worldveil/dejavu

Audio fingerprinting and recognition in Python

dejavu's Issues

optionally limit input data in seconds for fingerprinting

would be great to have a config option to limit the input seconds for fingerprinting.
ie, fingerprint only the first n seconds of the file.

Recognizing error for blank wave file.

Dear,
I think I meet issue after get a black wave file by arecord. Sometimes the confidence value will be get and sometimes not. Could anyone pls. have a check on this?

mp3_library/Brad-Sucks--Total-Breakdown.mp3 already fingerprinted, continuing...
Recording WAVE 'test_target/blank.wav' : Signed 16 bit Little Endian, Rate 44100 Hz, Stereo
Recording Done
test_target/another_song.wav
Confidence: 6
Not match.
test_target/another_song_2.wav
Confidence: 6
Not match.
test_target/blank.wav
Traceback (most recent call last):
File "./audio_analysis.py", line 54, in
confi_value = song['confidence']
TypeError: 'NoneType' object has no attribute 'getitem'

when number of songs beceme more 500 ,query will became very slow

when number of songs beceme more 500 ,query will became very slow ,more than several minutes,

fails to read files not in 44100hz

having a wav file in 48000hz I get this:

Traceback (most recent call last):
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
self._target(_self._args, *_self._kwargs)
File "dejavu/dejavu/control.py", line 84, in fingerprint_worker
channels = self.extract_channels(wavout_path)
File "dejavu/dejavu/control.py", line 108, in extract_channels
assert Fs == self.fingerprinter.Fs
AssertionError

Is is possible to down-sample the file to 44100? would it be still accurate?

Converter can't handle spaces in file/directory name

The converter seems to unconditionally replace any spaces ( ) in the full path with underscores (_).

This simply produces the following traceback:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "build/bdist.linux-x86_64/egg/dejavu/control.py", line 40, in fingerprint
  File "build/bdist.linux-x86_64/egg/dejavu/convert.py", line 28, in find_files
OSError: [Errno 2] No such file or directory

compare two songs

Is there a way to just compare two mp3s on the fly? Like,

dejavu.is_same_song(song1.mp3, song2.mp3)

It then returns True or False depending.

Use argparse or docopt for option parsing

I noticed that you use a lot of manually printed statement in dejavu.py to show help and parse from sys.argv. You can make it much nicer by using argparse or better yet docopt. It'll make your code much nicer, cleaner, and easier to read. :-)

Port to pysoundcard

Pysoundcard, (by the author of pyaudio) is a newer binding using CFFI, it has no problem installing in a virtualenv, and is more likely to work in pypy (because of CFFI).

https://github.com/bastibe/PySoundCard

issues with SQL backend and Unique Constraint

I'm having troubles with the Unique constraint..,or maybe I'm getting something wrong.. :S. here goes..
(talking about the plain SQL backend, not the ORM):
the schema dump looks like:

CREATE TABLE IF NOT EXISTS `fingerprints` (
  `hash` binary(10) NOT NULL,
  `song_id` mediumint(8) unsigned NOT NULL,
  `offset` int(10) unsigned NOT NULL,
  PRIMARY KEY (`hash`),
  UNIQUE KEY `song_id` (`song_id`,`offset`,`hash`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;

"hash" is a primary key, which means will be unique for each record regardless of the other columns. What happens is that I fingerprint 2 different files, and I get some hashes that are already in the DB but from the other file (with different offset). And because hash is unique, an IntegrityError is raised. Then, the hash of the new file is never saved.

Am I missing something?.

File recognizer not working as well as expected

I integrated dejavu into a web application successfully and wrote an Android application to consume the web service. My Android app records audio for 12 seconds then sends it to the web service for recognition. The web service matches the uploaded file using FileRecognizer. However, the results are highly inconsistent considering the recorded audio's quality is pretty decent. Can this be an issue with the FileRecognizer?

asking if new song is a duplicate returns a song

I ran some tests and I'm not sure what's wrong, but this always seems to happen:

So basically say I have a few completely different songs, and I add the fingerprints of a few of them. After that I check if a fingerprint is already added, by using a block of a totally new song which I haven't fingerprinted, and I always get a song id in return. If I run the recognition over a block from an already fingerprinted song, the recognition does work (pinpointing a correct song)

Could I be possibly doing something wrong?

issues with new json configuration format.

couldnt make it work with the latest changes.
Would be good to have an example configuration file, with all the possible options.

