製The size of this paper is 297 mm) 587241 A7 B7 V. Description of the invention (9) The chords obtained according to the chord generation program 104 are output by the output device and are matched with the main melody accompaniment. [Effects of the invention] The method for automatic identification and generation of chords implemented according to the present invention will have many advantages and features. The present invention enables users to input the main theme of songs through electronic peripheral devices such as microphones, computer keyboards, etc., through digital processing and conversion. Into a specific format, and then use the song theme in the song database to build a model to get the chords and tonality that suit the theme. The invention allows the user to obtain the chords to be accompanied by accompaniment by inputting the main melody of the song and training through the database. The chords can be picked up before the user does not have any music theory and ability to play musical instruments. The main melody of the song is obtained from the library with appropriate chord accompaniment. (Please read the note on the back? Matters before filling out this page) The paper size printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economy applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) A7 V. Description of the invention ([ Brief description of the drawing] Ξ: = The architecture diagram of the automatic identification and generation method of Ming chord. The flowchart of the automatic identification and generation method of Ming chord. The basic architecture diagram. The model of the chord decision tree of the automatic identification and generation method of Ming chord The chord decision tree of the method for automatic identification and generation of chords Figure 5 is the second schematic diagram of the chord decision tree model of the method for automatic identification and generation of chords of the present invention. Printed by the Consumer Cooperative of Intellectual Property Bureau, Ministry of Economic Affairs

Claims (1)

58^ 58^ ??. ! S η A8 B8 C8 D8 六、申請專利範圍 其中前述之主旋律輸入之可讀取資料,係為使用者藉由輸入系 統之週邊輸入裝置所輸入之音樂訊號。 5 ·如申請專利範圍第1項之利用歌曲主旋律擷取和弦之方法, 其中前述之主旋律輸入系統之可讀取之資料格式,可為WAVE、 RA、或其他數位旋律之訊號。. 6·如申請專利範圍第i項之利用歌曲主旋律擷取和弦之方法,其 中前述之資料庫訓練程序中之資料模型建立方式,可為動能規 ^、資料探勘、鏔神經網路、決策樹、模糊邏辑或其他可建立功 月t*之方法。 7 ·如申晴專利範圍第!項之利用歌曲主旋軸取和弦之方法 其^述之主旋律輸出及處理程序中之基頻追縱方式可 ,茶轉換、快速傅立葉轉換、小波轉換或其他可將數位:故 式經由數位訊號轉換以成為頻域分佈之方法。 θ格 8·如申請專·㈣項之利用歌曲主旋義取和 、皮“中月)述之資料庫訓練程序中’係根據前述資料庫 序之編碼資料來建立決策樹。 、 里程 :·:::專利範圍第δ項之利用歌曲主旋律擷取和弦之方祛 專利範圍第1項之利用歌曲主旋律擷取和弦之方 則方該決策樹可綱4 1奴則歌曲輕律鮮和弦之方 ^述之結果㈣序中,㈣生之和弦結果可藉= 本紙張尺度適國國 297公釐) 587241 Α8 Β8 C8 D8 六、申請專利範圍 出裝置來輸出並配合主旋律伴奏 (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐)58 ^ 58 ^ ??.! S η A8 B8 C8 D8 6. Scope of patent application The readable data of the aforementioned main melody input is the music signal input by the user through the peripheral input device of the input system. 5 · If the method for capturing chords using the main theme of a song is applied in item 1 of the scope of patent application, the readable data format of the aforementioned main theme input system can be WAVE, RA, or other digital melody signals. 6. If the method of extracting chords from the main theme of the song is used in item i of the scope of patent application, the data model building method in the aforementioned database training procedure can be kinetic energy regulation, data exploration, neural network, decision tree , Fuzzy logic, or other methods that can be used to build the power month t *. 7 · If the patent scope of Shen Qing! The method of using the main axis of the song to take chords is as follows: The main melody output and the fundamental frequency tracking method in the processing program are as follows: tea conversion, fast Fourier conversion, wavelet conversion, or other digital conversion: so digital conversion through digital signal To become a method of frequency domain distribution. θ cell 8. If you apply for a special program, you can use the main theme of the song to get the sum, and the database training program described in "Zhongyue" is used to build a decision tree based on the coded data of the aforementioned database order. Mileage: · ::: Patent Scope Item δ Extraction of Chords Using Song Main Melody Rule of Patent Scope Item 1 Extraction of Chords Using Song Main Melody Rule The decision tree can outline 4 1 Slave songs with light rhythm and fresh chords ^ In the preface of the results described, the chord result of the birth can be borrowed = 297 mm of the size of this paper) 587241 Α8 Β8 C8 D8 6. Apply for a patent range device to output and cooperate with the main melody accompaniment (please read the first Note: Please fill in this page again.) Printed on the paper by the Consumers' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. The paper size applies to China National Standard (CNS) A4 (210 X 297 mm).
TW90132861A 2001-12-28 2001-12-28 Automatic accord identification and generation method TW587241B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW90132861A TW587241B (en) 2001-12-28 2001-12-28 Automatic accord identification and generation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW90132861A TW587241B (en) 2001-12-28 2001-12-28 Automatic accord identification and generation method

