mirror of
https://github.com/correl/dejavu.git
synced 2024-12-29 11:09:26 +00:00
Refactored the fingerprint module
This commit is contained in:
Vin
parent
9aa80ff87c
commit
6a6ae94e3d
1 changed files with 116 additions and 105 deletions
221
dejavu/fingerprint.py
Normal file → Executable file
221
dejavu/fingerprint.py
Normal file → Executable file
|
|
@ -13,19 +13,117 @@ import time
|
|||
import hashlib
|
||||
import pickle
|
||||
|
||||
IDX_FREQ_I = 0
|
||||
IDX_TIME_J = 1
|
||||
|
||||
DEFAULT_FS = 44100
|
||||
DEFAULT_WINDOW_SIZE = 4096
|
||||
DEFAULT_OVERLAP_RATIO = 0.5
|
||||
DEFAULT_FAN_VALUE = 15
|
||||
|
||||
DEFAULT_AMP_MIN = 10
|
||||
PEAK_NEIGHBORHOOD_SIZE = 20
|
||||
MIN_HASH_TIME_DELTA = 0
|
||||
|
||||
def fingerprint(channel_samples,
|
||||
Fs=DEFAULT_FS,
|
||||
wsize=DEFAULT_WINDOW_SIZE,
|
||||
wratio=DEFAULT_OVERLAP_RATIO,
|
||||
fan_value=DEFAULT_FAN_VALUE,
|
||||
amp_min=DEFAULT_AMP_MIN):
|
||||
"""
|
||||
FFT the channel, log transform output, find local maxima, then return
|
||||
locally sensitive hashes.
|
||||
"""
|
||||
# FFT the signal and extract frequency components
|
||||
arr2D = mlab.specgram(
|
||||
channel_samples,
|
||||
NFFT=wsize,
|
||||
Fs=Fs,
|
||||
window=mlab.window_hanning,
|
||||
noverlap=int(wsize * wratio))[0]
|
||||
|
||||
# apply log transform since specgram() returns linear array
|
||||
arr2D = 10 * np.log10(arr2D)
|
||||
arr2D[arr2D == -np.inf] = 0 # replace infs with zeros
|
||||
|
||||
# find local maxima
|
||||
local_maxima = get_2D_peaks(arr2D, plot=False, amp_min=amp_min)
|
||||
|
||||
# return hashes
|
||||
return generate_hashes(local_maxima, fan_value=fan_value)
|
||||
|
||||
def get_2D_peaks(arr2D, plot=False, amp_min=DEFAULT_AMP_MIN):
|
||||
|
||||
# http://docs.scipy.org/doc/scipy/reference/generated/scipy.ndimage.morphology.iterate_structure.html#scipy.ndimage.morphology.iterate_structure
|
||||
struct = generate_binary_structure(2, 1)
|
||||
neighborhood = iterate_structure(struct, PEAK_NEIGHBORHOOD_SIZE)
|
||||
|
||||
# find local maxima using our fliter shape
|
||||
local_max = maximum_filter(arr2D, footprint=neighborhood) == arr2D
|
||||
background = (arr2D == 0)
|
||||
eroded_background = binary_erosion(background, structure=neighborhood, border_value=1)
|
||||
detected_peaks = local_max - eroded_background # this is a boolean mask of arr2D with True at peaks
|
||||
|
||||
# extract peaks
|
||||
amps = arr2D[detected_peaks]
|
||||
j, i = np.where(detected_peaks)
|
||||
|
||||
# filter peaks
|
||||
amps = amps.flatten()
|
||||
peaks = zip(i, j, amps)
|
||||
