1285880 九、發明說明： 【發明所屬之技術領域】 本發明係揭示-種可依據使用者對音樂喜好自動作曲產生音 未之方法與系統，尤指—種可自動作曲的方法及系統。 【先前技術】 音樂是人類生財不可或缺的元素，在日常生活的各個層面 都可接觸到大1的音樂。過往音軸作屬於專業性藝術創作領 ^，需要受過專業鱗之創作者方關作出相#程度受人喜愛的 音樂作品。目前具有樂音產生能力的電子設備好，但均僅具備 播放或廣奏I力’朗者只驗_人心、智活動演奏音樂，對於 未受過音料__人，並不能提供自動作㈣功能。、 P未受過音料業观之平fA即使喜歡音樂，也相為不熟 =疋律、即奏、音準及各種音樂風格，更*瞭解音樂創作之格式、 。法口而會有著無法創作屬於個人音樂的遺憾。當前*於各種電 毛達即便未叉過音樂專業訓練之平常人，現在要 種樂器或具有產生音縣力之電子設備也是㈣㈣，2 ^各種樂II或具有產生音親力之電子設備，並不意味可以創= 謂「手 \己的9 ▲舉例來說，現在的行動f話内乡麟有所言 =鈴聲韻功I」’但是對於未受過音料業訓賴人來說，幾 無法憑以行於自己的音樂或手機鈴聲。 5 1285880 因此’實在有必要提供—種新的可自動作曲的方法與系統， 内建有可自動學習使用者對各種音樂旋律、節奏、風格之偏好的 =系統軟體系統，並依據使用者個人喜好回饋資訊，自動調整 音樂之旋律、節奏、風格，協助使用者自動作曲並產生音樂，以 產生屬於使用者個人的音樂。 【發明内容】 本發明之主要目的係提供—種可依據使用者對音樂喜好自動 作曲產生音樂之方法與系統，係包含—微處理器，—記憶裝置， -輸入裝置’―音樂產生I置以及—可自動作曲產生音樂之人工 ^慧專豕純軟體系統，可產生各種不同旋律、節奏、風格之音 樂’並依據使用者個人喜好_資訊，自_整音樂之旋律、節 奏、風格，以產生屬於使用者個人的音樂。 本發明之另—目的係在提供建有可自動依據使用者對 不同旋律、節奏、風格音樂個人喜好_=#訊，自鱗習並作曲 產生音樂之專家系統軟體系統之電子設備系統，解決未受過音樂 專業训練者無法自行作曲產生音樂的習知技術缺點。 、 本發明之可自動作曲產生音樂之人工智慧專家系統軟體系 統，依功能區別’包含兩個主要的部分：—為演化計算運作模組， 另―則為音樂資料整合餅分模組。其巾，「演化計算運作模組」， 6 1285880 為目前習知之演化計算技術實作。可能使用的演化計算技術，包 括：基因演算法（Genetic Algorithms)、基因規劃法（Genetic1285880 IX. Description of the Invention: [Technical Field] The present invention discloses a method and system for automatically composing a sound according to a user's preference for music, and more particularly to a method and system for automatically composing music. [Prior Art] Music is an indispensable element of human wealth, and it can be accessed at all levels of daily life. In the past, the sound axis was a professional art creation, and it was necessary to be the creator of the professional scales to make a musical piece that was loved by the degree. At present, the electronic equipment with the ability to generate musical tones is good, but all of them only have the ability to play or play the role of the singer. The singer is only _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ P has not received the view of the audio industry. Even if you like music, it is also unfamiliar = law, sonar, pitch and various musical styles, and more * understand the format of music creation. There is a regret that it is impossible to create personal music. At present, in the ordinary people who have not crossed the music professional training, it is now necessary to plant musical instruments or electronic devices that produce sounds of the county. (4) (4), 2 ^ various music II or electronic devices with sound and power, and It doesn't mean that you can create = "hands and your own 9 ▲ For example, the current action f words inside the town has some words = ringing rhyme I" 'but for those who have not been trained in the music industry, With your own music or ringtones. 5 1285880 Therefore, it is really necessary to provide a new method and system for automatic composition, built-in system software system that can automatically learn the user's preferences for various music melody, rhythm and style, and according to the user's personal preferences. Reward information, automatically adjust the melody, rhythm, style of music, assist users to automatically compose music and produce music to produce music belonging to the user. SUMMARY OF THE INVENTION The main object of the present invention is to provide a method and system for automatically composing music according to a user's preference for music, including a microprocessor, a memory device, an input device, a music generation device, and - Automatically composing music to produce music ^ Hui specializes in pure software system, can produce a variety of different melody, rhythm, style of music 'and according to the user's personal preferences _ information, from the _ music melody, rhythm, style, to produce Music belonging to the user. Another object of the present invention is to provide an electronic device system with an expert system software system that automatically generates music based on the user's personal preference for different melody, rhythm, and style music, and self-study and composing music. Those who have received professional training in music cannot make their own technical shortcomings in producing music. The soft system of the artificial intelligence expert system capable of automatically composing music to produce music according to the present invention comprises two main parts: - an evolutionary computing operation module, and a music component integration cake module. Its towel, "Evolution Computational Operation Module", 6 1285880, is the current practice of evolutionary computing technology. Possible evolutionary computing techniques, including: Genetic Algorithms, Genetic Programming (Genetic)
Programming)、演化規劃法（Evolutionary Programming)......等 荨。此類演化計算技術，主要為一般用途（general purpose)之最 佳化演算法，其特色在於將欲處理之問題或函數，視為未知細節 之黑盒（blackbox)以進行最佳解之搜尋。另一主要模組r音樂資 料整合與評分模組」，則負責處理本專利所特有之功能部分。 籲 依功能區分，此模組可更進一步分為：樂音資訊表現 (representation)模組與樂曲評分模組。樂音資訊表現模組負責儲 存「樂音區塊（musicblocks)」以及此樂音區塊之評分資訊。 本專利中之一個樂音區塊可包含一或多個音符以及各項基本 資訊，如音高、音長……等，在演算處理的過程中，視為一基本 單位，有別於其他在已知文獻中所使用之以單一音符本身為基本 鲁單位之表現方式。同時，樂音資訊表現模組中，可選擇性地置入 樂音區塊生成、結合、與分離的各種運算子（〇perat〇r)以對_ 中_塊進行整合減理，例如生成音料贼完縣曲等。而 、樂曲評分模組，則為使用者對樂曲主觀評分之模組，其功能主要 為播放樂曲讓使用者魏、讓使用者表達對樂曲崎分、以及將 分數送回樂音資訊表現模組中，成為該模組中所儲存之各個樂音 區塊資訊的-部分。個者對樂曲之評分原則上為針對整個音樂 片段，但可選擇性地加人即時評分及與即日細饋功能，亦即僅對 7 1285880 目前正在齡之樂音區定評分；以及個者主触行樂音編 修之功能。 此專家祕軟m運賴程，基本上先由樂音資訊表現 模組’以亂數或風格導向（style 〇細ed)之方式，產生出音樂取 樣。再經由音樂㈣整合與評分模財之料評分模組，取得使 用者對所產生之音樂取樣之域評分。此時，雌音:#訊表現模 組所儲存㈣对’已具有音縣料基本㈣和使时對各音 樂取樣的評分。將現有資訊，導人—般騎之演化計算運作狗且， 例如：基因演算法’以演化計算之方法，將目前所有的音樂取樣 進行-個世代（genemtion)之演化步驟。频果將會_至樂音 資訊表現做’再重覆上叙流程，甚至設定之終止條件為止7 例如：使用者主動停止整個人化音樂之產生程序。 由於演化計算之各項相關技術，已經各個工程及科學領域， _乏糊於進行最佳化任務，至少包括：藥物設計、基因序列、 演化硬體及⑼、頻道分派、VLSI設計、賴設計、甚至股市預 測...等等。綜觀如此不同的領域，演化計算之相關技術，做為一 般用途的最佳化方法與卫具，其可行性及多樣性，就目前已知的 情形而言，乃毋庸置疑。而在所有使用演化計算方法的論文及應 ?中’演化計算方法之效能超越職l〇msearch亦為不爭之事實： 是故本專_演化計算運作池，來做騎佳化工具來使用。 據許多領域朗的絲，其魏也非mdQm s_h所鮮其項背1 8 1285880 因此，本發明採取廣為使用於各工程與工程領域的演化計算 最佳化工具，以及對樂音資料的特別處理模組，針對使用者主觀 評分之樂音資料進行最佳化搜尋與運算，應可產生使用此系統之 使用者所偏好之個人化音樂或樂曲片段。並可用於各項可發出聲 音之a又備與器具，例如：手機、鬧鐘、個人網頁播放音樂…等等。 本發明中另一功能區塊一音樂資料整合與評分模組。雖然如 刖所述，我們使用了 generalpurpose的最佳化演算法來進行搜尋 之任務，但要讓該計算模組適當地發揮能力，必須針對此系統的 特〖生·如產生之目的物為樂曲片段（資料特性）、與使用者互動需 考慮非電翻素（例如：料疲勞、注意力無法射過久·…） 等’發展出並非-般直接（straightf〇贿d)做法所能提供之資料呈 現方式（data representation)與知識系統㈣論喂s 能達到計射之目標。單此部分設計，已非目前熟知技術。 而另邛刀，則疋為了提供使用者自行依喜好產生樂曲片段 之服務，而將兩大魏區塊以特殊設計之方式接合，所呈現出之 系統整體運作流程。整個系統結構的整合部分，其功能區塊間互 動運作之流程本身轉目前習知的技術與設計，故為專利發明之 另一主要標的。 因此為達成上述之目的，依據本發明之一特色，即一種可 動作曲之方法’使用於一可自動作曲的系統中，該可自動作曲 9 1285880Programming), Evolutionary Programming, etc. 荨. Such evolutionary computational techniques, primarily optimized for general purpose, are characterized by treating the problem or function to be treated as a blackbox of unknown detail for the best solution search. Another major module, the Music Data Integration and Scoring Module, handles the functional parts unique to this patent. According to the function distinction, this module can be further divided into: music information representation (representation) module and music score module. The Music Information Performance Module is responsible for storing the "musicblocks" and the rating information of this music block. A music block in this patent may contain one or more notes and various basic information, such as pitch, length, etc., in the process of calculation processing, regarded as a basic unit, which is different from others. It is known in the literature that the single note itself is the expression of the basic unit. At the same time, in the music information presentation module, various operators (〇perat〇r) for generating, combining, and separating the music block can be selectively placed to integrate and reduce the _zhong_block, for example, generating a thief Finish the county song and so on. The music score module is a module for subjective scoring of the music, and its function is mainly to play the music to the user, let the user express the music, and return the score to the music information performance module. , becomes the part of the information of each tone block stored in the module. The score of the music is in principle directed to the entire music piece, but can be selectively added to the instant score and the day-to-day search function, that is, only the score of the current age of 7 1285880; and the main touch The function of editing music. This expert secret soft m journey, basically by the music information display module 'in a random or style-oriented (style 〇 fine ed) way to produce music samples. Then, through the music (4) integration and scoring model money scoring module, the user is scored on the domain of the generated music sampling. At this time, the female voice: # 讯 模 所 储存 储存 四 四 四 四 四 四 四 四 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The existing information, guided by the evolution of the rider, and the operation of the dog, for example, the genetic algorithm's method of evolutionary calculation, the current sampling of all the music is carried out - a generational evolution step. The frequency will be _ to the music information performance to 're-repeated the process, even set the termination conditions. 