Publications (1)

Publication Number Publication Date
TW587241B true TW587241B (en) 2004-05-11

Family

ID=34057350

Family Applications (1)

Application Number Title Priority Date Filing Date
TW90132861A TW587241B (en) 2001-12-28 2001-12-28 Automatic accord identification and generation method

Country Status (1)

Country Link
TW (1) TW587241B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI457867B (en) * 2011-09-22 2014-10-21 Casio Computer Co Ltd Musical performance evaluating device and musical performance evaluating method

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI457867B (en) * 2011-09-22 2014-10-21 Casio Computer Co Ltd Musical performance evaluating device and musical performance evaluating method

Similar Documents

Publication Publication Date Title
Benetos et al. Automatic music transcription: An overview
CN112382257B (en) Audio processing method, device, equipment and medium
Wang et al. MuseBERT: Pre-training Music Representation for Music Understanding and Controllable Generation.
CN113192471B (en) Musical main melody track recognition method based on neural network
Hung et al. Learning disentangled representations for timber and pitch in music audio
Wang Research on the improved method of fundamental frequency extraction for music automatic recognition of piano music
CN112289300A (en) Audio processing method and device, electronic equipment and computer readable storage medium
Dixon et al. Probabilistic and logic-based modelling of harmony
Salamon et al. Melody, bass line, and harmony representations for music version identification
Van Balen Audio description and corpus analysis of popular music
Tyburek et al. Comparison of the efficiency of time and frequency domain descriptors for the classification of selected wind instruments
TW587241B (en) Automatic accord identification and generation method
Yanchenko Classical music composition using hidden Markov models
Ramirez et al. Automatic performer identification in celtic violin audio recordings
Cheng Music information retrieval technology: Fusion of music, artificial intelligence and blockchain
CN115331648A (en) Audio data processing method, device, equipment, storage medium and product
Papadopoulos Joint estimation of musical content information from an audio signal
Qi et al. [Retracted] Music Singing Based on Computer Analog Piano Accompaniment and Digital Processing for 5G Industrial Internet of Things
You et al. Music tonality detection based on Krumhansl-Schmuckler profile
KR102227415B1 (en) System, device, and method to generate polyphonic music
Wu et al. Generating detailed music datasets with neural audio synthesis
Pauwels Exploiting prior knowledge during automatic key and chord estimation from musical audio
Qiu et al. Humming2Music: Being A Composer As Long As You Can Humming.
Zhang Mobile music recognition based on deep neural network
Gupta et al. NUS Speak-to-Sing: A Web Platform for Personalized Speech-to-Singing Conversion.

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees