5 TF Lite Inference.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "import numpy       as np\n",
    "import pretty_midi as pm\n",
    "import librosa     as lbr\n",
    "from   librosa.display import specshow\n",
    "import tensorflow  as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([{'name': 'Placeholder',\n",
       "   'index': 16,\n",
       "   'shape': array([17920]),\n",
       "   'shape_signature': array([17920]),\n",
       "   'dtype': numpy.float32,\n",
       "   'quantization': (0.0, 0),\n",
       "   'quantization_parameters': {'scales': array([], dtype=float32),\n",
       "    'zero_points': array([], dtype=int32),\n",
       "    'quantized_dimension': 0},\n",
       "   'sparsity_parameters': {}}],\n",
       " [{'name': 'frame_logits',\n",
       "   'index': 1332,\n",
       "   'shape': array([ 1, 32, 88]),\n",
       "   'shape_signature': array([ 1, 32, 88]),\n",
       "   'dtype': numpy.float32,\n",
       "   'quantization': (0.0, 0),\n",
       "   'quantization_parameters': {'scales': array([], dtype=float32),\n",
       "    'zero_points': array([], dtype=int32),\n",
       "    'quantized_dimension': 0},\n",
       "   'sparsity_parameters': {}},\n",
       "  {'name': 'onset_logits',\n",
       "   'index': 2635,\n",
       "   'shape': array([ 1, 32, 88]),\n",
       "   'shape_signature': array([ 1, 32, 88]),\n",
       "   'dtype': numpy.float32,\n",
       "   'quantization': (0.0, 0),\n",
       "   'quantization_parameters': {'scales': array([], dtype=float32),\n",
       "    'zero_points': array([], dtype=int32),\n",
       "    'quantized_dimension': 0},\n",
       "   'sparsity_parameters': {}},\n",
       "  {'name': 'offset_logits',\n",
       "   'index': 1336,\n",
       "   'shape': array([ 1, 32, 88]),\n",
       "   'shape_signature': array([ 1, 32, 88]),\n",
       "   'dtype': numpy.float32,\n",
       "   'quantization': (0.0, 0),\n",
       "   'quantization_parameters': {'scales': array([], dtype=float32),\n",
       "    'zero_points': array([], dtype=int32),\n",
       "    'quantized_dimension': 0},\n",
       "   'sparsity_parameters': {}},\n",
       "  {'name': 'velocity_values',\n",
       "   'index': 3957,\n",
       "   'shape': array([ 1, 32, 88]),\n",
       "   'shape_signature': array([ 1, 32, 88]),\n",
       "   'dtype': numpy.float32,\n",
       "   'quantization': (0.0, 0),\n",
       "   'quantization_parameters': {'scales': array([], dtype=float32),\n",
       "    'zero_points': array([], dtype=int32),\n",
       "    'quantized_dimension': 0},\n",
       "   'sparsity_parameters': {}}])"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "interp = tf.lite.Interpreter(model_path=\"onsets_frames_wavinput.tflite\")\n",
    "interp.get_input_details(), interp.get_output_details()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "frame_logits -23.69737 6.138942 \n",
      " [[[-10.075554   -8.374783   -6.84772   ... -16.522121  -14.521896\n",
      "   -11.675898 ]\n",
      "  [ -7.588108   -6.482976   -8.975293  ... -17.467564  -15.592136\n",
      "   -12.42377  ]\n",
      "  [ -6.4807262  -6.1015177 -11.316768  ... -20.001945  -19.071222\n",
      "   -14.897568 ]\n",
      "  ...\n",
      "  [ -7.1594357  -6.243394   -5.824191  ... -16.188425  -20.882141\n",
      "   -18.404758 ]\n",
      "  [ -7.621876   -5.7240925  -6.9756927 ... -15.3395    -20.947762\n",
      "   -16.949936 ]\n",
      "  [ -8.354395   -7.688925   -8.928634  ... -15.203535  -21.160841\n",
      "   -16.934267 ]]]\n",
      "\n",
      "onset_logits -26.280355 2.1618528 \n",