7 For example: the user actively stops the whole humanized music generation process. Due to the related technologies of evolutionary computing, various engineering and scientific fields have been optimistic, including at least: drug design, gene sequence, evolution hardware and (9), channel assignment, VLSI design, Lai design, Even stock market forecasts...etc. Looking at such different fields, the related technologies of evolutionary computing, as a generalized optimization method and sanitation, its feasibility and diversity are undoubtedly known in the current situation. In all the papers using the evolutionary calculation method and the performance of the 'evolutionary calculation method, it is an indisputable fact: it is the original _ evolutionary computing operation pool, to use the riding tool. According to many fields, the silk is not the same as mdQm s_h. 1 8 1285880 Therefore, the present invention adopts an evolutionary optimization tool widely used in various engineering and engineering fields, and special processing of music materials. The module, which optimizes the search and calculation of the user's subjectively rated music data, should produce personalized music or music pieces preferred by the user using the system. It can also be used for a variety of sounds and equipment, such as mobile phones, alarm clocks, personal web pages to play music, etc. Another functional block in the present invention is a music data integration and scoring module. Although we have used the generalpurpose optimization algorithm to perform the search task, as long as the computing module is properly functioning, it must be targeted for the purpose of this system. Fragments (data characteristics), interaction with users need to consider non-electrical factors (for example: material fatigue, attention can not be shot too long ...), etc. 'developing is not straightforward (straightf bribe d) practices can provide The data representation and the knowledge system (4) can achieve the goal of counting. This part of the design is not currently well known. In addition, in order to provide users with the service of generating music pieces according to their own preferences, the two Wei blocks are combined in a special design manner, and the overall operation process of the system is presented. The integrated part of the overall system structure, the process of interactive operation between functional blocks itself is transferred to the current known technology and design, and is therefore another main target of the patented invention. Therefore, in order to achieve the above object, according to one feature of the present invention, a method of actionable music is used in an automatically composable system, which can be automatically composed 9 1285880
可自動作曲產生音樂之專家系統軟«統重複擷取多個非特定之 #曰秦取樣數位貝料’並整合成為多個使用者整合評價資料；⑷ 可ί動作曲產生音樂之專家系統軟體系統依據多個使用者整合評 價貝料’產生有效之輸人、輸出的歸屬函數運算結果；⑹可自 動作曲產生音樂之專家系統軟體祕自記憶裝置中擷取一與該有 奴輸入、輸出的歸屬函數運算結果最近似之音樂取樣數位資 料在不偏離原有效之輸入、輸出的歸屬函數運算結果之集合内， 现數方式對錄近似之音綠樣數位資·機律、速度、節 奏風格之隨機修改，產生第i個隨機作曲結果音樂取樣數位資 =並傳送至音樂產生裝置；⑻可自動作曲產生音樂之專家系 統軟㈣統接受使用者儀輪人裝置輸人之個人賴第1個隨機 ‘:曲結果音樂取樣數位資料之評價分數；⑺可自動作曲產生音 y、專豕糸統車人體糸統利用該第1個隨機作曲結果音樂取樣數位 ί料之5平價分數產生新的有效之輸入、輸出的歸屬函數運算結 果，（8)可自動作曲產生音樂之專家系統軟體系統依據步驟（5) 1285880 至⑺之方法，依序產生第2個至第N個隨機作曲結果音樂取樣 數位貝料’並逐次修正有效之輸人、輸出畴屬函數運算結果， 直到使用者透過輸人裝置輸人—停止輪人訊號為止，·以及，（9) 可自動作曲產生音樂之專㈣統可在接受到該停止訊號後，將該 第N個隨機作曲結果音樂取樣數位資料儲存於記憶裝置中，以完 成自動作曲。 依據本發明之另-特色，係提供—種可自動作曲的系統用以 自動作曲。本發明之可自動作曲的系統，包括：—微處理器，主 要用以進仃運异，-記憶錢，主要用⑽存音樂取樣數位資料； -輸入裝置’主要用以輸人之音樂取樣評價分數卜音樂產生裝 置’主以接受音樂取樣數位資料，並產生可供使用者跨 音樂；以及-可自動作曲產生音樂之專家系統軟體系統，主· 以將音樂雜評齡數触簡之音躲樣之雕、卿 節奏、風格㈣資料整合被用者整合簡㈣，並依據輪入又、、 輸出的歸屬函數運算結果進行自動作曲。 ’ "由於本發_造_，能提供錢上_，並卻能 技術缺失，故依法申請專利。 豸知 1285880 【實施方式】 為能讓貴審查委員能更瞭解本發明之技術内容，特舉一較 佳具體實施例如下。 有關本發明之可自動作曲的方法及彡統之較佳實施例，請先 參閱第1圖之系統架構圖。本例係應用於一資訊系統中，用以提供 使用者自動作曲服務。