peaks_filtered = [x for x in peaks if x[2] > amp_min] # freq, time, amp
|
||||
|
||||
# get indices for frequency and time
|
||||
frequency_idx = [x[1] for x in peaks_filtered]
|
||||
time_idx = [x[0] for x in peaks_filtered]
|
||||
|
||||
if plot:
|
||||
# scatter of the peaks
|
||||
fig, ax = plt.subplots()
|
||||
ax.imshow(arr2D)
|
||||
ax.scatter(time_idx, frequency_idx)
|
||||
ax.set_xlabel('Time')
|
||||
ax.set_ylabel('Frequency')
|
||||
ax.set_title("Spectrogram of \"Blurred Lines\" by Robin Thicke");
|
||||
plt.gca().invert_yaxis()
|
||||
plt.show()
|
||||
|
||||
return zip(frequency_idx, time_idx)
|
||||
|
||||
def generate_hashes(peaks, fan_value=DEFAULT_FAN_VALUE):
|
||||
"""
|
||||
Hash list structure:
|
||||
sha1-hash[0:20] time_offset
|
||||
[(e05b341a9b77a51fd26, 32), ... ]
|
||||
"""
|
||||
fingerprinted = set() # to avoid rehashing same pairs
|
||||
hashes = []
|
||||
|
||||
for i in range(len(peaks)):
|
||||
for j in range(fan_value):
|
||||
if i+j < len(peaks) and not (i, i+j) in fingerprinted:
|
||||
|
||||
freq1 = peaks[i][IDX_FREQ_I]
|
||||
freq2 = peaks[i+j][IDX_FREQ_I]
|
||||
t1 = peaks[i][IDX_TIME_J]
|
||||
t2 = peaks[i+j][IDX_TIME_J]
|
||||
t_delta = t2 - t1
|
||||
|
||||
if t_delta >= MIN_HASH_TIME_DELTA:
|
||||
h = hashlib.sha1("%s|%s|%s" % (str(freq1), str(freq2), str(t_delta)))
|
||||
hashes.append((h.hexdigest()[0:20], t1))
|
||||
|
||||
# ensure we don't repeat hashing
|
||||
fingerprinted.add((i, i+j))
|
||||
return hashes
|
||||
|
||||
# TODO: move all of the below to a class with DB access
|
||||
|
||||
|
||||
class Fingerprinter():
|
||||
|
||||
IDX_FREQ_I = 0
|
||||
IDX_TIME_J = 1
|
||||
|
||||
DEFAULT_FS = 44100
|
||||
DEFAULT_WINDOW_SIZE = 4096
|
||||
DEFAULT_OVERLAP_RATIO = 0.5
|
||||
DEFAULT_FAN_VALUE = 15
|
||||
|
||||
DEFAULT_AMP_MIN = 10
|
||||
PEAK_NEIGHBORHOOD_SIZE = 20
|
||||
MIN_HASH_TIME_DELTA = 0
|
||||
|
||||
|
||||
def __init__(self, config,
|
||||
Fs=DEFAULT_FS,
|
||||
|
|
@ -55,104 +153,15 @@ class Fingerprinter():
|
|||
hashes = self.process_channel(samples, song_id=sid)
|
||||
print "Generated %d hashes" % len(hashes)
|
||||
self.db.insert_hashes(hashes)
|
||||
|
||||
|
||||
# TODO: put this in another module
|
||||
def match(self, samples):
|
||||
"""Used for matching unknown songs"""
|
||||
hashes = self.process_channel(samples)
|
||||
matches = self.db.return_matches(hashes)
|
||||
return matches
|
||||
|
||||
def process_channel(self, channel_samples, song_id=None):
|
||||
"""
|
||||
FFT the channel, log transform output, find local maxima, then return
|
||||
locally sensitive hashes.
