diff --git a/picard/util/bitreader.py b/picard/util/bitreader.py
new file mode 100644
index 000000000..4828245f0
--- /dev/null
+++ b/picard/util/bitreader.py
@@ -0,0 +1,176 @@
+# -*- coding: utf-8 -*-
+#
+# Picard, the next-generation MusicBrainz tagger
+#
+# Copyright (C) 2014, 2020 Christoph Reiter
+# Copyright (C) 2019, 2021 Philipp Wolfer
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License
+# as published by the Free Software Foundation; either version 2
+# of the License, or (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+
+# This implementation is taken from mutagen, see
+# https://github.com/quodlibet/mutagen/blob/master/mutagen/_util.py
+# https://github.com/quodlibet/mutagen/blob/master/mutagen/tak.py
+
+
+class BitReaderError(Exception):
+    pass
+
+
+class _BitReader(object):
+
+    def __init__(self, fileobj):
+        self._fileobj = fileobj
+        self._buffer = 0
+        self._bits = 0
+        self._pos = fileobj.tell()
+
+    def bits(self, count):
+        """Reads `count` bits and returns an uint.
+
+        May raise BitReaderError if not enough data could be read or
+        IOError by the underlying file object.
+        """
+        raise NotImplementedError
+
+    def bytes(self, count):
+        """Returns a bytearray of length `count`. Works unaligned."""
+
+        if count < 0:
+            raise ValueError
+
+        # fast path
+        if self._bits == 0:
+            data = self._fileobj.read(count)
+            if len(data) != count:
+                raise BitReaderError("not enough data")
+            return data
+
+        return bytes(bytearray(self.bits(8) for _ in range(count)))
+
+    def skip(self, count):
+        """Skip `count` bits.
+
+        Might raise BitReaderError if there wasn't enough data to skip,
+        but might also fail on the next bits() instead.
+        """
+
+        if count < 0:
+            raise ValueError
+
+        if count <= self._bits:
+            self.bits(count)
+        else:
+            count -= self.align()
+            n_bytes = count // 8
+            self._fileobj.seek(n_bytes, 1)
+            count -= n_bytes * 8
+            self.bits(count)
+
+    def get_position(self):
+        """Returns the amount of bits read or skipped so far"""
+
+        return (self._fileobj.tell() - self._pos) * 8 - self._bits
+
+    def align(self):
+        """Align to the next byte, returns the amount of bits skipped"""
+
+        bits = self._bits
+        self._buffer = 0
+        self._bits = 0
+        return bits
+
+    def is_aligned(self):
+        """If we are currently aligned to bytes and nothing is buffered"""
+
+        return self._bits == 0
+
+
+class MSBBitReader(_BitReader):
+    """BitReader implementation which reads bits starting at LSB in each byte.
+    """
+
+    def bits(self, count):
+        """Reads `count` bits and returns an uint, MSB read first.
+
+        May raise BitReaderError if not enough data could be read or
+        IOError by the underlying file object.
+        """
+
+        if count < 0:
+            raise ValueError
+
+        if count > self._bits:
+            n_bytes = (count - self._bits + 7) // 8
+            data = self._fileobj.read(n_bytes)
+            if len(data) != n_bytes:
+                raise BitReaderError("not enough data")
+            for b in bytearray(data):
+                self._buffer = (self._buffer << 8) | b
+            self._bits += n_bytes * 8
+
+        self._bits -= count
+        value = self._buffer >> self._bits
+        self._buffer &= (1 << self._bits) - 1
+        return value
+
+
+class LSBBitReader(_BitReader):
+    """BitReader implementation which reads bits starting at LSB in each byte.
+    """
+
+    def _lsb(self, count):
+        value = self._buffer & 0xff >> (8 - count)
+        self._buffer = self._buffer >> count
+        self._bits -= count
+        return value
+
+    def bits(self, count):
+        """Reads `count` bits and returns an uint, LSB read first.
+
+        May raise BitReaderError if not enough data could be read or
+        IOError by the underlying file object.
+        """
+        if count < 0:
+            raise ValueError
+
+        value = 0
+        if count <= self._bits:
+            value = self._lsb(count)
+        else:
+            # First read all available bits
+            shift = 0
+            remaining = count
+            if self._bits > 0:
+                remaining -= self._bits
+                shift = self._bits
+                value = self._lsb(self._bits)
+
+            # Now add additional bytes
+            n_bytes = (remaining - self._bits + 7) // 8
+            data = self._fileobj.read(n_bytes)
+            if len(data) != n_bytes:
+                raise BitReaderError("not enough data")
+            for b in bytearray(data):
+                if remaining > 8:  # Use full byte
+                    remaining -= 8
+                    value = (b << shift) | value
+                    shift += 8
+                else:
+                    self._buffer = b
+                    self._bits = 8
+                    b = self._lsb(remaining)
+                    value = (b << shift) | value
+
+        return value
diff --git a/picard/util/imageinfo.py b/picard/util/imageinfo.py
index ad381ef7a..ea552976e 100644
--- a/picard/util/imageinfo.py
+++ b/picard/util/imageinfo.py
@@ -24,7 +24,7 @@
 from io import BytesIO
 import struct
 
-from mutagen.tak import _LSBBitReader as LSBBitReader
+from picard.util.bitreader import LSBBitReader
 
 
 class IdentificationError(Exception):
diff --git a/test/test_util_bitreader.py b/test/test_util_bitreader.py
new file mode 100644
index 000000000..413ecdb41
--- /dev/null
+++ b/test/test_util_bitreader.py
@@ -0,0 +1,45 @@
+# -*- coding: utf-8 -*-
+#
+# Picard, the next-generation MusicBrainz tagger
+#
+# Copyright (C) 2021 Philipp Wolfer
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License
+# as published by the Free Software Foundation; either version 2
+# of the License, or (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+
+from io import BytesIO
+
+from test.picardtestcase import PicardTestCase
+
+from picard.util.bitreader import (
+    LSBBitReader,
+    MSBBitReader,
+)
+
+
+class LsbBitReaderTest(PicardTestCase):
+
+    def test_msb_bit_reader(self):
+        data = BytesIO(b'\x8B\xC0\x17\x10')
+        reader = MSBBitReader(data)
+        self.assertEqual(8944, reader.bits(14))
+        self.assertEqual(369, reader.bits(14))
+        self.assertEqual(0, reader.bits(4))
+
+    def test_lsb_bit_reader(self):
+        data = BytesIO(b'\x8B\xC0\x17\x10')
+        reader = LSBBitReader(data)
+        self.assertEqual(139, reader.bits(14))
+        self.assertEqual(95, reader.bits(14))
+        self.assertEqual(1, reader.bits(4))