      " [[[ -9.927405   -7.369683   -9.975944  ... -14.781149  -20.59234\n",
      "   -21.875896 ]\n",
      "  [ -8.494212   -6.497216   -9.745885  ... -16.438147  -20.16504\n",
      "   -21.469685 ]\n",
      "  [ -5.74159    -5.374315   -7.007977  ... -15.074481  -19.662813\n",
      "   -19.571705 ]\n",
      "  ...\n",
      "  [ -7.404469  -13.952096   -9.1285095 ... -17.606022  -16.07436\n",
      "   -18.246044 ]\n",
      "  [-10.06579   -16.22526   -12.722668  ... -19.301077  -15.797797\n",
      "   -17.987482 ]\n",
      "  [-11.48592   -14.447435  -13.82669   ... -17.178944  -15.940237\n",
      "   -18.457497 ]]]\n",
      "\n",
      "offset_logits -24.51408 0.44074488 \n",
      " [[[-17.340656  -11.373997  -14.253749  ... -15.655101  -11.701646\n",
      "   -15.163293 ]\n",
      "  [-11.423455  -10.765892  -11.932632  ... -13.555202  -14.262076\n",
      "   -14.44061  ]\n",
      "  [ -9.7559595 -10.9763775 -13.181602  ... -17.424536  -21.783028\n",
      "   -18.9529   ]\n",
      "  ...\n",
      "  [ -4.0059843  -4.406693   -4.9031034 ... -12.996494  -20.538347\n",
      "   -18.19133  ]\n",
      "  [ -7.0618043  -5.8196406  -8.942878  ... -15.272884  -24.51408\n",
      "   -24.502539 ]\n",
      "  [ -7.028188   -6.166064  -10.240178  ... -19.323587  -22.409927\n",
      "   -20.684498 ]]]\n",
      "\n",
      "velocity_values 0.48288545 0.89712584 \n",
      " [[[0.823717   0.78170323 0.74166745 ... 0.8308126  0.752248   0.76715326]\n",
      "  [0.85326445 0.7862402  0.70128155 ... 0.8375125  0.77962613 0.6923882 ]\n",
      "  [0.822463   0.77855116 0.66571385 ... 0.83219975 0.79527944 0.6846243 ]\n",
      "  ...\n",
      "  [0.8218269  0.7565063  0.74646467 ... 0.8963327  0.81083107 0.5659447 ]\n",
      "  [0.7581388  0.7540133  0.7273959  ... 0.86502516 0.7427253  0.708699  ]\n",
      "  [0.6144901  0.6770095  0.6732216  ... 0.8069209  0.7065943  0.74213696]]]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "interp.allocate_tensors()\n",
    "inputLen = interp.get_input_details()[0]['shape'][0]\n",
    "interp.set_tensor(interp.get_input_details()[0]['index'], np.array(np.random.random_sample(inputLen), dtype=np.float32))\n",
    "interp.invoke()\n",
    "for output in interp.get_output_details():\n",
    "    first32Chords = interp.get_tensor(output['index'])\n",
    "    print(output['name'], first32Chords[0].min(), first32Chords[0].max(), '\\n', first32Chords, end='\\n\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Load the Song."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Song duration\t0 min : 43 sec\n"
     ]
    }
   ],
   "source": [
    "rate, songName = 16_000, './Test songs/MIDI-Unprocessed_22_R1_2011_MID--AUDIO_R1-D8_12_Track12_wav.wav'\n",
    "\n",
    "song = lbr.effects.trim(lbr.load(songName, rate)[0])[0]\n",
    "songLen = int(lbr.get_duration(song, rate))\n",
    "print('Song duration\\t{} min : {} sec'.format(songLen // 60, songLen % 60))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Predict Probabilities."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "frame_logits -27.153994 0.8419235 \n",
      " [[[-14.042323  -12.16228   -15.464712  ... -19.305365  -17.50471\n",
      "   -15.492955 ]\n",
      "  [-14.720356  -12.98374   -16.597113  ... -22.14126   -18.51053\n",
      "   -17.276787 ]\n",
      "  [-16.611492  -13.866851  -16.829737  ... -23.895563  -20.367254\n",
      "   -18.967611 ]\n",
      "  ...\n",
      "  [-13.751821  -13.166228  -14.316219  ... -18.07019   -17.144405\n",
      "   -15.896842 ]\n",
