def read_wav(wav_file, output_file):
in_processor = RNNPianoNoteProcessor()
peak_picking = PeakPickingProcessor(threshold=0.35,
smooth=0.09,
combine=0.05)
#output = write_notes
output = write_midi
out_processor = [peak_picking, output]
processor = IOProcessor(in_processor, out_processor)
processor.process(wav_file, output_file)
def beat_extractor(queue_beat):
kwargs = dict(
fps=100,
correct=True,
infile=None,
outfile=None,
max_bpm=170,
min_bpm=60,
#nn_files = [BEATS_LSTM[0]],
transition_lambda=100,
num_frames=1,
online=True,
verbose=1)
def beat_callback(beats, output=None):
if len(beats) > 0:
# Do something with the beat (for now, just print the array to stdout)
queue_beat.put(beats[0])
#print(beats)
#print('Process to write betas: %s' % os.getpid())
in_processor = RNNBeatProcessor(**kwargs)
beat_processor = DBNBeatTrackingProcessor(**kwargs)
out_processor = [beat_processor, beat_callback]
processor = IOProcessor(in_processor, out_processor)
process_online(processor, **kwargs)
def main():
""" music_process """
# define parser
p = argparse.ArgumentParser(
formatter_class=argparse.RawDescriptionHelpFormatter,
description='''
python music_process.py ./data/Jian_chrous.wav -o ./data/jian_beat.txt
''')
# version
p.add_argument('--version',
action='version',
version='FDGame-music-process-0.1')
# input/output options
# p.add_argument('single')
# p.add_argument('./data/Jian_chrous.wav')
io_arguments_single(p)
ActivationsProcessor.add_arguments(p)
# signal processing arguments
SignalProcessor.add_arguments(p, norm=False, gain=0)
# peak picking arguments
DBNBeatTrackingProcessor.add_arguments(p)
NeuralNetworkEnsemble.add_arguments(p, nn_files=None)
# parse arguments
args = p.parse_args()
# set immutable arguments
args.fps = 100
# print arguments
if args.verbose:
print(args)
# input processor
if args.load:
# load the activations from file
in_processor = ActivationsProcessor(mode='r', **vars(args))
else:
# use a RNN to predict the beats
in_processor = RNNBeatProcessor(**vars(args))
# output processor
if args.save:
# save the RNN beat activations to file
out_processor = ActivationsProcessor(mode='w', **vars(args))
else:
# track the beats with a DBN and output them
beat_processor = DBNBeatTrackingProcessor(**vars(args))
out_processor = [beat_processor, write_beats]
# create an IOProcessor
processor = IOProcessor(in_processor, out_processor)
# and call the processing function
args.func(processor, **vars(args))
def main():
parser = argparse.ArgumentParser("merges a csv of survey responses, and a sqlite3 database of responses.")
parser.add_argument("dumpfile", help="The output of \"flask dumpdb --outfile=dump.json\"")
parser.add_argument("samples", help="folder with the actual mp3 samples")
parser.add_argument('outfile', help='output file (EX: train_dataset.npz)')
parser.add_argument('--fps', action='store', type=float, default=100, help='frames per second [default=100]')
args = parser.parse_args()
trials = json.load(open(args.dumpfile, "r"))['dataset']
data = []
labels = []
sample_names = []
for trial in trials:
final_response = [r['timestamp'] for r in trial['data']['final_response']]
sample_url = trial['url']
o = urlparse(sample_url)
sample_name = os.path.split(o.path)[-1]
preprocessor = TfRhythmicGroupingPreProcessor()
infile = os.path.join(args.samples, sample_name)
print(infile)
label_processor = LabelOutputProcessor(final_response, args.fps)
# create an IOProcessor
processor = IOProcessor(preprocessor, label_processor)
sample_data, sample_labels = _process((processor, infile, None, vars(args)))
data.append(sample_data)
if sample_data.shape != data[0].shape:
print("Shapes do not match, ", data[0].shape, sample_data.shape, "Skipping.")
data.pop()
continue
labels.append(sample_labels)
sample_names.append(sample_name)
np.savez(args.outfile, x=data, labels=labels, sample_names=sample_names)
def main():
"""DBNBeatTracker"""
# define parser
p = argparse.ArgumentParser(
formatter_class=argparse.RawDescriptionHelpFormatter,
description='''
The DBNBeatTracker.py program detects all beats in an audio file according to
the method described in:
"A Multi-Model Approach to Beat Tracking Considering Heterogeneous Music
Styles"
Sebastian Böck, Florian Krebs and Gerhard Widmer.
