增强现实AR>> acoustic-signal-processing>> 返回
项目作者： trungd

项目描述 ：
Implementations of voice, music and other acoustic signal processing techniques with explicative GUI.
高级语言： Java
项目地址: git://github.com/trungd/acoustic-signal-processing.git
创建时间： 2017-10-12T07:06:07Z
项目社区：https://github.com/trungd/acoustic-signal-processing
开源协议：
下载


acoustic-signal-processing

Voice/Music acoustic signal processing techniques with explicative GUI.

Implementations

  • [ ] Hidden Markov Model with MFCC as feature vector for simple voice recognition

  • [x] Non-negative Matrix Factorization for Sound Separation

  • [x] Autocorrelation / Subharmonic Summation for pitch detection

Features

  • [x] Real-time or file input

  • [x] Display multiple charts as of entire period or specified frame length.

  • [x] Display calculation and recognition results: loudness, fundamental frequency, Japanese vowel prediction,…

Chart List

Waveform (current frame / entire period)

Waveform of entire period with marker indicating current position.

Waveform of current frame.

Spectrum

Spectrogram

Spectrogram with display of voiced / unvoiced period differentiated using zero-crossing rate and fundamental frequency.

Autocorrelation

Autocorrelation and fundamental frequency position (second peak)

Spectrum + Cepstrum (frame)

Display filter spectrum (default: lifter order 13) in comparison with original spectrum.

Recognition Result

Using saved training result to determine a / i / u / e / o period.

Libraries