Training data:
F.txt
The training data for Bach's fugue.twinkle.txt
: Twinkle Twinkle Little Star.
Outputs from single_voice_prediction.ipynb
:
original.mid
: The original data (F.txt), only the lowest voice.pred.mid
: The original (F.txt) plus its predictions, only the lowest voice. (You may compare this withoriginal.mid
to see when the prediction starts.)twinkle_pred.mid
: The twinkle.txt data plus its predictions, only the 0th (first column).original_all_voices.mid
: The original data (F.txt), all voices. Not used in the report.pred_all_voices.mid
: The original (F.txt) plus its predictions, all voices. These voices were predicted separately. Not used in the report.
Outputs from multiple_voice_prediction.ipynb
:
simplified_generated.mid
: The multiple-voice prediction.
Performance results:
p4Chain.py
: add split methods for test and train set, and generate both regression outputs for all data used and only train data usedp4multi.py
: the same as abovePerformance.rmd
: calculate piano format output for both algorithms and make plotsstatistics_calculation.py
: calculate statistics parameteres