How do I setup the database server credentials now?.

dejavu to find position in the song

Hello.

If I not need to find the song, but I need to find the point in miliseconds with a offset of a song. It is posible with our library?

You know the music that are playing and you like continue in your mobile in the point.

Thank you.
Sorry my bad english

Invalid match and poor confidence

Hi,

I would like to use this library to find the position only in one song, but it's not working at all for the song I'm using. I don't know if it's the problem of the song or the algorithm, but even the tests are failing with invalid match and zero confidence.

The song is very popular (Taylor Swift - Shake It Off), but it's licensed. You can try to get it with a few commands.

youtube-dl https://www.youtube.com/watch?v=nfWlot6h_JM
ffmpeg -i Taylor\ Swift\ -\ Shake\ It\ Off-nfWlot6h_JM.mp4 Taylor\ Swift\ -\ Shake\ It\ Off-nfWlot6h_JM.wav

Then what I did was that I just modified test_dejavu.sh to scan wav files and then execute. I used wav files, because mp3 had a strange length, but wav seems to be ok.

Can you help me to fix this issue?

Thanks!

Speaker Recognition

Could this system be used for speaker recognition?

Collecting microphone signal from HTML5 MediaStreamAudioSourceNode

Hey Man! Such a great project you have done!

This is not exactly an issue, just a doubt. Do you have any example (or hint) on how to collect the audio input from a microphone over the browser and stream it to your module (via websockets or something like that, real-time) ?

I'm trying to integrate dejavu to a Django project that among other things, listen to the microphone and reacts to specific songs.

Any help would be much appreciated!

Regargs!

Failed fingerprinting

Hello I have code that previously worked but now something change(I presume fingerprint directory with pool) and I can't make it work in fingerprint directory I receive following traceback error:

Failed fingerprinting
Traceback (most recent call last):
  File "/home/krasi/work/dj/src/dejavu/dejavu/__init__.py", line 74, in fingerprint_directory
    song_name, hashes = iterator.next()
  File "/usr/lib/python2.7/multiprocessing/pool.py", line 659, in next
    raise value
OSError: [Errno 2] No such file or directory
Failed fingerprinting
Traceback (most recent call last):
  File "/home/krasi/work/dj/src/dejavu/dejavu/__init__.py", line 74, in fingerprint_directory
    song_name, hashes = iterator.next()
  File "/usr/lib/python2.7/multiprocessing/pool.py", line 659, in next
    raise value
OSError: [Errno 2] No such file or directory

FIXED: Missing ffmpeg. In ubuntu 14 it also missing from the repositories.
solved with adding custom PPA:

sudo add-apt-repository ppa:mc3man/trusty-media
sudo apt-get update
sudo apt-get install ffmpeg gstreamer0.10-ffmpeg

Meaning of output (Diff and Offset)

I have pulled Dejavu and fingerprinted a file.

What does the Diff and offset in this output mean?

"Diff is -25999 with 2 offset-aligned matches"

Thanks

OSError: [Errno 2] No such file or directory

When running the following:

directory = "/Users/jhill/Music"
extension = "mp3"
djv.fingerprint_directory(directory, ["." + extension], 1)

I get the following error:

Failed fingerprinting
Traceback (most recent call last):
  File "/Users/jhill/Sites/dejavu/dejavu/__init__.py", line 67, in fingerprint_directory
    song_name, hashes = iterator.next()
  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/multiprocessing/pool.py", line 655, in next
    raise value
OSError: [Errno 2] No such file or directory

I have confirmed that the file does indeed exist. Is the real error being obscured by the multiprocessing? What should I do to fix this error?

"TypeError" exception when a song is not recognized

when checking a song and there is no positive match, using:
..
recognizer = Recognizer(dejavu.fingerprinter, config)
res = recognizer.read("14622976.wav")

throws the following exception, would be nice if a "NotFound" exception is throwed instead.
*EDIT: or return None.