      "  [-12.448567  -13.176397  -12.176679  ... -14.821554  -15.790749\n",
      "   -14.629224 ]\n",
      "  [ -7.995164  -12.186578   -8.2076435 ... -10.337193  -14.158825\n",
      "   -12.389364 ]]]\n",
      "\n",
      "onset_logits -23.404188 1.6334646 \n",
      " [[[-13.625739  -13.542617  -10.938008  ... -15.825341  -14.930651\n",
      "   -15.726932 ]\n",
      "  [-14.180186  -10.902331  -11.793695  ... -15.237801  -15.9614105\n",
      "   -16.198803 ]\n",
      "  [-15.372886  -12.241614  -13.743533  ... -16.694813  -16.858574\n",
      "   -17.099842 ]\n",
      "  ...\n",
      "  [-12.024006  -12.641288  -16.27623   ... -22.412968  -21.697542\n",
      "   -21.875744 ]\n",
      "  [ -8.250113  -12.320369  -13.584641  ... -23.404188  -18.041811\n",
      "   -22.526606 ]\n",
      "  [ -3.2675467  -5.561775   -6.705703  ... -12.441843  -11.890234\n",
      "   -13.985788 ]]]\n",
      "\n",
      "offset_logits -28.826075 -1.7003431 \n",
      " [[[-14.37775   -16.083757  -15.840225  ... -14.813948  -15.796591\n",
      "   -17.52423  ]\n",
      "  [-12.82962   -15.37506   -15.864007  ... -15.296923  -14.8387375\n",
      "   -20.143234 ]\n",
      "  [-15.582333  -16.667921  -18.57708   ... -17.206629  -16.002592\n",
      "   -21.493275 ]\n",
      "  ...\n",
      "  [-18.38626   -18.895933  -17.152014  ... -15.227375  -18.431753\n",
      "   -26.381958 ]\n",
      "  [-16.781967  -16.303637  -15.836984  ... -16.285673  -18.523483\n",
      "   -23.93745  ]\n",
      "  [-11.53842   -10.332179  -12.626075  ... -15.759255  -21.430754\n",
      "   -24.323797 ]]]\n",
      "\n",
      "velocity_values 0.024562836 0.74945533 \n",
      " [[[0.3414591  0.13483727 0.5112071  ... 0.35711384 0.22428694 0.7163497 ]\n",
      "  [0.33953536 0.06248683 0.5217909  ... 0.29869884 0.13797796 0.69671273]\n",
      "  [0.3210908  0.03165787 0.4977731  ... 0.27831206 0.12431842 0.691208  ]\n",
      "  ...\n",
      "  [0.5649249  0.5394853  0.24028349 ... 0.3219682  0.40771052 0.67696387]\n",
      "  [0.6056961  0.5770917  0.25661293 ... 0.3153806  0.45923358 0.6750634 ]\n",
      "  [0.6369287  0.6639227  0.62672186 ... 0.7301516  0.681034   0.6862562 ]]]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "interp.set_tensor(interp.get_input_details()[0]['index'], song[:inputLen])\n",
    "interp.invoke()\n",
    "for output in interp.get_output_details():\n",
    "    first32Chords = interp.get_tensor(output['index'])\n",
    "    print(output['name'], first32Chords[0].min(), first32Chords[0].max(), '\\n', first32Chords, end='\\n\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3 Keep Only Notes with Probabilities >= some threshold.\n",
    "From https://arxiv.org/pdf/1710.11153.pdf:\n",
    "\n",
    "Page 1 \"Abstract\":\n",
    "\n",
    "`During inference, we restrict the predictions from the framewise detector by not allowing a new note to start unless the onset detector also agrees that an onset for that pitch is present in the frame.`\n",
    "\n",
    "Page 2, Section 3 \"Model Configuration\":\n",
    "\n",
    "`An activation from the frame detector is only allowed to start a note if the onset detector agrees that an onset is present in that frame.`\n",
    "\n",
    "Page 3, Section 3.1 \"Velocity Estimation\":\n",
    "\n",
    "`At inference time the output is ... then transformed to a midi velocity by the following mapping:`\n",
    "$$v_{midi} = 80 * v_{predicted} + 10$$\n",
