|
|
|
|
|
import os.path |
|
|
|
import time as reqtime |
|
import datetime |
|
from pytz import timezone |
|
|
|
from itertools import groupby |
|
import copy |
|
|
|
import gradio as gr |
|
|
|
import random |
|
|
|
from midi_to_colab_audio import midi_to_colab_audio |
|
import TMIDIX |
|
|
|
import matplotlib.pyplot as plt |
|
|
|
in_space = os.getenv("SYSTEM") == "spaces" |
|
|
|
|
|
|
|
def pitches_counts(melody_score): |
|
|
|
pitches = [p[4] for p in melody_score] |
|
|
|
pcounts = [] |
|
|
|
count = 0 |
|
pp = -1 |
|
|
|
for p in pitches: |
|
if p == pp: |
|
count += 1 |
|
pcounts.append(count) |
|
else: |
|
count = 0 |
|
pcounts.append(count) |
|
pp = p |
|
|
|
return pcounts |
|
|
|
|
|
|
|
def find_similar_song(songs, src_melody): |
|
|
|
src_pcount = pitches_counts(src_melody) |
|
|
|
ratios = [] |
|
|
|
for s in songs: |
|
patch = s[1] |
|
|
|
trg_melody = [e for e in s[3] if e[6] == patch] |
|
trg_pcount = pitches_counts(trg_melody) |
|
|
|
pcount = 0 |
|
|
|
for i, c in enumerate(src_pcount): |
|
if c == trg_pcount[i]: |
|
pcount += 1 |
|
|
|
ratios.append(pcount / len(src_pcount)) |
|
|
|
max_ratio = max(ratios) |
|
|
|
return songs[ratios.index(max_ratio)], max_ratio |
|
|
|
|
|
|
|
def mix_chord(chord, tones_chord, mel_patch, mel_pitch, next_note_dtime): |
|
|
|
cho = [] |
|
|
|
for k, g in groupby(sorted(chord, key=lambda x: x[6]), lambda x: x[6]): |
|
|
|
if k != 128: |
|
if k == mel_patch: |
|
|
|
cg = list(g) |
|
|
|
c = copy.deepcopy(cg[0]) |
|
|
|
if cg[0][2] > next_note_dtime: |
|
c[2] = next_note_dtime |
|
|
|
c[4] = mel_pitch |
|
c[5] = 105 + (mel_pitch % 12) |
|
|
|
cho.append(c) |
|
|
|
else: |
|
cg = list(g) |
|
|
|
tclen = len(tones_chord) |
|
|
|
tchord = tones_chord |
|
|
|
if len(cg) > tclen: |
|
tchord = tones_chord + [random.choice(tones_chord) for _ in range(len(cg)-tclen)] |
|
|
|
for i, cc in enumerate(cg): |
|
|
|
c = copy.deepcopy(cc) |
|
|
|
if cc[2] > next_note_dtime: |
|
c[2] = next_note_dtime |
|
|
|
c[4] = ((c[4] // 12) * 12) + tchord[i] |
|
c[5] += c[4] % 12 |
|
|
|
cho.append(c) |
|
|
|
else: |
|
cho.extend(list(g)) |
|
|
|
return cho |
|
|
|
|
|
|
|
def MixMelody(input_midi, input_find_best_match): |
|
print('=' * 70) |
|
print('Req start time: {:%Y-%m-%d %H:%M:%S}'.format(datetime.datetime.now(PDT))) |
|
start_time = reqtime.time() |
|
|
|
print('=' * 70) |
|
|
|
fn = os.path.basename(input_midi.name) |
|
fn1 = fn.split('.')[0] |
|
|
|
print('-' * 70) |
|
print('Input file name:', fn) |
|
print('Find best matches', input_find_best_match) |
|
print('-' * 70) |
|
|
|
|
|
raw_score = TMIDIX.midi2single_track_ms_score(input_midi.name) |
|
|
|
|
|
|
|
|
|
raw_escore = TMIDIX.advanced_score_processor(raw_score, return_enhanced_score_notes=True)[0] |
|
|
|
if len(raw_escore) > 0: |
|
|
|
|
|
|
|
|
|
src_escore = TMIDIX.recalculate_score_timings(TMIDIX.augment_enhanced_score_notes([e for e in raw_escore if e[6] < 80])) |
|
|
|
src_cscore = TMIDIX.chordify_score([1000, src_escore]) |
|
|
|
src_melody = [c[0] for c in src_cscore][:256] |
|
|
|
src_melody_pitches = [p[4] for p in src_melody] |
|
|
|