|
||||
"""
|
||||
# FFT the signal and extract frequency components
|
||||
arr2D = mlab.specgram(
|
||||
channel_samples,
|
||||
NFFT=self.window_size,
|
||||
Fs=self.Fs,
|
||||
window=mlab.window_hanning,
|
||||
noverlap=self.noverlap)[0]
|
||||
|
||||
# apply log transform since specgram() returns linear array
|
||||
arr2D = 10 * np.log10(arr2D)
|
||||
arr2D[arr2D == -np.inf] = 0 # replace infs with zeros
|
||||
|
||||
# find local maxima
|
||||
local_maxima = self.get_2D_peaks(arr2D, plot=False)
|
||||
|
||||
# return hashes
|
||||
return self.generate_hashes(local_maxima, song_id=song_id)
|
||||
|
||||
def get_2D_peaks(self, arr2D, plot=False):
|
||||
|
||||
# http://docs.scipy.org/doc/scipy/reference/generated/scipy.ndimage.morphology.iterate_structure.html#scipy.ndimage.morphology.iterate_structure
|
||||
struct = generate_binary_structure(2, 1)
|
||||
neighborhood = iterate_structure(struct, Fingerprinter.PEAK_NEIGHBORHOOD_SIZE)
|
||||
|
||||
# find local maxima using our fliter shape
|
||||
local_max = maximum_filter(arr2D, footprint=neighborhood) == arr2D
|
||||
background = (arr2D == 0)
|
||||
eroded_background = binary_erosion(background, structure=neighborhood, border_value=1)
|
||||
detected_peaks = local_max - eroded_background # this is a boolean mask of arr2D with True at peaks
|
||||
|
||||
# extract peaks
|
||||
amps = arr2D[detected_peaks]
|
||||
j, i = np.where(detected_peaks)
|
||||
|
||||
# filter peaks
|
||||
amps = amps.flatten()
|
||||
peaks = zip(i, j, amps)
|
||||
peaks_filtered = [x for x in peaks if x[2] > self.amp_min] # freq, time, amp
|
||||
|
||||
# get indices for frequency and time
|
||||
frequency_idx = [x[1] for x in peaks_filtered]
|
||||
time_idx = [x[0] for x in peaks_filtered]
|
||||
|
||||
if plot:
|
||||
# scatter of the peaks
|
||||
fig, ax = plt.subplots()
|
||||
ax.imshow(arr2D)
|
||||
ax.scatter(time_idx, frequency_idx)
|
||||
ax.set_xlabel('Time')
|
||||
ax.set_ylabel('Frequency')
|
||||
ax.set_title("Spectrogram of \"Blurred Lines\" by Robin Thicke");
|
||||
plt.gca().invert_yaxis()
|
||||
plt.show()
|
||||
|
||||
return zip(frequency_idx, time_idx)
|
||||
|
||||
def generate_hashes(self, peaks, song_id=None):
|
||||
"""
|
||||
Hash list structure:
|
||||
sha1-hash[0:20] song_id, time_offset
|
||||
[(e05b341a9b77a51fd26, (3, 32)), ... ]
|
||||
"""
|
||||
fingerprinted = set() # to avoid rehashing same pairs
|
||||
hashes = []
|
||||
|
||||
for i in range(len(peaks)):
|
||||
for j in range(self.fan_value):
|
||||
if i+j < len(peaks) and not (i, i+j) in fingerprinted:
|
||||
|
||||
freq1 = peaks[i][Fingerprinter.IDX_FREQ_I]
|
||||
freq2 = peaks[i+j][Fingerprinter.IDX_FREQ_I]
|
||||
t1 = peaks[i][Fingerprinter.IDX_TIME_J]
|
||||
t2 = peaks[i+j][Fingerprinter.IDX_TIME_J]
|
||||
t_delta = t2 - t1
|
||||
|
||||
if t_delta >= Fingerprinter.MIN_HASH_TIME_DELTA:
|
||||
h = hashlib.sha1("%s|%s|%s" % (str(freq1), str(freq2), str(t_delta)))
|
||||
hashes.append((h.hexdigest()[0:20], (song_id, t1)))
|
||||
|
||||
# ensure we don't repeat hashing
|
||||
fingerprinted.add((i, i+j))
|
||||
return hashes
|
||||
|
||||
def insert_into_db(self, key, value):
|
||||
self.db.insert(key, value)
|
||||
|
||||
# TODO: this function has nothing to do with fingerprinting. is it needed?
|
||||
def print_stats(self):
|
||||
|
||||
iterable = self.db.get_iterable_kv_pairs()
|
||||
|
|
@ -168,10 +177,12 @@ class Fingerprinter():
|
|||
for song_id, count in counter.iteritems():
|
||||
song_name = self.song_names[song_id]
|
||||
print "%s has %d spectrogram peaks" % (song_name, count)
|
||||
|
||||
|
||||
# this does... what? this seems to only be used for the above function
|
||||
def set_song_names(self, wpaths):
|
||||
self.song_names = wpaths
|
||||
|
||||
|
||||
# TODO: put this in another module
|
||||
def align_matches(self, matches, starttime, record_seconds=0, verbose=False):
|
||||
"""
|
||||
Finds hash matches that align in time with other matches and finds
|
||||
|
|
|
|||
Loading…
Reference in a new issue