sound.length not returning correctly

[DMB3]

sound.length seems to always return 1.0 on Android.

https://groups.google.com/forum/?fromgroups=#!topic/kivy-users/Dzf2ubmxILU

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

[tito]

Branch started to fix the issue, but i got weird results. First, we can know the duration of a sound only when it's correctly loaded, any call to the MediaPlayer.getDuration will lead to error like:

E/MediaPlayer(12471): Attempt to call getDuration without a valid mediaplayer
E/MediaPlayer(12471): error (-38, 0)

The question is: should we sync the loading (as pygame does by default) on android? We might get some lag then.
Or should we normalize our Sound API to be async, and then use the Android async feature too?

Also, the duration was correct when playing, but in milliseconds. Then i tried to / 1000. to get seconds, and the division result is just a very weird number. Maybe the java part is wrong.

Anyway, will continue later.

[tito]

This is the branch: https://github.com/kivy/python-for-android/tree/fix-sound-length

[pkor]

Has there been any change for this? I'm having the same issue and sound length is critical.

[MichaelCurrie]

I'm hitting this error now, and I'm wondering if it has something to do with Python 2 implicitly casting a float to an integer. Just a thought.

p.s. Once this issue is fixed it would be nice to inform this thread about it too.

[MichaelCurrie]

@tito Does this kivy/kivy#2776 mean this is taken care of?

[MichaelCurrie]

As a workaround, I pre-saved all my sound lengths to CSV using the pydub library.

• All my sounds are .ogg so you'll have to change this if this is not the case for you.
• I've assumed the media files are all in a folder ../media relative to the location of the below file.

import os, csv
from glob import glob

from pydub import AudioSegment

from flatten import flatten

cur_dir = os.path.dirname(__file__)
media_dir = os.path.join(os.path.dirname(cur_dir), 'media')
sound_paths = (glob(os.path.join(x[0], '*.ogg')) for x in os.walk(media_dir))
# Flatten it, since the lists might contain other lists
south_paths_flattened = list(flatten(sound_paths))

with open('sound_lengths.csv', 'w') as csvfile:
    csvwriter = csv.writer(csvfile, delimiter=',', lineterminator='\n')

    for sound_path in south_paths_flattened:
        print("sound_path: %s" % sound_path)
        try:
            sound = AudioSegment.from_ogg(sound_path)
        except Exception:
            print("Could not load file.")
        else:
            csvwriter.writerow([os.path.normpath(sound_path),
                                str(len(sound) / 1000.)])

Then I load sounds using the static method load of FixedSoundLoader, just the same as if I was going to load the sound with SoundLoader.load.

class FixedSoundLoader(object):
    """
    A custom factory class that substitutes for SoundLoader to avoid issue
    https://github.com/kivy/python-for-android/issues/67

    """
    sound_lengths = {}

    @staticmethod
    def load_sound_lengths():
        with open('sound_lengths.csv', 'r') as csvfile:
            csvreader = csv.reader(csvfile, delimiter=',')
            for row_number, row in enumerate(csvreader):
                print(repr(row))
                FixedSoundLoader.sound_lengths[row[0]] = float(row[1])

    @staticmethod
    def load(sound_file_path):
        sound_file_path = os.path.normpath(sound_file_path)

        if not FixedSoundLoader.sound_lengths:
            FixedSoundLoader.load_sound_lengths()

        sound = SoundLoader.load(sound_file_path)

        try:
            sound_length = FixedSoundLoader.sound_lengths[sound_file_path]
        except KeyError:
            # If we don't have this sound on record in the CSV, hope
            # that the length method actually words
            pass
        else:
            # It's actually the _get_length method we must overrride, as per
            # https://github.com/kivy/kivy/blob/master/
            #                          kivy/core/audio/__init__.py
            sound._get_length = lambda : sound_length

        return sound

[inclement]

Closing as this issue relates to the old toolchain which is no longer supported. Please open a new issue if you experience this issue with the current master branch.