Traceback (most recent call last):
  File "checksong.py", line 14, in <module>
    res = recognizer.read("14622976.wav")
  File "dejavu/dejavu/recognize.py", line 40, in read
        return self.fingerprinter.align_matches(matches, starttime, verbose=verbose)
  File "dejavu/dejavu/fingerprint.py", line 206, in align_matches
    songname = self.db.get_song_by_id(song_id)    [SQLDatabase.FIELD_SONGNAME]
TypeError: 'NoneType' object has no attribute '__getitem__'

When new songs are added for fingerprinting, fingerprinted column of songs table doesn't get updated properly

When songs are added for fingerprinting, songs get fingerprinted and rows are added to tables, but after completing fingerprinting, fingerprinted column of songs table is not updated to 1. It remains as 0.
So after adding songs to db, when same songs are given again as input for fingerprinting, duplicate rows are added to db.

Infinite loop if MySQL server connection timed out.

When the MySQL server hangs up on the connection and dejavu tries to use it, it will print out Error in query(), 2006: MySQL server has gone away until stopped by the user.

Traceback after using an interrupt:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "dejavu/recognize.py", line 38, in read
  File "dejavu/fingerprint.py", line 62, in match

  File "dejavu/database.py", line 320, in return_matches
  File "dejavu/database.py", line 288, in query

The database connection should most likely be dropped after usage, and (re)created when needed. Instead of connecting at initialization and holding onto forever.

ambiguity in hash generation

In fingerprint.py, line 139

fingerprinted = set()  # to avoid rehashing same pairs

where elements later are defined as

# ensure we don't repeat hashing
fingerprinted.add((i, i + j))

However, this check is redundant as the combination of i and j will always be unique. What's the intention here?

song_name column in database does not have full path

I have a collection I want to fingerprint with this structure:

/music/artist/album/*.mp3

the fingerprinted records in the songs table has onyl the file names, I think should have the full path. otherwise, is not possible to locate the original file if the collection is not a flat directory.

Hashs number (fingerprints number)

Hi,
First of all, let me congratulate you for the great work.

So, i read your explanation about dejavu, and i've been testing your solution.
I noticed that the solution can generate a lot of fingerprints (240k for "Mirrors" like you said on the explanation).

I've built a database with 200+ songs, and the average number of fingerprints for each song is 8k (the maximum is 16k). First i thought it was the songs, but i also added the music "Mirrors" to my database, and it only has 10k fingerprints.

Do you know what i might have done wrong? I've used the default params to generate the fingerprints:
DEFAULT_AMP_MIN = 10
PEAK_NEIGHBORHOOD_SIZE = 20
MIN_HASH_TIME_DELTA = 0
MAX_HASH_TIME_DELTA = 200

Thank you in advance.

Cumps,
Fábio

Change Source of Raw Input

Hi, i really love your work, i was wondering one thing.
If we could read a raw input like you do with pyaudio, but instead from the Microphone we could add network streams, let say for example that we have something like this:

class Player(object):
    def __init__(self, channel):
        self.pipeline = gst.Pipeline("RadioPipe")
        self.player = gst.element_factory_make("playbin", "player")
        pulse = gst.element_factory_make("pulsesink", "pulse")
        fakesink = gst.element_factory_make("fakesink", "fakesink")
    filesink = gst.element_factory_make('filesink','sink')
    filesink.set_property('location','./output') 
        self.player.set_property('uri', channel)
        self.player.set_property("audio-sink", filesink)
        self.player.set_property("video-sink", fakesink)
    self.pipeline.add(self.player)
    def play(self):
        self.pipeline.set_state(gst.STATE_PLAYING)
    def stop(self):
        self.pipeline.set_state(gst.STATE_PAUSED)

In which i can save the output of the stream directly to the hardrive instead of a file.
Have any plans for something like this?

Allow more options to be supplied to MySQLdb.

Currently the only configuration options allowed to be changed (without relying on monkeypatching) are hostname, username, password and database.

At least having an option for supplying the port would be nice. Other options can be found at the documentation for MySQLdb.

not able to run, Failed to pin

my code is

if __name__ == '__main__':
    from dejavu import Dejavu
    con = {
    "database": {
        "host": "127.0.0.1",
        "user": "root",
        "passwd": "", 
        "db": "dejavu"
    }
}
    djv = Dejavu(con)
    djv.fingerprint_directory("mp3", [".mp3"], 3)

Failed fingerprinting

Traceback (most recent call last):

File "C:\Users\user\Desktop\dejavu\dejavu\__init__.py", line 74, in fingerprint_directory


song_name, hashes = iterator.next()

 File "C:\Python27\lib\multiprocessing\pool.py", line 659, in next

  raise value

WindowsError: [Error 2] The system cannot find the file specified

checked database,files in mp3 folder and its ok, i am using python 2.7 and installed all dependencies

Multiple files recognizer

Hello, is it possible somehow to make dejavu runs recognizer on all files in a choosen directory instead of one particular file?