第，圖顯示本例主要包括有一可自動作曲產 癱生音樂之專家系統1〇、一音樂取樣資料庫11、樂音資訊表現模組 12、-音樂資料整合與評分模組13、一音樂產生裝置加、音勒、 一輸入裝置3G、-使用者音樂取樣評價分數31以及—作曲結果4〇。 本例之了自動作曲的方法及系統係具備接受使用者音樂取樣 評價並自動作曲功能，因此可自動作曲產生音樂之專家系統_ 由樂音資絲麵組12、音樂㈣整合鱗分顯1細主要部分 構成，其中可自動作曲產生音樂之專家系統1〇中之樂音資訊表現 #模組12可自音樂取樣資料庫Μ中擷取某一非特定之音樂取樣，傳 送至音樂產生裝置20以產生音樂21。使用者在義完可自動作曲 產生音樂之專家系統10隨機選取之音樂21後，可透過輸入裝置3〇 輸人-使用者音樂取樣評價分細。可自動作曲產生音樂之專家 系、_在接受到使用者音樂取樣評價分數31後，即將使用者音樂 取樣許價錄31奴音料健合鱗分额13。音樂資料整合 與評分模組13在接受到使用者音樂取樣評價分數⑴後，即將此使 12 1285880 用者音樂取樣評價分數31與被評價之音樂取樣之調性、旋律、速 度、節奏、風格索引資料整合為使用者整合評價資料5〇。重複進 行上述步驟後，音樂資料整合與評分模組13將多筆使用者整合評 價身料50導入可自動作曲產生音樂之專家系統1〇一般用途之演化 計异運作模組，以演化計算之方法，將目前所有的音樂取樣進行 一代（generation)之演化步驟，產生有效之輸入、輸出的歸屬函 數運算結果。其結果將會回饋至樂音資訊表現模組12，樂音資訊 Φ 表現模組彳2可在產生之有效之輸入、輸出的歸屬函數運算結果範 圍内，隨機產生音樂取樣，再重覆上述之流程，逐次產生新的輸 入、輸出的歸屬函數運算結果，並依據新的輸入、輸出的歸屬函 數運算結果，隨機產生新的音樂取樣，直到使用者主動停止整個 人化音樂之產生程序為止。 接下來請參閱第3圖關於本發明使用者音樂取樣評價分數之 實施例。使用者在聆聽完音樂產生裝置2〇所播放之音樂21後，可 •依據個人主觀偏見，依據第3圖實施例中所示之選單指示，利用輸 入裝置30輸入使用者個人對該音樂21之使用者音樂取樣評價分數 31。此外，在另一實施例中，使用者可以在無顯示裝置顯示如第3 ‘ 圖之選單提示晝面下，直接依據輸入裝置30所註記之評價分數文 字’直接利用輸入裝置3〇輸入使用者個人對該音樂21之使用者音 樂取樣評價分數31。 接下來請一併參閱第4圖關於本發明關於使用者整合評價資 13 1285880 =之實施例。可自動作曲產生音樂之專家系觸在接受到使用者 =取樣評價分數31後，即將使用者音樂取樣評價分數31送交音 樂資料整合與評分模組13。音雜㈣合與評分额13在接受到 使用者音樂取樣評價分數31後，㈣此使用者音練樣評價分數 31與被評價之音樂取樣之雛、旋律、速度、節奏、風格索引資 料整σ為使用者整合評價資料5Q。使用者整合評價資料如中有關 =樂取樣之調性、旋律、速度、節奏、風格等#料均依據系統預 :方式編製索引，如：C大調調性為1、降D大調雛為2、速度莊 嚴為1—、行版速度為2、民謠風格為]、節奏藍賴格為2.·…等等。 由^母-音樂取樣皆可依據上述諸音樂躲編織引，亦可視為 數學上之乡魏，在與使用者细裝獅輸人之個者音樂取樣 A貝刀數31結合後’即成為—使用者整合評價資料5〇。 以下明參考第2圖關於本發明之流程圖，以便了解如何達到自 動作曲之技術。 步驟S201 : 自動作曲產生音樂之專家系統1()之樂音資訊表現模組12於微 處理為啟動後自動自輯裝置音縣樣資料11愤麵取某一 非特疋之音樂取樣數位資料，並且同時讀取該音樂取樣資料之調 性、旋律、速度、節奏、風格等索引資料。當然在另一實施例中， 自動作曲產生音樂之專家系統10可產生一選單並顯示於一顯示 1285880 器，使用者可以透過輸入數值碼或點選選單的方式，預先選取音 樂取樣數位資料的索引類型，如··民謠風格、節奏藍調風格等等， 以限縮自動作曲產生音樂之專家系統1〇隨機擷取音樂取樣數位資 料之範圍。 步驟S202 : 自動作曲產生音樂之專家系統1〇之樂音資訊表現模組12傳送 φ 至音樂產生裝置20產生音樂21以供使用者聆聽評價。 步驟S203 : 使用者跨聽後輸人裝置3Q輸人_音樂雜之評價分數 31。請-併參考第3圖之係依據本發明關於用者音樂取樣評價分數 之實施例。此時使用者可依據個人對該音樂取樣之喜好，選取符 合自己喜好鱗價分數。在另—實施财，此處之輸人裝置可以 #為鍵盤、滑鼠或觸控式螢幕等或其他可供使用者輸入評價分數之 輸入裝置；_人之簡分數除直接讀值方式呈料，亦可以 選單、問卷等習於該項記憶者所熟知的方式進行之。 步驟S204 : 可自動作曲產生音樂之專家系統娜該音樂取樣之評價分數 31送交音㈣職合與評分·彳3。音樂:#合與評分模组η 在接受咖者音樂取樣則_31後，梅_者音樂取樣 15 1285880 評價分數31與被評價之音樂取樣之調性、旋律、速度、節奏、風 格索引資料整合為使用者整合評價資_。請—併參考第4圖係依 據本發明關於使用者整合評價資料之實施例。 步驟S205 : 可自動作曲產生音樂之專家系統10之樂音資訊表現模組彳2重 複前述步驟S201至步驟S202之步驟程序n次，並播放n段音樂21 φ 供使用者聆聽評價。 步驟S206 : 可自動作曲產生音樂之專家系統1〇依次重複接受使用者问個 音樂取樣評價分數31，並送交音樂資料整合與評分模組13整合成 為N個使用者整合評價資料5〇。 步驟S207 : 可自動作曲產生音樂之專家系統1〇依據N個使用者整合評價 貝料50產生有效之輸入、輸出的歸屬函數運算結果。在本實施例 中係利用基因演算法（Genetic Alg〇rithms)為例，請一併參考第4 圖係依據本發_於制者整合評價㈣之實_。在本實施例 中所有使用者整合評價資料5〇可視為一多為矩陣，可自動作曲 產生音樂之專家系統1〇即以基因演算法^神㈣對 此矩陣求最佳解並同時求得函數範圍。 16 I285880 步驟S208 : 可自動作曲產生音樂之專家系統1〇之樂音資訊表現模組12自 .動自記織置音樂取樣資料庫彳彳㈣取—與該最佳解最近似之音 • 樂取樣數位資料。 步驟S209 : 可自動作曲產生音樂之專家系統1〇之樂音資訊表現模組12在 不偏離原有效之輸人、輸畴屬函數結果集仙，以亂數對 該最近嫌最佳解之音樂取樣數位資料進行隨機修改，產生第_ 隨機作曲結果音樂取樣數位資料。 步驟S210 : 可自動作曲產生音樂之專㈣統购音資絲現模組12將隨 機作曲結果音樂取樣數位資料傳送至音樂產生裝置2〇產生音樂 • 21。 步驟S211 : 使用者聆聽後利用輸入裝置30輸入對該隨機作曲結果音樂取 樣數位資料之評價分數31。 步驟S212 : 如步驟S204，可自動作曲產生音樂之專家系統1〇音樂資料整 17 1285880 價r與被評價之隨機編糊取 機作喊果整合評㈣袖t、風齡㈣㈣合為第1使用者隨 步驟S213 : 如步驟S2G7可自動作曲產生音樂之專㈣、_依據前述步 驟伽之N個使用者整合評價資卿與此處請用者隨機作曲 •結果整合評價資料50以基因演算法（Genetic Aig〇rithms)產生新 的最佳解以及函數範圍。 步驟S214 : 如步驟S209，彳自動作曲產生音樂之專家系統1〇樂音資訊表 現杈組12在不偏離新的有效效之輸入、輸出歸屬函數運算結果集 合内，以亂數對該最近似新的最佳解之音樂取樣數位資料進行隨 • 機修改，產生第2個隨機作曲結果音樂取樣數位資料。 步驟S215 : 使用者跨聽第2個隨機作曲結果音樂取樣數位資料透過音樂 產生裝置20產生之音樂21後’利用輸入裝置3〇輸入對該隨機作曲 結果音樂取樣數位資料之評價分數31。 步驟S216 : 1285880 依據S212方式’可自動作曲產生音樂之專家系統1〇音樂資料 整合與評分模組13將該評價分數31與被評價之隨機作曲結果音樂 取樣之調性、旋律、速度、節奏、風格索引資料整合為第2使；者 隨機作曲結果整合評價資料5〇。 步驟S217 : 可自動作曲產生音樂之專家系統10音樂資料整合與評分模組 • 13依據步驟S213之方式依據N個使用者整合評價資料與第彳及第2 使用者隨機作曲結果整合評價資料以基因演算法（Genetic Algorithms)求得最佳解及函數範圍。 步驟S218: 依據前述步驟S214至步驟S217第2個隨機作曲結果音樂取樣 數位資料產生方式，逐次產生第3個到第N個隨機作曲結果音樂取 φ 樣數位資料’並逐次以基因演算法（Genetic Algorithms)求得最佳 解及函數範圍。 步驟S219 : 當使用者連續聆聽多個依據前述步驟218所產生之隨機作曲 音樂取樣數位資料所產生之音樂21後，可透過輸入裝置3〇輸入一 停止輸入訊號。可自動作曲產生音樂之專家系統10在接受到此一 停止訊號後即將第N個隨機作曲結果音樂取樣予以儲存，並完成自 19 1285880 動作曲 綜上所述，本發明無論就目的、手段及功效，在在均顯示其 迥異於習知技術之特徵，實為一極具實用價值之發明，想請責 審查委㈣察，早日縣專利，俾減社會，實感德便： 意的是，上述諸多實施例僅係為了便於朗碎例 二 所主張之權利朗自應以申料職_述 ，杈明 述實施例。 “、、而非僅限於上 【圖式簡單說明】 第1圖係本發明之系統架構圖。 第2圖係依據本發明之流程圖。 