- 學位論文
以微控制器(MCU)實現基於小波轉換之伴奏系統
Implementation of an accompaniment system based on DWT via microcontroller unit (MCU)
摘要
在音樂演奏中，節拍(BPM，beats per minute)是最基礎也最重要的元素，節拍不穩定就會嚴重的影響歌曲的呈現，使音樂雜亂無章，然而，大多數的伴奏系統或節拍器都是以事先設定好的固定節拍設計，無法根據演奏者的節拍自動調整並跟隨，因此，本研究提出了以微控制器(MCU)實現基於小波轉換之伴奏系統，分析吉他彈奏的節拍速度，使用離散小波轉換分析吉他訊號的時域與頻域特徵，計算節拍BPM數值，並在爵士鼓上架設自動打擊系統，跟隨吉他手進行伴奏，設計踏板開關讓吉他手根據演奏需求，自行切換打擊節奏組合，並透過手機APP即時顯示BPM數值，幫助使用者掌握彈奏的節拍狀況；在本研究中，選用低階與高階兩種不同的微控器進行實驗，低階是選用HT32F52352，高階則是STM32F746ZGT6，對數位訊號處理當中的取樣頻率與小波轉換進行實驗與比較，找出在不同取樣頻率下，何種多貝西小波函數(dbN)與分解階層數更適合本系統分析吉他節拍，最後實驗結果驗證256Hz的取樣頻率、db5多貝西小波函數與3階分解階層的分析效果最好。
並列摘要
In music performance, the beat (BPM, beats per minute) is the most basic and important element, and unstable rhythm will seriously affect the presentation of the song, making the music chaotic, so when learning a musical instrument, it is often used to train the stability of the beat. However, most accompaniment systems or metronomes are designed with a predetermined fixed beat and cannot automatically adjust to the player's beat and follow it. Therefore, this research proposes an accompaniment system based on DWT via microcontroller unit (MCU) ,that Analyze the tempo of guitar playing, use discrete wavelet transform to analyze the time domain and frequency domain characteristics of the guitar signal, calculate the beat BPM value, and set up an automatic percussion system on the jazz drum to follow the guitarist to accompaniment, design a pedal switch to allow the guitarist to switch the percussion rhythm combination according to the performance needs, and display the BPM value in real time through the mobile APP to help users grasp the rhythm of the playing; In this research, two different microcontrollers, low-level and high- level, are used for experiments. The low- level is HT32F52352, and the high-end is STM32F746ZGT6. The sampling frequency and wavelet transform in digital signal processing are tested and compared to find out the differences. Which Daubechies wavelet function (dbN) and decomposition level are more suitable for analyzing guitar rhythm and audio at the sampling frequency. The experimental result verifies that the sampling frequency of 256Hz, the db5 wavelet function and the 3level order decomposition level are the most effective signal processing methods. An additional pressure sensor was placed on the drums to monitor the percussion status, while the microphone radio signal was monitored to prove that the accompaniment system was able to keep up with the guitarist's tempo.
參考文獻
延伸閱讀
- An, C. M. (2019). 具有諧波抑制特性應用之微型化分支線耦合器設計. 明新學報, 43(2), 1-14. https://www.airitilibrary.com/Article/Detail?DocID=1728760X-201907-201907240010-201907240010-1-14
- 蕭俊祥、陳垣圻（2011）。應用小波函數於主動式噪音控制性能之改善。載於中華民國振動與噪音工程學會（主編），中華民國振動與噪音工程學會論文集（頁210-215）。中華民國振動與噪音工程學會。https://doi.org/10.30028/CSSVANNUAL.201106.0210
- 吳昱融（2016）。微小化、雙頻兼具諧波抑制180ﾟ分合波器之研製〔碩士論文，逢甲大學〕。華藝線上圖書館。https://doi.org/10.6341/fcu.M0328881
- 宋旻達（2012）。Internet-Based Remote Programmable Embedded Micro-Controller System Design And Implementation〔碩士論文，元智大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0009-2801201415014991
- 葉昱廷（2016）。Design and Implementation of a Bidirectional Resonant DC Converter〔碩士論文，國立交通大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0030-0803201714350861