哼唱選歌(Query by Singing/Humming, QBSH)為音樂檢索的領域中相當重要的應用，是一種將使用者哼唱的人聲作為搜尋的目標，再從資料庫中搜尋出最符合使用者哼唱的歌曲的技術方法。基於本研究對使用者經驗(User experience)的了解，現今使用者已經將使用服務的觀點從需要安裝軟體至本機端電腦的概念轉變為偏好透過網際網路使用服務的概念。本研究主要是以本實驗室舊有的QBSH辨識核心為基礎，建置一個全新的使用介面以讓使用者不需要複雜地安裝套件軟體即可透過網頁服務(Web Service)來使用QBSH的服務。除此之外，本研究亦結合了Web2.0的概念，提供了一個能夠讓使用者加入新的歌曲至歌曲資料庫的使用流程，使用者不再只是內容搜尋者的角色，也是內容的提供者。
　　本研究的第二個目的在於優化系統的運作架構與流程上的改良。雖然本實驗室所開發的QBSH辨識核心相當穩健，但由於整體系統運作的可靠度較難掌握，目前仍有許多問題存在於伺服器之間的溝通與後端歌曲資料庫的同步等環節。本論文將會提出在伺服器之間如何利用系統排程達到同步的效果，以及運用控制伺服器的時間點來解決伺服器可能隨時離線或是新加入的狀況。

QBSH (Query by Singing/Humming) system is a way for music retrieval. QBSH system takes the tune from user’s singing or humming as the search target, and then the system returns the most similar song to the user. Based on the understanding of user experience, users tend to use web-based services rather than to have software installed to their computers. The goal of this research is to build a web-based QBSH system based on the existing QBSH core developed in our lab. This allows the users to use our QBSH system through the Internet without a complicated installation process. In addition, this system incorporates the concept of Web 2.0 and allows the users to add new songs to the song database. Thus users do not only search for songs but are also providers of the contents.
The second goal of this research is to optimize the structure and the operating process of the QBSH system. Although the core of melody recognition is quite robust, the reliability of QBSH system is still hard to control. Problems exist in the communication among the servers and in song database synchronization. So this research will apply system scheduling to achieve the goal of synchronizing all servers’ status. Also, severs will be allow to be new adding or into offline anytime by controlling servers’ time.

[1] J.-S. Roger Jang, Jiang-Chun Chen, Ming-Yang Kao, “MIRACLE: A Music Information Retrieval System with Clustered Computing Engines”,2nd Annual International Symposium on Music Information Retrieval 2001, Indiana University, Bloomington, Indiana, USA, October 2001.
[2] Jiang-Chun Chen, J.-S. Roger Jang, “Parallel Processing of Content-based music Retrieval”, MS Thesis, National Tsing Hua University, Taiwan, 2001.
[3] Hsao-Siung Chiu, J.-S. Roger, “Research and Implementation of Melody Recognition on Clustered PCs”, MS Thesis, National Tsing Hua University, Taiwan, 2008.
[4] Jang, J.-S. Roger, Hong-Ru Lee, Ming-Yang Kao, “Content-based Music Retrieval Using Linear Scaling and Branch-and-bound Tree Search”, IEEE International Conference on Multimedia and Expo, Waseda University, Tokyo, Japan, August 2001.
[5] Jang, J.-S Roger, Ming-Yang Kao, “A Query-by-Singing System based on Dynamic Programming”, International Workshop on Intelligent Systems Resolutions (the 8th Bellman Continuum), PP. 85-89, Hsinchu, Taiwan, Dec 2000.
[6] Adobe Flash Player, http://zh.wikipedia.org/zh-tw/Adobe_Flash_Player#Flash_Player.E7.BD.B7.E5.B7.A5
[7] Lawrencer R. Rabiner, "On the Use of Autocorrelation Analysis for Pitch Detection", IEEE Trans. ASSP, vol. 25, pp. 24-33, Feb 1977.
[8] Android, http://blog.chinatimes.com/tomsun/archive/2010/08/31/533666.html
[9] Midomi, http://www.midomi.com/
[10] Tunebot, http://tunebot.cs.northwestern.edu/
[11] Musiceline, http://musicline.de/
[12] Jang, J.-S. Roger, 線上教材：Audio Signal Processing and Recognition (音訊處理與辨識). http://mirlab.org/jang/books
/audioSignalProcessing/, last accessed 17 Sep 2010
[13] Youtube, http://www.youtube.com