src_harm_tones_chords = TMIDIX.harmonize_enhanced_melody_score_notes(src_melody) |
|
|
|
|
|
|
|
matched_songs = [a for a in all_songs if a[2] == max(32, len(src_melody))] |
|
|
|
random.shuffle(matched_songs) |
|
|
|
max_match_ratio = -1 |
|
|
|
if input_find_best_match: |
|
new_song, max_match_ratio = find_similar_song(matched_songs, src_melody) |
|
else: |
|
new_song = random.choice(matched_songs) |
|
|
|
print('Selected Monster Mono Melodies MIDI:', new_song[0]) |
|
print('Selected melody match ratio:', max_match_ratio) |
|
print('Selected melody instrument:', TMIDIX.Number2patch[new_song[1]], '(', new_song[1], ')') |
|
print('Melody notes count:', new_song[2]) |
|
print('Matched melodies pool count', len(matched_songs)) |
|
|
|
MIDI_Summary = 'Selected Monster Mono Melodies MIDI: ' + str(new_song[0]) + '\n' |
|
MIDI_Summary += 'Selected melody match ratio: ' + str(max_match_ratio) + '\n' |
|
MIDI_Summary += 'Selected melody instrument: ' + str(TMIDIX.Number2patch[new_song[1]]) + ' (' + str(new_song[1]) + ')' + '\n' |
|
MIDI_Summary += 'Melody notes count: ' + str(new_song[2]) + '\n' |
|
MIDI_Summary += 'Matched melodies pool count: ' + str(len(matched_songs)) |
|
|
|
fn1 += '_' + str(new_song[0]) + '_' + str(TMIDIX.Number2patch[new_song[1]]) + '_' + str(new_song[1]) + '_' + str(new_song[2]) |
|
|
|
trg_patch = new_song[1] |
|
|
|
trg_song = copy.deepcopy(new_song[3]) |
|
TMIDIX.adjust_score_velocities(trg_song, 95) |
|
|
|
cscore = TMIDIX.chordify_score([1000, trg_song]) |
|
|
|
print('=' * 70) |
|
print('Done loading source and target MIDIs...!') |
|
print('=' * 70) |
|
print('Mixing...') |
|
|
|
mixed_song = [] |
|
|
|
midx = 0 |
|
|
|
for i, c in enumerate(cscore): |
|
cho = copy.deepcopy(c) |
|
|
|
patches = sorted(set([e[6] for e in c])) |
|
|
|
if trg_patch in patches: |
|
|
|
if midx < len(src_melody)-1: |
|
next_note_dtime = src_melody[midx+1][1] - src_melody[midx][1] |
|
else: |
|
next_note_dtime = 255 |
|
|
|
mixed_song.extend(mix_chord(c, src_harm_tones_chords[midx], trg_patch, src_melody_pitches[midx], next_note_dtime)) |
|
|
|
midx += 1 |
|
|
|
else: |
|
|
|
if i < len(cscore)-1: |
|
next_note_dtime = cscore[i+1][0][1] - cscore[i][0][1] |
|
else: |
|
next_note_dtime = 255 |
|
|
|
mixed_song.extend(mix_chord(cho, src_harm_tones_chords[midx], trg_patch, src_melody_pitches[midx], next_note_dtime)) |
|
|
|
if midx == len(src_melody): |
|
break |
|
|
|
print('=' * 70) |
|
print('Done!') |
|
print('=' * 70) |
|
|
|
|
|
print('Rendering results...') |
|
|
|
print('=' * 70) |
|
print('Sample INTs', mixed_song[:5]) |
|
print('=' * 70) |
|
|
|
output_score, patches, overflow_patches = TMIDIX.patch_enhanced_score_notes(mixed_song) |
|
|
|
detailed_stats = TMIDIX.Tegridy_ms_SONG_to_MIDI_Converter(output_score, |
|
output_signature = 'Harmonic Melody MIDI Mixer', |
|
output_file_name = fn1, |
|
track_name='Project Los Angeles', |
|
list_of_MIDI_patches=patches, |
|
timings_multiplier=16 |
|
) |
|
|
|
new_fn = fn1+'.mid' |
|
|
|
|
|
audio = midi_to_colab_audio(new_fn, |
|
soundfont_path=soundfont, |
|
sample_rate=16000, |
|
volume_scale=10, |
|
output_for_gradio=True |
|
) |
|
|
|