Proceedings of the 15th International Society for Music Information
Retrieval Conference (ISMIR), 2014.
It does not use the multi-model (Section 2.2.) and selection stage (Section
2.3), i.e. this version corresponds to the pure DBN version of the
algorithm for which results are given in Table 2.
Instead of the originally proposed state space and transition model for the
DBN, the following is used:
"An Efficient State Space Model for Joint Tempo and Meter Tracking"
Florian Krebs, Sebastian Böck and Gerhard Widmer.
Proceedings of the 16th International Society for Music Information
Retrieval Conference (ISMIR), 2015.
This program can be run in 'single' file mode to process a single audio
file and write the detected beats to STDOUT or the given output file.
$ DBNBeatTracker.py single INFILE [-o OUTFILE]
If multiple audio files should be processed, the program can also be run
in 'batch' mode to save the detected beats to files with the given suffix.
$ DBNBeatTracker.py batch [-o OUTPUT_DIR] [-s OUTPUT_SUFFIX] FILES
If no output directory is given, the program writes the files with the
detected beats to the same location as the audio files.
The 'pickle' mode can be used to store the used parameters to be able to
exactly reproduce experiments.
''')
# version
p.add_argument('--version',
action='version',
version='DBNBeatTracker.py.2016')
# input/output options
io_arguments(p, output_suffix='.beats.txt', online=True)
ActivationsProcessor.add_arguments(p)
# signal processing arguments
SignalProcessor.add_arguments(p, norm=False, gain=0)
# peak picking arguments
DBNBeatTrackingProcessor.add_arguments(p)
NeuralNetworkEnsemble.add_arguments(p, nn_files=None)
# parse arguments
args = p.parse_args()
# set immutable arguments
args.fps = 100
# print arguments
if args.verbose:
print(args)
# input processor
if args.load:
# load the activations from file
in_processor = ActivationsProcessor(mode='r', **vars(args))
else:
# use a RNN to predict the beats
in_processor = RNNBeatProcessor(**vars(args))
# output processor
if args.save:
# save the RNN beat activations to file
out_processor = ActivationsProcessor(mode='w', **vars(args))
else:
# track the beats with a DBN
beat_processor = DBNBeatTrackingProcessor(**vars(args))
# output handler
from madmom.utils import write_events as writer
# sequentially process everything
out_processor = [beat_processor, writer]
# create an IOProcessor
processor = IOProcessor(in_processor, out_processor)
# and call the processing function
args.func(processor, **vars(args))
def main():
"""DBNBeatTracker"""
# define parser
p = argparse.ArgumentParser(
formatter_class=argparse.RawDescriptionHelpFormatter,
description='''
The DBNBeatTracker program detects all beats in an audio file according to
the method described in:
"A Multi-Model Approach to Beat Tracking Considering Heterogeneous Music
Styles"
Sebastian Böck, Florian Krebs and Gerhard Widmer.
Proceedings of the 15th International Society for Music Information
Retrieval Conference (ISMIR), 2014.
It does not use the multi-model (Section 2.2.) and selection stage (Section
2.3), i.e. this version corresponds to the pure DBN version of the
algorithm for which results are given in Table 2.
Instead of the originally proposed state space and transition model for the
DBN, the following is used:
"An Efficient State Space Model for Joint Tempo and Meter Tracking"
Florian Krebs, Sebastian Böck and Gerhard Widmer.
Proceedings of the 16th International Society for Music Information
Retrieval Conference (ISMIR), 2015.
This program can be run in 'single' file mode to process a single audio
file and write the detected beats to STDOUT or the given output file.
$ DBNBeatTracker single INFILE [-o OUTFILE]
If multiple audio files should be processed, the program can also be run
in 'batch' mode to save the detected beats to files with the given suffix.
$ DBNBeatTracker batch [-o OUTPUT_DIR] [-s OUTPUT_SUFFIX] FILES
If no output directory is given, the program writes the files with the
detected beats to the same location as the audio files.
The 'pickle' mode can be used to store the used parameters to be able to
exactly reproduce experiments.