第3圖係依據本發明關於用者音樂取樣評價分數之實施^ •第4圖係依據本發明關於使用者整合評價資料之實施例歹 【主要元件符號說明】 可自動作曲產生音樂之專家系統1〇 音樂取樣資料庫11 樂音資訊表現模組12 音樂資料整合與評分模組13 20 1285880 音樂產生裝置20 音樂21 輸入裝置30 使用者音樂取樣評價分數31 作曲結果40 • 使用者整合資料50The expert system that can automatically compose and produce music is soft and repetitively captures a number of non-specific #曰秦sampled digital materials and integrates them into multiple users' integrated evaluation materials; (4) can be used to generate music expert software system According to the evaluation of multiple users, the results of the function of the input function of the input and output of the beetle are generated. (6) The expert system of the automatic composing music can be used to capture the attribution of the slave input and output. The most similar music sampling digital data of the function operation result is in the set of the result of the attribution function of the input and output which are not deviated from the original effective, and the current mode is recorded as the approximate sound green sample number, the law, the speed, the rhythm style random Modify, generate the i-th random composition result music sample digits = and transfer to the music production device; (8) can automatically compose the music to produce the expert system of the soft (four) system accepts the user's instrument wheel device to lose the individual's first random ' : the result score of the music sampled digital data; (7) can automatically compose the sound y, the special car body system uses the 1 random composition result music sampling digits 5 parity scores to generate new valid input and output attribution function operation results, (8) automatic composing music production expert system software system according to steps (5) 1285880 to (7) The method sequentially generates the second to the Nth random composition result, and the music samples the digital data, and sequentially corrects the effective input and output domain function calculation results until the user inputs the input device through the input device. Up to, and, (9) can automatically compose music to produce music (4) can receive the stop signal, the Nth random composition result music sample digital data is stored in the memory device to complete the automatic composition. According to another feature of the present invention, a system for automatically composing is provided for self-operating music. The automatic composing system of the invention comprises: a microprocessor, which is mainly used for importing and transferring different things, - memory money, mainly using (10) storing music to sample digital data; - inputting device 'mainly used for inputting music sampling evaluation The score music generating device 'mainly accepts music to sample digital data and generates a software system for the user to cross music; and - can automatically compose music to generate music, the main system to hide the sound of the music number Sample carving, Qing rhythm, style (4) Data integration is integrated with the user (4), and automatically composes according to the result of the rounding and output function of the attribution function. ‘ " Because the hair _ _ _ can provide money _, but can be technically missing, so apply for a patent according to law.知知 1285880 [Embodiment] In order to enable the reviewing committee to better understand the technical contents of the present invention, a specific embodiment is described below. For a preferred embodiment of the method and system for automatic composing of the present invention, please refer to the system architecture diagram of Figure 1. This example is applied to an information system to provide users with automatic composition services. First, the figure shows that this example mainly includes an expert system for automatically composing and producing music, a music sampling database 11, a music information presentation module 12, a music data integration and scoring module 13, and a music generating device. Plus, sound, one input device 3G, - user music sample evaluation score 31 and - composition result 4 〇. In this example, the method and system for automatic composition is equipped with an expert system that accepts the user's music sampling evaluation and automatic composition function, so that the music can be automatically composed and produced. _ The music is composed of the music layer (12), and the music (4) is integrated with the scale. Part of the composition, wherein the expert system for automatically composing music can be used to capture a certain non-specific music sample from the music sampling