print('Done!') |
|
print('=' * 70) |
|
|
|
|
|
|
|
output_midi_title = str(fn1) |
|
output_midi_summary = str(MIDI_Summary) |
|
output_midi = str(new_fn) |
|
output_audio = (16000, audio) |
|
|
|
for o in output_score: |
|
o[1] *= 16 |
|
o[2] *= 16 |
|
|
|
output_plot = TMIDIX.plot_ms_SONG(output_score, plot_title=output_midi_title, return_plt=True) |
|
|
|
print('Output MIDI file name:', output_midi) |
|
print('Output MIDI title:', output_midi_title) |
|
print('Output MIDI summary:', MIDI_Summary) |
|
print('=' * 70) |
|
|
|
|
|
|
|
|
|
print('-' * 70) |
|
print('Req end time: {:%Y-%m-%d %H:%M:%S}'.format(datetime.datetime.now(PDT))) |
|
print('-' * 70) |
|
print('Req execution time:', (reqtime.time() - start_time), 'sec') |
|
|
|
return output_midi_title, output_midi_summary, output_midi, output_audio, output_plot |
|
|
|
|
|
|
|
if __name__ == "__main__": |
|
|
|
PDT = timezone('US/Pacific') |
|
|
|
print('=' * 70) |
|
print('App start time: {:%Y-%m-%d %H:%M:%S}'.format(datetime.datetime.now(PDT))) |
|
print('=' * 70) |
|
|
|
soundfont = "SGM-v2.01-YamahaGrand-Guit-Bass-v2.7.sf2" |
|
|
|
all_songs = TMIDIX.Tegridy_Any_Pickle_File_Reader('Monster_Mono_Melodies_MIDI_Dataset_65536_32_256') |
|
print('=' * 70) |
|
|
|
app = gr.Blocks() |
|
with app: |
|
gr.Markdown("<h1 style='text-align: center; margin-bottom: 1rem'>Harmonic Melody MIDI Mixer</h1>") |
|
gr.Markdown("<h1 style='text-align: center; margin-bottom: 1rem'>Harmonize and mix any MIDI melody</h1>") |
|
gr.Markdown( |
|
"![Visitors](https://api.visitorbadge.io/api/visitors?path=asigalov61.Harmonic-Melody-MIDI-Mixer&style=flat)\n\n" |
|
"This is a demo for TMIDIX Python module from tegridy-tools and Monster Mono Melodies MIDI Dataset\n\n" |
|
"Check out [tegridy-tools](https://github.com/asigalov61/tegridy-tools) on GitHub!\n\n" |
|
"Check out [Monster-MIDI-Dataset](https://github.com/asigalov61/Monster-MIDI-Dataset) on GitHub!\n\n" |
|
) |
|
gr.Markdown("## Upload your MIDI or select a sample example MIDI below") |
|
|
|
input_midi = gr.File(label="Input MIDI", file_types=[".midi", ".mid", ".kar"]) |
|
input_find_best_match = gr.Checkbox(label="Find best match", value=True) |
|
|
|
run_btn = gr.Button("mix melody", variant="primary") |
|
|
|
gr.Markdown("## Output results") |
|
|
|
output_midi_title = gr.Textbox(label="Output MIDI title") |
|
output_midi_summary = gr.Textbox(label="Output MIDI summary") |
|
output_audio = gr.Audio(label="Output MIDI audio", format="wav", elem_id="midi_audio") |
|
output_plot = gr.Plot(label="Output MIDI score plot") |
|
output_midi = gr.File(label="Output MIDI file", file_types=[".mid"]) |
|
|
|
|
|
run_event = run_btn.click(MixMelody, [input_midi, input_find_best_match], |
|
[output_midi_title, output_midi_summary, output_midi, output_audio, output_plot]) |
|
|
|
gr.Examples( |
|
[["Abracadabra-Sample-Melody.mid", True], |
|
["Sparks-Fly-Sample-Melody.mid", True], |
|
], |
|
[input_midi, input_find_best_match], |
|
[output_midi_title, output_midi_summary, output_midi, output_audio, output_plot], |
|
MixMelody, |
|
cache_examples=True, |
|
) |
|
|
|
app.queue().launch() |