pyldpc 0.7.0

Simulation of Low Density Parity Check Codes ldpc

version 0.7.0

In Brief:

• Generates coding and decoding matrices.

• Probabilistic decoding: Belief Propagation algorithm.

• Images transmission simulation (channel model: AGWN).

• Sound transmission simulation (channel model :AGWN).

Image coding-decoding example:

Sound coding-decoding example:

Sound Transmission

Installation

From pip:

$ pip install --upgrade pyldpc

Requiries: numpy, scipy, automatically installed with pip.

Tutorials:

Jupyter notebooks:

Many changes in tutorials in v.0.7.0

• Users’ Guide:

1- LDPC Coding-Decoding Simulation

2- Images Coding-DecodingTutorial

3- Sound Coding-DecodingTutorial

4- LDPC Matrices Construction Tutorial

• For LDPC construction details:

1- pyLDPC Construction

2- LDPC Images Functions Construction

3- LDPC Sound Functions Construction

version 0.7.0

Contains:

1. Coding and decoding matrices Generators:

   • Regular parity-check matrix using Callager’s method.

   • Coding Matrix G both non-systematic and systematic.

2. Coding function adding Additive White Gaussian Noise.

3. Decoding functions using Probabilistic Decoding (Belief propagation algorithm):

   • Default BP algorithm.

   • