Music

This dataset contains source and processed audio files designed primarily for use with Time-Scale Modification (TSM) research.  88 source files were processed at 10 different time-scale ratios by 6 different TSM algorithms resulting in training set of 5280 files.  Each file was then subjectively evaluated by a minimum of 7 participants providing opinion scores.  Mean Opinion Scores (MeanOS) and Median Opinion Scores (MedianOS) are provided for each file.  An additional 20 files were processed at 4 time-scale ratios by an additional 3 methods were also subjectively evaluated resulting in a t

102 views
  • Machine Learning
  • Last Updated On: 
    Thu, 02/13/2020 - 01:07

    The Million Song Dataset is a freely-available collection of audio features and metadata for a million contemporary popular music tracks. Its purposes are:

    • To encourage research on algorithms that scale to commercial sizes
    • To provide a reference dataset for evaluating research
    • As a shortcut alternative to creating a large dataset with APIs (e.g. The Echo Nest's)
    • To help new researchers get started in the MIR field

     

    385 views
  • Digital signal processing
  • Last Updated On: 
    Thu, 11/08/2018 - 10:34