Best regards

RuntimeError: could not open display

I downloaded master (tar) and extracted.
Installed all mentioned dependencies and created a dejavu.cnf file with the following contents:

config = {
"database": {
"host": "127.0.0.1",
"user": "dejavu",
"passwd": "dejavudbpass",
"db": "dejavu",
}
}

Then I run python go.py and I get the following errors:

Traceback (most recent call last):
File "go.py", line 1, in
from dejavu import Dejavu
File "/usr/src/dejavu/dejavu/init.py", line 3, in
import fingerprint
File "/usr/src/dejavu/dejavu/fingerprint.py", line 3, in
import matplotlib.pyplot as plt
File "/usr/lib64/python2.6/site-packages/matplotlib/pyplot.py", line 78, in
new_figure_manager, draw_if_interactive, show = pylab_setup()
File "/usr/lib64/python2.6/site-packages/matplotlib/backends/init.py", line 25, in pylab_setup
globals(),locals(),[backend_name])
File "/usr/lib64/python2.6/site-packages/matplotlib/backends/backend_gtkagg.py", line 10, in
from matplotlib.backends.backend_gtk import gtk, FigureManagerGTK, FigureCanvasGTK,
File "/usr/lib64/python2.6/site-packages/matplotlib/backends/backend_gtk.py", line 8, in
import gtk; gdk = gtk.gdk
File "/usr/lib64/python2.6/site-packages/gtk-2.0/gtk/init.py", line 64, in
_init()
File "/usr/lib64/python2.6/site-packages/gtk-2.0/gtk/init.py", line 52, in _init
_gtk.init_check()
RuntimeError: could not open display

I'm not a python developer, so I need some enlightenment.
I checked under fingerprint.py and I see all "plot" falues are false:

# find local maxima
local_maxima = get_2D_peaks(arr2D, plot=False, amp_min=amp_min)

def get_2D_peaks(arr2D, plot=False, amp_min=DEFAULT_AMP_MIN):

I tried adding "plot": "False" to the JSON cnf file, but same luck.
Please help!
All I need to do is make a .sh script to automatically compare some recordings with one sample.

Thanks in advance,

Alex

memory usage

I'm having issues fingerprinting a 2 hour long mp3. From what i can see, it fills up the memory (RAM) and then the script crashes. The filesize is around 140MB and I was testing in a virtual machine with ubuntu/5GB ram/3GB swap.
Any thoughts on this?

little improvement

Hi,

I've been wondering for a while now. You use sha1 to generate the fingerprints, which generates string with 40 chars, correct? But you only pick the first 20 (FINGERPRINT_REDUCTION now).
So, basically you are losing information, which will increase the change of collisions (fingerprints with different values with the same 20 chars).

So i thought... Is it not better to use longs?
since the values of the fingerprints are not very high, you can use 24 bits. and since t_delta will be between 0 and 200 by default you use 16 bits and generate the fingerprint like this:

if t_delta >= MIN_HASH_TIME_DELTA and t_delta <= MAX_HASH_TIME_DELTA: h = (freq1 << 40) + (freq2 << 16) + t_delta yield (h, t1)

also, comparing long fingerprints is faster than comparing string fingerprints, right?

Question about fps & recognition

Hi,

I test the library for recognition of series, works fine, thanks for your job.

I have 2 questions:

• It is possible to have recognition working with // FPS encoding ? Look like dvd / tv channels don't have the same FPS.
• It's there a difference between file recognition and mic recognition ? As i don't have soundcard working with my ubuntu, i can't use the mic recording, so i record with my laptop and send the file to my server for recognition.

Thanks.

pressing Ctrl-C does not finish gracefully the processes

if I try to stop the fingerprinting pressing Ctrl-C, leaves zombie processes.

