Hi
working on a music mining project that requires me to extract feature using *Mel-frequency cepstral coefficients*.
Now, i need advice on how should i create my database to store all such extracted features ?
For understanding, i will have features such as this:
1. Estimated Tempo
2. Spectral flatness
3. Prominent tones across spectral bands
4. bandwidths for identifying Instruments used per track
My goal is to extract all the instrument waveforms in a music, into separate tracks. So, that if i want i can select specific insturments and play only those.
Please suggest what sort of a database option are available? How do i build it.
Any pointers/ papers related to this will also be appreciated.
I thank you for your time in advance.
Database for storing features extracted from audio waveform
Moderator: kcleung