warn when file is too short, fixes slhck#87

ffmpeg_normalize/_cmd_utils.py

@@ -15,6 +15,8 @@
     CUR_OS.startswith(i) for i in
     ['CYGWIN', 'MSYS', 'Linux', 'Darwin', 'SunOS', 'FreeBSD', 'NetBSD'])
 NUL = 'NUL' if IS_WIN else '/dev/null'
+DUR_REGEX = re.compile(r'Duration: (?P<hour>\d{2}):(?P<min>\d{2}):(?P<sec>\d{2})\.(?P<ms>\d{2})')
+TIME_REGEX = re.compile(r'out_time=(?P<hour>\d{2}):(?P<min>\d{2}):(?P<sec>\d{2})\.(?P<ms>\d{2})')
 
 # https://gist.github.com/Hellowlol/5f8545e999259b4371c91ac223409209
 def to_ms(s=None, des=None, **kwargs):
@@ -35,9 +37,6 @@ def to_ms(s=None, des=None, **kwargs):
     return result
 
 class CommandRunner():
-    DUR_REGEX = re.compile(r'Duration: (?P<hour>\d{2}):(?P<min>\d{2}):(?P<sec>\d{2})\.(?P<ms>\d{2})')
-    TIME_REGEX = re.compile(r'out_time=(?P<hour>\d{2}):(?P<min>\d{2}):(?P<sec>\d{2})\.(?P<ms>\d{2})')
-
     def __init__(self, cmd, dry=False):
         self.cmd = cmd
         self.dry = dry
@@ -76,12 +75,12 @@ def run_ffmpeg_command(self):
             stderr.append(line.strip())
             self.output = "\n".join(stderr)
 
-            if not total_dur and CommandRunner.DUR_REGEX.search(line):
-                total_dur = CommandRunner.DUR_REGEX.search(line).groupdict()
+            if not total_dur and DUR_REGEX.search(line):
+                total_dur = DUR_REGEX.search(line).groupdict()
                 total_dur = to_ms(**total_dur)
                 continue
             if total_dur:
-                result = CommandRunner.TIME_REGEX.search(line)
+                result = TIME_REGEX.search(line)
                 if result:
                     elapsed_time = to_ms(**result.groupdict())
                     yield int(elapsed_time / total_dur * 100)

ffmpeg_normalize/_media_file.py

@@ -7,7 +7,7 @@
 
 from ._streams import AudioStream, VideoStream, SubtitleStream
 from ._errors import FFmpegNormalizeError
-from ._cmd_utils import NUL, CommandRunner
+from ._cmd_utils import NUL, CommandRunner, DUR_REGEX, to_ms
 from ._logger import setup_custom_logger
 logger = setup_custom_logger('ffmpeg_normalize')
 
@@ -69,8 +69,14 @@ def parse_streams(self):
 
         output_lines = [line.strip() for line in output.split('\n')]
 
+        duration = None
         for line in output_lines:
 
+            if 'Duration' in line:
+                duration = DUR_REGEX.search(line).groupdict()
+                duration = to_ms(**duration) / 1000
+                logger.debug("Found duration: " + str(duration) + " s")
+
             if not line.startswith('Stream'):
                 continue
 
@@ -82,13 +88,14 @@ def parse_streams(self):
             else:
                 continue
 
+
             if 'Audio' in line:
                 logger.debug("Found audio stream at index {}".format(stream_id))
                 sample_rate_match = re.search(r'(\d+) Hz', line)
                 sample_rate = int(sample_rate_match.group(1)) if sample_rate_match else None
                 bit_depth_match = re.search(r's(\d+)p?,', line)
                 bit_depth = int(bit_depth_match.group(1)) if bit_depth_match else None
-                self.streams['audio'][stream_id] = AudioStream(self, stream_id, sample_rate, bit_depth)
+                self.streams['audio'][stream_id] = AudioStream(self, stream_id, sample_rate, bit_depth, duration)
 
             elif 'Video' in line:
                 logger.debug("Found video stream at index {}".format(stream_id))

ffmpeg_normalize/_streams.py

@@ -33,7 +33,7 @@ def __init__(self, media_file, stream_id):
         super(SubtitleStream, self).__init__(media_file, 'subtitle', stream_id)
 
 class AudioStream(MediaStream):
-    def __init__(self, media_file, stream_id, sample_rate=None, bit_depth=None):
+    def __init__(self, media_file, stream_id, sample_rate=None, bit_depth=None, duration=None):
         super(AudioStream, self).__init__(media_file, 'audio', stream_id)
 
         self.loudness_statistics = {
@@ -44,6 +44,14 @@ def __init__(self, media_file, stream_id, sample_rate=None, bit_depth=None):
 
         self.sample_rate = sample_rate
         self.bit_depth = bit_depth
+        self.duration = duration
+
+        if self.duration and self.duration <= 3:
+            logger.warn(
+                "Audio stream has a duration of less than 3 seconds. "
+                "Normalization may not work. "
+                "See https://github.com/slhck/ffmpeg-normalize/issues/87 for more info."
+            )
 
     def __repr__(self):
         return "<{}, audio stream {}>".format(

test/test.py

@@ -141,6 +141,10 @@ def test_progress(self):
         self.assertTrue("100" in output)
         self.assertTrue(os.path.isfile('normalized/test.mkv'))
 
+    def test_duration(self):
+        output, _ = ffmpeg_normalize_call(['test/test.wav', '--debug'])
+        self.assertTrue("Found duration: " in output)
+
     def tearDown(self):
         for file in ['test.mkv', 'test.wav', 'test.mp3', 'test.aac', 'test.mp4']:
             if os.path.isfile('normalized/' + file):