^CTraceback (most recent call last):
File "test.py", line 12, in
dejavu.fingerprint("inputdata", "wav", [".mp3"], 2)
File "dejavu/dejavu/control.py", line 62, in fingerprint
p.join()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 145, in join
Process Process-1:
res = self._popen.wait(timeout)
File "/usr/lib64/python2.7/multiprocessing/forking.py", line 154, in wait
Traceback (most recent call last):
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in _bootstrap
return self.poll(0)
File "/usr/lib64/python2.7/multiprocessing/forking.py", line 135, in poll
pid, sts = os.waitpid(self.pid, flag)
KeyboardInterrupt
self.run()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
self._target(_self._args, *_self._kwargs)
File "dejavu/dejavu/control.py", line 78, in fingerprint_worker
wavout_path = self.converter.convert(filename, extension, Converter.WAV, output)
File "dejavu/dejavu/convert.py", line 47, in convert
mp3file = AudioSegment.from_mp3(orig_path)
File "pydub/pydub/audio_segment.py", line 249, in from_mp3
return cls.from_file(file, 'mp3')
File "pydub/pydub/audio_segment.py", line 236, in from_file
subprocess.call(convertion_command, stderr=open(os.devnull))
File "/usr/lib64/python2.7/subprocess.py", line 524, in call
return Popen(_popenargs, *_kwargs).wait()
File "/usr/lib64/python2.7/subprocess.py", line 1357, in wait
pid, sts = _eintr_retry_call(os.waitpid, self.pid, 0)
File "/usr/lib64/python2.7/subprocess.py", line 478, in _eintr_retry_call
return func(*args)
KeyboardInterrupt

^CError in atexit._run_exitfuncs:
Traceback (most recent call last):
File "/usr/lib64/python2.7/atexit.py", line 24, in _run_exitfuncs
func(_targs, *_kargs)
File "/usr/lib64/python2.7/multiprocessing/util.py", line 319, in _exit_function
p.join()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 145, in join
res = self._popen.wait(timeout)
File "/usr/lib64/python2.7/multiprocessing/forking.py", line 154, in wait
return self.poll(0)
File "/usr/lib64/python2.7/multiprocessing/forking.py", line 135, in poll
pid, sts = os.waitpid(self.pid, flag)
KeyboardInterrupt
Error in sys.exitfunc:
Traceback (most recent call last):
File "/usr/lib64/python2.7/atexit.py", line 24, in _run_exitfuncs
func(_targs, *_kargs)
File "/usr/lib64/python2.7/multiprocessing/util.py", line 319, in _exit_function
p.join()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 145, in join
res = self._popen.wait(timeout)
File "/usr/lib64/python2.7/multiprocessing/forking.py", line 154, in wait
return self.poll(0)
File "/usr/lib64/python2.7/multiprocessing/forking.py", line 135, in poll
pid, sts = os.waitpid(self.pid, flag)
KeyboardInterrupt

Specify a license for the project.

It would be wonderful if a license could be attached to the project so that users and contributors know what they are allowed to do with the code found in the repository.

The blog post mentioned in the readme suggests you want this to be an open source project, so using an open source license in that case would be fitting.

Downsampling

How could I downsample the input? This works great on music, but if you are only interested in human voice, you only need to go with 16000hz.

In this case, this should also optimize overall speed.

Spaces in directory names are not handled

Spaces in directory names to fingerprint is not properly handled by the code.
djv.fingerprint_directory("mp3", [".mp3"])
It simply ends without processing files or displaying any error messages.

[SQLBackend] Strange number of fingerprints saved to DB

When comparing the results of the ORMBackend and the SQLBackend, I found something strange:

for the same file (tests/test1.mp3 in my fork) this is the number of rows I get saved in the fingerprint table, for each backend:

ORM: 6457
SQL: 4851

but, the strange thing is, if I print the hashes quantity just before the insert, the total is 6457. ( 3406 for first channel and 3051 for the second one).

in "database_sql.py.insert_hashes", I can confirm that is trying to insert 6457 hashes. But when I check the mysql database, I get only 4851!.

(this can be easily reproduced running the unittests).

make fingerprinted .wav file storage optional

after fingerprinting a wav file (generated from a mp3), the wav file is not needed anymore. For matching the fingerprint, the original mp3 is enough. This way, a lot of space can be saved..

for example, I have a 10.000 songs in mp3 collection that i want to fingerprint, and keeping the wav files wastes a lot of space. After generating each fingerprint, I could just remove the wav file.

