diff --git a/mopidy/audio/scan.py b/mopidy/audio/scan.py
index a1c15e2a..4540bc05 100644
--- a/mopidy/audio/scan.py
+++ b/mopidy/audio/scan.py
@@ -40,6 +40,10 @@ class Scanner(object):
         finally:
             self._reset()
 
+        # TODO: this should be an option or just moved out.
+        if data[gst.TAG_DURATION] < 100:
+            raise exceptions.ScannerError(
+                'Rejecting file with less than 100ms audio data.')
         return data
 
     def _setup(self, uri):
diff --git a/mopidy/scanner.py b/mopidy/scanner.py
index dd21fdb4..56639bc9 100644
--- a/mopidy/scanner.py
+++ b/mopidy/scanner.py
@@ -19,9 +19,9 @@ sys.argv[1:] = []
 import pygst
 pygst.require('0.10')
 import gst
-import gst.pbutils
 
 from mopidy import config as config_lib, exceptions, ext
+from mopidy.audio import scan
 from mopidy.models import Track, Artist, Album
 from mopidy.utils import log, path, versioning
 
@@ -103,11 +103,12 @@ def main():
     logging.info('Found %d new or modified tracks.', len(uris_update))
     logging.info('Scanning new and modified tracks.')
 
-    scanner = Scanner(config['local']['scan_timeout'])
+    scanner = scan.Scanner(config['local']['scan_timeout'])
     for uri in uris_update:
         try:
             data = scanner.scan(uri)
             data[b'mtime'] = os.path.getmtime(path.uri_to_path(uri))
+            # TODO: check minumum time
             track = translator(data)
             local_updater.add(track)
             logging.debug('Added %s', track.uri)
@@ -183,46 +184,5 @@ def translator(data):
     return Track(**track_kwargs)
 
 
-class Scanner(object):
-    def __init__(self, timeout=1000):
-        self.discoverer = gst.pbutils.Discoverer(timeout * 1000000)
-
-    def scan(self, uri):
-        try:
-            info = self.discoverer.discover_uri(uri)
-        except gobject.GError as e:
-            # Loosing traceback is non-issue since this is from C code.
-            raise exceptions.ScannerError(e)
-
-        data = {}
-        audio_streams = info.get_audio_streams()
-
-        if not audio_streams:
-            raise exceptions.ScannerError('Did not find any audio streams.')
-
-        for stream in audio_streams:
-            taglist = stream.get_tags()
-            if not taglist:
-                continue
-            for key in taglist.keys():
-                # XXX: For some crazy reason some wma files spit out lists
-                # here, not sure if this is due to better data in headers or
-                # wma being stupid. So ugly hack for now :/
-                if type(taglist[key]) is list:
-                    data[key] = taglist[key][0]
-                else:
-                    data[key] = taglist[key]
-
-        # Never trust metadata for these fields:
-        data[b'uri'] = uri
-        data[b'duration'] = info.get_duration() // gst.MSECOND
-
-        if data[b'duration'] < 100:
-            raise exceptions.ScannerError(
-                'Rejecting file with less than 100ms audio data.')
-
-        return data
-
-
 if __name__ == '__main__':
     main()
diff --git a/tests/scanner_test.py b/tests/scanner_test.py
index 1102c525..47cc8116 100644
--- a/tests/scanner_test.py
+++ b/tests/scanner_test.py
@@ -3,8 +3,9 @@ from __future__ import unicode_literals
 import unittest
 
 from mopidy import exceptions
+from mopidy.audio.scan import Scanner
 from mopidy.models import Track, Artist, Album
-from mopidy.scanner import Scanner, translator
+from mopidy.scanner import translator
 from mopidy.utils import path as path_lib
 
 from tests import path_to_data_dir