Musical note extraction algorithm
This report investigates on the transcription of polyphonic musical signal. Onset time is determined to divide the whole signal into segments for better recognition ofnotes. Constant QTransform (CQT) is then used to transform each segment from time domain to frequency domain. A threshold level is...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2016
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/68996 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | This report investigates on the transcription of polyphonic musical signal. Onset time is
determined to divide the whole signal into segments for better recognition ofnotes.
Constant QTransform (CQT) is then used to transform each segment from time domain
to frequency domain. A threshold level is set to find the fundamental and harmonic
frequency peaks. The kcq values that correspond to these peaks are generated and stored
for latter recognition process. Recognition methods such as Top-down Analysis and
Tone Model method are implemented to detect the possible note. Piano music, such as
Twinkle Twinkle Little Star and Skip to my Lou are tested. The recognition of onemember
score and two-member score give 100% successful detection. These results are
presented in term of the identified notes and their duration and loudness. Further
investigation on other instruments, such as flute and guitar are tested. Polyphonic music
played using both guitar and piano is analyzed too. Perfect detection once again
confirms the accuracy of this recognition algorithm. |
|---|