SMD-DEMO is a Ruby command-line audio feature evaluation framework for ZCR and CFA that provides results analysis from imported audio file(either MP3 or M4A). For mixed audio(i.e. Speech and Music), it also provides a web-based visualization using Peaks.js developed by BBC R&D (https://github.com/bbcrd/peaks.js). This is a MSc Project in 2014 Summer, supported by UCL CS & BBC R&D.
- Import Audio
- Collect Results
- Visualize Results
- startTime: endTime - time_slot #start_offset is + 25% time_slot
- endTime #end_offset is - 25% time_slot
- editable: true
- color: waveform color
- labelText: id of segment
Linux or Mac OS
Peak.js: Web-based audio visualization(https://github.com/bbcrd/peaks.js)
Taglib: Ruby binding C++ library for reading metadata from audio files(http://robinst.github.io/taglib-ruby/)
Parallel: Ruby library for parallel processing (https://github.com/grosser/parallel)
AVCONV: Fast Video and Audio Converter(http://manpages.ubuntu.com/manpages/precise/man1/avconv.1.html)
SSConvert: Converter between various spreadsheet file formats(http://manpages.ubuntu.com/manpages/karmic/man1/ssconvert.1.html)
Sinatra: Ruby domain-specific language for creating web applications(http://www.sinatrarb.com/intro.html)
Input Audio: MP3 or M4A