Skip to content


Bitstream is a Python library to manage binary data as bitstreams:

>>> from bitstream import BitStream
>>> BitStream(b"Hello World!")

If you need to deal with existing binary file formats, or design your own binary formats or experiment with data compression algorithms, etc. and if the Python standard library doesn't work for you, you may be interested in bitstream. Read this section and have a look at the example applications to see if it is what you need.

The main features are:

Easy to use

Bitstreams are a simple abstraction. They behave like communication channels: you can only write data at one end of it and read data at the other end, in the same order. So you only need to know how to create a stream, write into it and read it to use this library:

>>> stream = BitStream()
>>> stream.write(b"Hello")
>>> stream.write(b" World!")
>>>, 5) # doctest: +BYTES
>>>, 7) # doctest: +BYTES
b' World!'

This simple way to manage binary data is good enough for a surprisingly large number of use cases. It should be much easier to use than struct and array, the modules that the standard Python library provides for this task.

Works at the bit and byte level.

Compact codes (for example Huffman codes) do not always represent data with an entire number of bytes. Since bitstream supports bits and not merely bytes, such codes are implemented with the same API. For example, the unary coding of a sequence natural numbers requires only a few lines:

>>> data = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> stream = BitStream()
>>> for number in data:
...     stream.write(number * [True] + [False])
>>> stream

Supports Python & NumPy types

BitStream has built-in support for the common data types with a standard binary layout: bools, bytes, fixed-size integers and floating-point integers.

>>> stream = BitStream()
>>> stream.write(True, bool)
>>> stream.write(False, bool)
>>> from numpy import int8
>>> stream.write(-128, int8)
>>> stream.write(b"AB", bytes)
>>> stream
>>>, 2)
[True, False]
>>>, 1)
array([-128], dtype=int8)
>>>, 2) # doctest: +BYTES

NumPy arrays are a convenient way to deal with sequences of homogeneous data:

>>> from numpy import *
>>> dt = 1.0 / 44100.0
>>> t = r_[0.0:1.0:dt]
>>> data = cos(2*pi*440.0*t)       
>>> stream = BitStream(data)

Refer to the Built-in types section for more details.

Advanced features

  • Performance. Bitstream is a Python C-extension module that has been optimized for the common use cases. Hopefully, it will be fast enough for your needs! Under the hood, the Cython language and compiler are used to generate this extension module.

  • Custom types. The list of supported types and binary representation may be enlarged at will: new readers and writers can be implemented and associated to specific data types.

    See also: Custom types.

  • Snapshots. At times, the stream abstraction is too simple, for example when you need to lookahead into the stream without consuming its content. Snapshots are an extension of the stream model that solve this kind of issue since they provide a "time machine" to restore a stream to an earlier state.

    See also: Snapshots.

Open Source

Bitstream is distributed under a MIT license. The development takes place on GitHub and releases are distributed on PyPI.