    "`The final mapping is arbitrary, but we found this leads to pleasing audio renderings.`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Evgeniy\\Anaconda3\\lib\\site-packages\\librosa\\display.py:974: MatplotlibDeprecationWarning: The 'basey' parameter of __init__() has been renamed 'base' since Matplotlib 3.3; support for the old name will be dropped two minor releases later.\n",
      "  scaler(mode, **kwargs)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 929,
       "width": 941
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "actProb, onProb, offProb, volProb = np.empty((1, 88)), np.empty((1, 88)), np.empty((1, 88)), np.empty((1, 88))\n",
    "outputStep = interp.get_output_details()[0]['shape'][1] * 512\n",
    "paddedSong = np.append(song, np.zeros(-(song.size - inputLen) % outputStep, dtype=np.float32))\n",
    "for i in range((paddedSong.size - inputLen) // outputStep + 1):\n",
    "    interp.set_tensor(interp.get_input_details()[0]['index'], paddedSong[i * outputStep : i * outputStep + inputLen])\n",
    "    interp.invoke()\n",
    "    actProb = np.vstack((actProb, interp.get_tensor(interp.get_output_details()[0]['index'])[0]))\n",
    "    onProb  = np.vstack(( onProb, interp.get_tensor(interp.get_output_details()[1]['index'])[0]))\n",
    "    offProb = np.vstack((offProb, interp.get_tensor(interp.get_output_details()[2]['index'])[0]))\n",
    "    volProb = np.vstack((volProb, interp.get_tensor(interp.get_output_details()[3]['index'])[0]))\n",
    "\n",
    "#############\n",
    "threshold = 0\n",
    "#############\n",
    "\n",
    "plt.figure(figsize=(16, 16))\n",
    "def PlotRoll(i, yProb, name):\n",
    "    plt.subplot(2, 1, i)\n",
    "    yProb[yProb < threshold] = 0\n",
    "    specshow(yProb[0:lbr.time_to_frames(10, rate)].T, x_axis='time', y_axis='cqt_note', sr=rate, fmin=lbr.note_to_midi('A0'))\n",
    "    plt.title('Piano-roll of predicted {} probabilities'.format(name))\n",
    "\n",
    "PlotRoll(1, onProb, 'onset')\n",
    "PlotRoll(2, np.maximum(onProb, actProb), 'note')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Evgeniy\\Anaconda3\\lib\\site-packages\\librosa\\filters.py:239: UserWarning: Empty filters detected in mel frequency basis. Some channels will produce empty responses. Try increasing your sampling rate (and fmax) or reducing n_mels.\n",
      "  \"Empty filters detected in mel frequency basis. \"\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(['Bb', 'Eb'], ['G', 'D', 'C', 'A', 'F', 'Eb', 'Bb'])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 712,
       "width": 949
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "midi = pm.PrettyMIDI(initial_tempo=lbr.beat.tempo(song, rate).mean())\n",
    "midi.lyrics += [pm.Lyric('Automatically transcribed from audio:\\r\\n\\t' + songName, 0),\n",
    "                pm.Lyric('Used software created by Boris Shakhovsky', 0)]\n",
    "track = pm.Instrument(program=pm.instrument_name_to_program('Acoustic Grand Piano'), name='Acoustic Grand Piano')\n",
    "midi.instruments += [track]\n",
    "\n",
    "''' Based on https://github.com/tensorflow/magenta/blob/master/magenta/music/sequences_lib.py#L1844\n",
    "    magenta.music.midi_ionote_sequence_to_midi_file(\n",
    "    magenta.music.sequences_libpianoroll_to_note_sequence(fps=rate/512, min_duration=0, min_midi_pitch=21 ... '''\n",
    "\n",
    "intervals, frameLenSecs = {}, lbr.frames_to_time(1, rate) # Time is in absolute seconds, not relative MIDI ticks\n",
    "########################################################\n",
    "onsets = (onProb > threshold).astype(np.int8)\n",
    "frames = onsets | (actProb > threshold).astype(np.int8) # Ensure that any frame with an onset prediction is considered active.\n",
