knn_instrument_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy, scipy, matplotlib.pyplot as plt, sklearn, librosa, mir_eval, urllib, IPython.display, stanford_mir\n",
    "plt.rcParams['figure.figsize'] = (14,5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[&larr; Back to Index](index.html)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercise: Instrument Classification using K-NN"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This exercise is loosely based upon \"Lab 1\" from previous MIR workshops ([2010](https://ccrma.stanford.edu/workshops/mir2010/Lab1_2010.pdf))."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For more on K-NN, see the [notebook on K-NN](knn.ipynb)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For help from a similar exercise, [follow the steps in the feature sonification exercise](feature_sonification.ipynb#Step-1:-Retrieve-Audio) first."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Goals"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. Extract spectral features from an audio signal.\n",
    "2. Train a K-Nearest Neighbor classifier.\n",
    "3. Use the classifier to classify beats in a drum loop."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 1: Retrieve Audio, Detect Onsets, and Segment"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Download the file `simple_loop.wav` onto your local machine."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "filename = 'simple_loop.wav'\n",
    "urllib.urlretrieve?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the audio file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "librosa.load?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Play the audio file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "IPython.display.Audio?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Detect onsets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "librosa.onset.onset_detect?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Convert onsets from frames to seconds (and samples):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "librosa.frames_to_time?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "librosa.frames_to_samples?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Listen to a click track, with clicks located at each onset, plus the original audio:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mir_eval.sonify.clicks?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "IPython.display.Audio?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 2: Extract Features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each segment, compute the zero crossing rate and spectral centroid."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "librosa.zero_crossings?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "librosa.feature.spectral_centroid?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Scale the features to be in the range [-1, 1]:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sklearn.preprocessing.MinMaxScaler?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sklearn.preprocessing.MinMaxScaler.fit_transform?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 3: Train K-NN Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use `stanford_mir.download_drum_samples` to download ten kick drum samples and ten snare drum samples. Each audio file contains a single drum hit at the beginning of the file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "stanford_mir.download_drum_samples?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each audio file, extract one feature vector. Concatenate all of these feature vectors into one feature table."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "numpy.concatenate?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 4: Run the Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create a K-NN classifer model object:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sklearn.neighbors.KNeighborsClassifier?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train the classifier:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sklearn.neighbors.KNeighborsClassifier.fit?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, run the classifier on the test input audio file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sklearn.neighbors.KNeighborsClassifier.predict?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 5: Sonify the Classifier Output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Play a \"beep\" for each detected kick drum. Repeat for the snare drum."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mir_eval.sonify.clicks?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## For Further Exploration"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In addition to the features used above, extract the following features:\n",
    "\n",
    "- spectral centroid\n",
    "- spectral spread\n",
    "- spectral skewness\n",
    "- spectral kurtosis\n",
    "- spectral rolloff\n",
    "- MFCCs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Re-train the classifier, and re-run the classifier over the test audio signal. Do the results change?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Repeat the steps above for more audio files."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[&larr; Back to Index](index.html)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}