database, and transmit it to the music generating device 20 to generate music. twenty one. After the user can automatically compose the music 21 randomly selected by the expert system 10 for generating music, the input device 3 can input the user-user music sample evaluation score. The expert who can automatically compose music to produce music, _ after receiving the user's music sampling evaluation score 31, will be the user's music sampling price record 31 slave sound material fitness scale score 13. After receiving the user music sampling evaluation score (1), the music data integration and scoring module 13 will make the 12 1285880 user music sampling evaluation score 31 and the evaluated music sampling tone, melody, speed, rhythm, style index. The data is integrated into the user's integrated evaluation data. After repeating the above steps, the music data integration and scoring module 13 introduces a plurality of user integrated evaluation bodies 50 into an expert system for automatically composing and generating music, and a general-purpose evolutionary computing module for evolutionary calculation. The current music sampling is performed in a generational evolution step to generate valid input and output attribution function calculation results. The result will be fed back to the music information presentation module 12, and the music information Φ performance module 彳2 can randomly generate music samples within the range of the generated input and output attribution function calculation results, and repeat the above process. The new input and output attribution function calculation results are successively generated, and new music samples are randomly generated according to the new input and output attribution function calculation results until the user actively stops the entire humanized music generation program. Next, please refer to Fig. 3 for an embodiment of the user's music sampling evaluation score of the present invention. After listening to the music 21 played by the music generating device 2, the user can input the user's personal music to the music 21 by using the input device 30 according to the personal subjective bias and according to the menu instruction shown in the embodiment of FIG. User music sample evaluation score 31. In addition, in another embodiment, the user can directly input the user by using the input device 3 directly according to the evaluation score text marked by the input device 30 under the menu prompt of the third display. The individual samples the evaluation score 31 for the user music of the music 21. Next, please refer to FIG. 4 for an embodiment of the present invention regarding user integration evaluation 13 1285880. The expert who can automatically compose the music to generate the music, after receiving the user = sample evaluation score 31, sends the user music sample evaluation score 31 to the music data integration and scoring module 13. After the user's music sampling evaluation score 31 is received, the user's music sample evaluation score 31 and the evaluated music sampling chick, melody, speed, rhythm, style index data σ Integrate evaluation data 5Q for users. The user's integrated evaluation data such as the tonality, melody, speed, rhythm, style, etc. of the music sampling are all indexed according to the system pre-: way, such as: C major tonality is 1, the major of D major is 2 The speed is solemnly 1 - the speed of the line is 2, the style of the folk song is], the rhythm blue lyric is 2.... and so on. The mother-music sampling can be based on the above-mentioned music to avoid weaving, and can also be regarded as the hometown of mathematics. In combination with the user's fine-packed lion input, the music sampling A shell number 31 is combined to become The user integrates the evaluation data 5〇. The flow chart of the present invention will now be described with reference to Figure 2 to understand how to achieve the technique of self-actuation. Step S201: The music information presentation module 12 of the expert system 1 () of the music generated by the action music automatically selects a non-special music sample digital data after the micro-processing is started, and automatically takes a non-special music