More Pythonic packaging and upload to PyPi

It would be beneficial to make this package compatible with how normal Python packages are organized and distributed. Good inspiration:

• http://www.kennethreitz.org/essays/repository-structure-and-python
• http://docs.python-guide.org/en/latest/
• https://docs.python.org/2/distutils/setupscript.html

Ken Reitz has some good example packages on his page. Then you can simply upload it to PyPi and share with the world!

module import typo in control.py

in control.py line 2:

from dejavu.converter import Converter

should be:

from dejavu.convert import Converter

Recognizing issue with Ubuntu 14.04 via microphone

I'm appreciate for you great code in audio fingerprint analyzing, but I have met some issue during microphone recognizing.
My sound cards setup is:
cat /proc/asound/cards
0 [HDMI ]: HDA-Intel - HDA Intel HDMI
HDA Intel HDMI at 0xf7c34000 irq 49
1 [PCH ]: HDA-Intel - HDA Intel PCH
HDA Intel PCH at 0xf7c30000 irq 50

And when I am trying to recognizing the music via mic, error occurs:
python dejavu.py recognize mic 20
logged here
ALSA lib pcm_dsnoop.c:618:(snd_pcm_dsnoop_open) unable to open slave
ALSA lib pcm_dmix.c:1022:(snd_pcm_dmix_open) unable to open slave
ALSA lib pcm.c:2239:(snd_pcm_open_noupdate) Unknown PCM cards.pcm.rear
ALSA lib pcm.c:2239:(snd_pcm_open_noupdate) Unknown PCM cards.pcm.center_lfe
ALSA lib pcm.c:2239:(snd_pcm_open_noupdate) Unknown PCM cards.pcm.side
bt_audio_service_open: connect() failed: Connection refused (111)
bt_audio_service_open: connect() failed: Connection refused (111)
bt_audio_service_open: connect() failed: Connection refused (111)
bt_audio_service_open: connect() failed: Connection refused (111)
ALSA lib pcm_dmix.c:1022:(snd_pcm_dmix_open) unable to open slave
Cannot connect to server socket err = No such file or directory
Cannot connect to server request channel
jack server is not running or cannot be started

Could you pls. help me to find out anything missing? Thanks.

error when fingerprinting on windows 7?

Hey guys. Any idee why i get this error when running go.py for fingerprinting?

Traceback (most recent call last):
File "", line 1, in
File "C:\Python27\lib\multiprocessing\forking.py", line 380, in main
prepare(preparation_data)
File "C:\Python27\lib\multiprocessing\forking.py", line 495, in prepare
'parents_main', file, path_name, etc
File "D:\pipi\go.py", line 17, in
djv.fingerprint_directory("", [".mp3"])
File "D:\pipi\dejavu__init__.py", line 44, in fingerprint_directory
pool = multiprocessing.Pool(nprocesses)
File "C:\Python27\lib\multiprocessing__init__.py", line 232, in Pool
return Pool(processes, initializer, initargs, maxtasksperchild)
File "C:\Python27\lib\multiprocessing\pool.py", line 159, in init
self._repopulate_pool()
File "C:\Python27\lib\multiprocessing\pool.py", line 222, in _repopulate_pool
w.start()
File "C:\Python27\lib\multiprocessing\process.py", line 130, in start
self._popen = Popen(self)
File "C:\Python27\lib\multiprocessing\forking.py", line 258, in init
cmd = get_command_line() + [rhandle]
File "C:\Python27\lib\multiprocessing\forking.py", line 358, in get_command_li
ne
Traceback (most recent call last):
File "", line 1, in
is not going to be frozen to produce a Windows executable.''')
File "C:\Python27\lib\multiprocessing\forking.py", line 380, in main
RuntimeErrorTraceback (most recent call last):
: prepare(preparation_data)
File "", line 1, in

File "C:\Python27\lib\multiprocessing\forking.py", line 495, in prepare
Attempt to start a new process before the current process
has finished its bootstrapping phase.

        This probably means that you are on Windows and you have
        forgotten to use the proper idiom in the main module:

            if __name__ == '__main__':
                freeze_support()
                ...