    "#######################################################\n",
    "\n",
    "def EndPitch(pitch, endFrame):\n",
    "    #######################################################################################\n",
    "    if volProb[intervals[pitch], pitch] < 0 or volProb[intervals[pitch], pitch] > 1: return\n",
    "    #######################################################################################\n",
    "    track.notes += [pm.Note(int(max(0, min(1, volProb[intervals[pitch], pitch])) * 80 + 10), pitch + 21,\n",
    "                            intervals[pitch] * frameLenSecs, endFrame * frameLenSecs)]\n",
    "    del intervals[pitch]\n",
    "\n",
    "# Add silent frame at the end so we can do a final loop and terminate any notes that are still active:\n",
    "for i, frame in enumerate(np.vstack([frames, np.zeros(frames.shape[1])])):\n",
    "    for pitch, active in enumerate(frame):\n",
    "        if active:\n",
    "            if pitch not in intervals:\n",
    "                if onsets is None: intervals[pitch] = i\n",
    "                elif onsets[i, pitch]: intervals[pitch] = i # Start a note only if we have predicted an onset\n",
    "                #else: Even though the frame is active, there is no onset, so ignore it\n",
    "            elif onsets is not None:\n",
    "                if (onsets[i, pitch] and not onsets[i - 1, pitch]):\n",
    "                    EndPitch(pitch, i)   # Pitch is already active, but because of a new onset, we should end the note\n",
    "                    intervals[pitch] = i # and start a new one\n",
    "        elif pitch in intervals: EndPitch(pitch, i)\n",
    "\n",
    "if track.notes: assert len(frames) * frameLenSecs >= track.notes[-1].end, 'Wrong MIDI sequence duration'\n",
    "\n",
    "plt.figure(figsize=(16, 12))\n",
    "#plt.subplot(2, 1, 1)\n",
    "\n",
    "notes = midi.get_pitch_class_histogram()\n",
    "plt.bar(np.arange(12), notes)\n",
    "plt.xticks(np.arange(12), ['C', 'C#', 'D', 'Eb', 'E', 'F', 'F#', 'G', 'Ab', 'A', 'Bb', 'B'])\n",
    "plt.title('Pitch class distribution (sort of a proxy for key)')\n",
    "plt.xlabel('Note')\n",
    "plt.ylabel('Proportion')\n",
    "plt.xlim(-.5, 11.5)\n",
    "\n",
    "gamma = [n for _, n in sorted([(count, ['C', 'C#', 'D', 'Eb', 'E', 'F', 'F#', 'G', 'Ab', 'A', 'Bb', 'B'][i])\n",
    "                               for i, count in enumerate(notes)], reverse=True)[:7]]\n",
    "blacks = sorted(n for n in gamma if len(n) > 1)\n",
    "blacks, gamma"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4.1 Key Signature Estimation, Option 1.\n",
    "### From Kumhansl and Schmuckler as reported here:\n",
    "http://rnhart.net/articles/key-finding/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Em'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "chroma = lbr.feature.chroma_cqt(song, rate).sum(1)\n",
    "major = [np.corrcoef(chroma, np.roll([6.35, 2.23, 3.48, 2.33, 4.38, 4.09, 2.52, 5.19, 2.39, 3.66, 2.29, 2.88], i))[0, 1] for i in range(12)]\n",
    "minor = [np.corrcoef(chroma, np.roll([6.33, 2.68, 3.52, 5.38, 2.60, 3.53, 2.54, 4.75, 3.98, 2.69, 3.34, 3.17], i))[0, 1] for i in range(12)]\n",
    "keySignature = (['C', 'C#', 'D', 'Eb', 'E', 'F', 'F#', 'G', 'Ab', 'A', 'Bb', 'B'][\n",
    "    major.index(max(major)) if max(major) > max(minor) else minor.index(max(minor)) - 3]\n",
    "                + ('m' if max(major) < max(minor) else ''))\n",