sampling digital data. Read the index data of the tonality, melody, speed, rhythm, style, etc. of the music sample data. Of course, in another embodiment, the expert system 10 for generating music from the action music can generate a menu and display it on a display 1285880. The user can pre-select the index of the music sample digital data by inputting a numerical code or clicking a menu. Types, such as · folk style, rhythm and blue tone style, etc., to limit the automatic composition of the music to produce the expert system 1 〇 randomly take the range of music sampling digital data. Step S202: The music information presentation module 12 of the expert system for generating music from the action music transmits φ to the music generating device 20 to generate the music 21 for the user to listen to the evaluation. Step S203: The user crosses the input device 3Q to input the evaluation score 31 of the music. Please - and refer to Figure 3 for an embodiment of the evaluation scores for user music sampling in accordance with the present invention. At this time, the user can select the grading score according to his/her preference for the music. In another implementation, the input device here can be a keyboard, a mouse or a touch screen or other input device for the user to input the evaluation score; _ the simple score of the person except the direct reading method It is also possible to conduct menus, questionnaires, etc. in a manner familiar to the memory. Step S204: The expert system that can automatically compose music to generate music. The evaluation score of the music sample is 31. The voice is sent (4) and the score is 彳 3. Music: #合和评分 Module η After accepting the coffee sample sampling _31, Mei _ music sampling 15 1285880 Evaluation score 31 and the assessed music sampling tonality, melody, speed, rhythm, style index data integration Integrate evaluation resources for users. Please - and with reference to Figure 4, an embodiment of the user-integrated evaluation data in accordance with the present invention. Step S205: The music information presentation module 可2 of the expert system 10 for automatically composing music can repeat the steps of the foregoing steps S201 to S202 for n times, and play n pieces of music 21 φ for the user to listen to the evaluation. Step S206: The expert system for automatically composing music to generate music 1 重复 repeatedly accepts the user to ask for a music sampling evaluation score 31, and sends it to the music data integration and scoring module 13 to integrate the evaluation data for N users. Step S207: An expert system capable of automatically composing music to generate music. According to the N user integration evaluation, the bedding material 50 generates an effective input and output attribution function operation result. In this embodiment, a genetic algorithm (Genetic Alg〇rithms) is used as an example. Please refer to the fourth figure in accordance with the present invention. In this embodiment, all the user integrated evaluation data can be regarded as a matrix, and the expert system for automatically composing music can be used to obtain an optimal solution for the matrix and to obtain a function at the same time. range. 16 I285880 Step S208: An expert system for automatically composing music to generate music. The music information presentation module 12 is self-recorded and woven by the music sampling database. (4) Take the sound that is closest to the best solution. data. Step S209: The expert system for automatically composing music to generate music 1 is not subject to the original effective input and output function result set, and the music of the most recent best solution is sampled by random numbers. The digital data is randomly modified to produce the first _ random composition result music sample digital data. Step S210: Automatically composing music to produce music. (4) The unified audiophile module 12 transmits the random sample music sample digital data to the music generating device 2 to generate music. Step S211: The user inputs the evaluation score 31 of the music sample digital data of the random composition result by the input device 30 after listening. Step S212: In step S204, the expert system for automatically composing music is generated. 