''')
# version
p.add_argument('--version',
action='version',
version='DBNBeatTracker.2016')
# input/output options
io_arguments(p, output_suffix='.beats.txt', online=True)
ActivationsProcessor.add_arguments(p)
# signal processing arguments
SignalProcessor.add_arguments(p, norm=False, gain=0)
# peak picking arguments
DBNBeatTrackingProcessor.add_arguments(p)
NeuralNetworkEnsemble.add_arguments(p, nn_files=None)
# parse arguments
args = p.parse_args()
# set immutable arguments
args.fps = 100
# print arguments
if args.verbose:
print(args)
print("The following mesxxsage shows the args :\n", args,
"\n\n\nshows the vars(args) message:\n", vars(args))
'''
The args's message:
Namespace(correct=True, fps=100, func=<function process_online at 0x00000000055E9AC8>,
gain=0, infile=None, load=False, max_bpm=215.0, min_bpm=55.0, nn_files=None, norm=False,
num_frames=1, num_tempi=None, num_threads=1, observation_lambda=16, online=True,
origin='stream', outfile=<open file '<stdout>', mode 'w' at 0x0000000002DFB0C0>,
save=False, sep=None, threshold=0, transition_lambda=100, verbose=None)
'''
# input processor
if args.load:
# load the activations from file
in_processor = ActivationsProcessor(mode='r', **vars(args))
else:
# use a RNN to predict the beats
in_processor = RNNBeatProcessor(**vars(args))
# output processor
if args.save:
# save the RNN beat activations to file
out_processor = ActivationsProcessor(mode='w', **vars(args))
else:
# track the beats with a DBN and output them
beat_processor = DBNBeatTrackingProcessor(**vars(args))
out_processor = [beat_processor, write_beats]
print("ok")
# create an IOProcessor
processor = IOProcessor(in_processor, out_processor)
# and call the processing function
args.func(processor, **vars(args))
"Otherwise the model's performance will be very poor.\n")
# initialize network
print("Initializing tagging network ...")
model = select_model(args.model)
net = model.build_model(batch_size=1)
# initialize neural network
TAGGER = Network(net, print_architecture=False)
# load model parameters network
if args.params is not None:
dump_file = args.params
else:
out_path = os.path.join(os.path.join(EXP_ROOT), model.EXP_NAME)
dump_file, log_file = get_dump_file_paths(out_path, 1)
dump_file = dump_file.replace(".pkl", "_it0.pkl")
print("Loading model from: %s" % dump_file)
TAGGER.load(dump_file)
# set prediction rate
PREDICT_EVERY_K = args.predict_every_k
# dummy prediction to compile model
print("Compiling prediction function ...")
TAGGER.predict(SLIDING_WINDOW[np.newaxis, np.newaxis])
print("Starting prediction loop ...")
processor = IOProcessor(in_processor=processor_pipeline2, out_processor=output_processor)
process_online(processor, infile=None, outfile=None, sample_rate=32000)
def main():
parser = argparse.ArgumentParser(
"merges a csv of survey responses, and a sqlite3 database of responses."
)
parser.add_argument(
"dumpfile", help="The output of \"flask dumpdb --outfile=dump.json\"")
parser.add_argument("samples", help="folder with the actual mp3 samples")
parser.add_argument('outfile', help='output file (EX: train_dataset.npz)')
parser.add_argument('--fps',
action='store',
type=float,
default=100,
help='frames per second [default=100]')
args = parser.parse_args()
experiments = util.load_by_url(args.dumpfile)
responses_by_url = util.get_final_responses_list(experiments)
# trials = json.load(open(args.dumpfile, "r"))['dataset']
n_features = 314
n_frames = 800
data = []
labels = []
sample_names = []
for sample_url, final_responses in responses_by_url.items():
o = urlparse(sample_url)
sample_name = os.path.split(o.path)[-1]
preprocessor = TfRhythmicGroupingPreProcessor()
infile = os.path.join(args.samples, sample_name)
print(infile)
label_processor = LabelOutputProcessor(final_responses, args.fps)
# create an IOProcessor
processor = IOProcessor(preprocessor, label_processor)
sample_data, sample_labels = _process(
(processor, infile, None, vars(args)))
if sample_data.shape[0] < n_frames or sample_data.shape[
1] != n_features:
print("SKIPPING: Shapes is {} but it must be (>={}, {}), ".format(
sample_data.shape, n_frames, n_features))
continue
sample_data = np.expand_dims(sample_data, axis=2)
data.append(sample_data[:n_frames])
labels.append(sample_labels[:n_frames])
sample_names.append(sample_name)
print("DATASET:")
print('{} clips'.format(len(data)))
print('{} frames of audio at {} fps'.format(n_frames, args.fps))
print('{} features per frame'.format(n_features))
print("saving {} ...".format(args.outfile))
data = np.array(data)
labels = np.array(labels)
np.savez(args.outfile, x=data, labels=labels, sample_names=sample_names)
print("done.")