    "keySignature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Am'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "chroma = lbr.feature.chroma_cqt(C=onsets.T, fmin=lbr.note_to_hz('A0')).sum(1)\n",
    "major = [np.corrcoef(chroma, np.roll([6.35, 2.23, 3.48, 2.33, 4.38, 4.09, 2.52, 5.19, 2.39, 3.66, 2.29, 2.88], i))[0, 1] for i in range(12)]\n",
    "minor = [np.corrcoef(chroma, np.roll([6.33, 2.68, 3.52, 5.38, 2.60, 3.53, 2.54, 4.75, 3.98, 2.69, 3.34, 3.17], i))[0, 1] for i in range(12)]\n",
    "keySignature = (['C', 'C#', 'D', 'Eb', 'E', 'F', 'F#', 'G', 'Ab', 'A', 'Bb', 'B'][\n",
    "    major.index(max(major)) if max(major) > max(minor) else minor.index(max(minor)) - 3]\n",
    "                + ('m' if max(major) < max(minor) else ''))\n",
    "keySignature"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4.2 Key Signature Estimation, Option 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Gm'"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MajorMinor = lambda mj, mn: mj if gamma.index(mj) < gamma.index(mn) else mn + 'm'\n",
    "\n",
    "if len(blacks) == 0: keySignature = MajorMinor('C', 'A')\n",
    "\n",
    "elif len(blacks) == 1:\n",
    "    if blacks[0] == 'F#':\n",
    "        assert 'F' not in gamma\n",
    "        keySignature = MajorMinor('G', 'E')\n",
    "    elif blacks[0] == 'Bb':\n",
    "        assert 'B' not in gamma\n",
    "        keySignature = MajorMinor('F', 'D')\n",
    "    else: assert False\n",
    "\n",
    "elif len(blacks) == 2:\n",
    "    if blacks == ['C#', 'F#']:\n",
    "        assert 'C' not in gamma and 'F' not in gamma\n",
    "        keySignature = MajorMinor('D', 'B')\n",
    "    elif blacks == ['Bb', 'Eb']:\n",
    "        assert 'B' not in gamma and 'E' not in gamma\n",
    "        keySignature = MajorMinor('Bb', 'G')\n",
    "    else: assert False\n",
    "\n",
    "elif len(blacks) == 3:\n",
    "    if blacks == ['Ab', 'C#', 'F#']:\n",
    "        assert 'C' not in gamma and 'F' not in gamma and 'G' not in gamma\n",
    "        keySignature = MajorMinor('A', 'F#')\n",
    "    elif blacks == ['Ab', 'Bb', 'Eb']:\n",
    "        assert 'B' not in gamma and 'E' not in gamma and 'A' not in gamma\n",
    "        keySignature = MajorMinor('Eb', 'C')\n",
    "    else: assert False\n",
    "\n",
    "elif len(blacks) == 4:\n",
    "    if blacks == ['Ab', 'C#', 'Eb', 'F#']:\n",
    "        assert 'C' not in gamma and 'D' not in gamma and 'F' not in gamma and 'G' not in gamma\n",
    "        keySignature = MajorMinor('E', 'C#')\n",
    "    elif blacks == ['Ab', 'Bb', 'C#', 'Eb']:\n",
    "        assert 'B' not in gamma and 'E' not in gamma and 'A' not in gamma and 'D' not in gamma\n",
    "        keySignature = MajorMinor('Ab', 'F')\n",
    "    else: assert False\n",
    "\n",
    "elif 'B' in gamma and 'E' in gamma: keySignature = MajorMinor('B', 'Ab')\n",
    "elif 'C' in gamma and 'F' in gamma: keySignature = MajorMinor('C#', 'Bb')\n",
    "else: assert False\n",
    "\n",
    "keySignature"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5. Finally, Write MIDI.\n",
    "No instrument information is extracted, and all transcribed notes get combined into one part."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'./Test songs/MIDI-Unprocessed_22_R1_2011_MID--AUDIO_R1-D8_12_Track12_wav.mid saved'"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "midi.key_signature_changes += [pm.KeySignature(pm.key_name_to_key_number(keySignature), 0)]\n",
    "songName = '.'.join(songName.split('.')[:-1]) + '.mid'\n",
    "midi.write(songName)\n",
    "'{} saved'.format(songName)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}