        The "freeze_support()" line can be omitted if the program
        is not going to be frozen to produce a Windows executable.  File "C:

\Python27\lib\multiprocessing\forking.py", line 380, in main

'__parents_main__', file, path_name, etc

prepare(preparation_data)
File "D:\pipi\go.py", line 17, in
File "C:\Python27\lib\multiprocessing\forking.py", line 495, in prepare
djv.fingerprint_directory("", [".mp3"])
'parents_main', file, path_name, etc
File "D:\pipi\dejavu__init__.py", line 44, in fingerprint_directory
File "D:\pipi\go.py", line 17, in
pool = multiprocessing.Pool(nprocesses)
File "C:\Python27\lib\multiprocessing__init__.py", line 232, in Pool
djv.fingerprint_directory("", [".mp3"])
return Pool(processes, initializer, initargs, maxtasksperchild)
File "D:\pipi\dejavu__init__.py", line 44, in fingerprint_directory
File "C:\Python27\lib\multiprocessing\pool.py", line 159, in init
pool = multiprocessing.Pool(nprocesses)
self.repopulate_pool()
File "C:\Python27\lib\multiprocessing\pool.py", line 222, in repopulate_pool
File "C:\Python27\lib\multiprocessing__init.py", line 232, in Pool
w.start()
File "C:\Python27\lib\multiprocessing\process.py", line 130, in start
self._popen = Popen(self)
File "C:\Python27\lib\multiprocessing\forking.py", line 258, in init
return Pool(processes, initializer, initargs, maxtasksperchild)
cmd = get_command_line() + [rhandle]
File "C:\Python27\lib\multiprocessing\pool.py", line 159, in init
File "C:\Python27\lib\multiprocessing\forking.py", line 358, in get_command_li
ne
self._repopulate_pool()
is not going to be frozen to produce a Windows executable.''')
RuntimeError File "C:\Python27\lib\multiprocessing\pool.py", line 222, in _repo
pulate_pool

MicrophoneRecognizer wav files - not recognizing

Hi, I have difficulty recgnize wav files from microphone. It works with mp3 both mic and file recognize, but for wav files it only works for file recognizer. I suppose that when fingerprinting with pydub the fingerprints has different parameters than defaults. Any Idea to solve that issue without converting the files to mp3.
P.S. Sorry the problem was with .pyc files. Everything is ok.

Speed improvements

Are there any possible tweaks that could be done to the algorithm to improve performance.

dejavu currently seems to net about x4 real-time fingerprinting speeds.

I've been testing with lower sample rate, different window size, overlap ratio and I think there are certainly speedups to be found if we can find the sweet spot between speed and accuracy (None of my tests have acceptable accuracy as of yet, and varying speedups).

Since I'm not too familiar with the exact algorithm I thought it might be a better idea to involve everyone in it.

The rate of hash used is low when recognizing

According to the code, when making hash,choose two point and this two point has random delta of time,but as common ,we often recognize the sound which short of 10 seconds, so if the two point which have the delta of time are more than 10 second ,the hash would never be used,which cause too many unused hashes.
if we can just choose the two point which has little delta of the time ? so when recognize a 10 second sound, all the hashes made will be affected in the database

AttributeError when song not found

Similar to issue #1, instead it now throws an AttributeError with the following traceback

>>> recognizer.read("test.wav")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "build/bdist.linux-x86_64/egg/dejavu/recognize.py", line 40, in read
  File "build/bdist.linux-x86_64/egg/dejavu/fingerprint.py", line 205, in align_matches

regression, filenames are not saved anymore with full path

previously [1], I pushed some changes, one of those it was adding the fullpath to the song name field in the database. (otherwise, would be imposible to match a file, if there are same names files in different folders).
That change was reverted with the last merge. Is this the desired behaviour for the database? or should it be pushed again?.

[1] #12

identify songs by hash to not re-fingerprint files?

does it make sense instead of checking by the filename (which is not reliable.. because the files may be moved, or renamed..), use some kind of hash without the metadata?.

ie: would this [1] work?. I think It would, but only for mp3s.., what about other formats?.

[1] https://github.com/sptim/mp3hash

fequency peaks and time

In fingerprint.py I have noticed that in line 72 that x[0] is freq , x[1] is time and x[2] is amp.
( peaks_filtered = [x for x in peaks if x[2] > amp_min] # freq, time, amp)

But in line 75 and 76 x[1] is taken as frequency and x[0] as time
frequency_idx = [x[1] for x in peaks_filtered]
time_idx = [x[0] for x in peaks_filtered]

I am confused that frequency and times are getting swapped?
when I print frequency_idx I can see values which are incrementing (similar to time)
and when I print time_idx it seems like frequency values whithout any order.