1 〇 music data, whole 17 1285880, price r, and the evaluated random splicing machine for the integration of the screaming fruit (four) sleeve t, wind age (four) (four) combined for the first use With step S213: as in step S2G7, the music can be automatically composed (4), _ according to the aforementioned steps, the N user integrated evaluation evaluation and the user's random composing and the result integration evaluation data 50 to the gene algorithm ( Genetic Aig〇rithms) produces new optimal solutions and function ranges. Step S214: In step S209, the expert system 1 for automatically composing music to generate music is displayed in the result set of the input and output attribution function without deviating from the new effective effect, and the most approximate new one is obtained by random numbers. The optimal solution of the music sample digital data is modified by the machine to generate the second random composition result music sample digital data. Step S215: The user crosses the second random composition result music sample digital data through the music 21 generated by the music generating device 20, and then uses the input device 3 to input the evaluation score 31 of the music sample digital data for the random composition result. Step S216: 1285880 According to the S212 mode, the expert system for automatically composing music can be used. The music data integration and scoring module 13 compares the evaluation score 31 with the evaluated random composition result, the tone, the melody, the speed, the rhythm of the music sampling. The style index data is integrated into the second place; the random composition result is integrated with the evaluation data 5〇. Step S217: The music data integration and scoring module of the expert system 10 for automatically composing music is generated. 13 According to the method of step S213, the evaluation data is integrated with the data of the N users and the randomized composition of the second user and the second user. The algorithm (Genetic Algorithms) finds the best solution and the range of functions. Step S218: According to the second random composition result of the foregoing step S214 to step S217, the music sample digital data generation manner is sequentially generated, and the third to the Nth random composition result music is taken to take the φ-like digital data successively and the gene algorithm is successively performed (Genetic) Algorithms) find the best solution and function range. Step S219: After the user continuously listens to the music 21 generated by the random composition music sampling digital data generated in the foregoing step 218, a stop input signal can be input through the input device 3. The expert system 10 capable of automatically composing and generating music, after receiving the stop signal, samples the Nth random composition result music and stores it, and completes the action summary from 19 1285880, the present invention regardless of the purpose, means and function In the show, it is different from the characteristics of the conventional technology, it is a very practical invention, I would like to ask the review committee (four) to check, the early county patent, reduce the society, the real sense of deli: I mean, many of the above The examples are only for the purpose of facilitating the claim of the second example, and the embodiments are described in the context of the application. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a system architecture diagram of the present invention. Figure 2 is a flow chart according to the present invention. Figure 3 is a sample evaluation score for user music according to the present invention. Implementation of the figure ^ Figure 4 is an embodiment of the user-integrated evaluation data according to the present invention. [Explanation of main component symbols] Expert system for automatically composing music to generate music 1 Music sampling database 11 Music information presentation module 12 Music data Integration and scoring module 13 20 1285880 Music generating device 20 Music 21 Input device 30 User music sampling evaluation score 31 Composing result 40 • User integration data 50
21twenty one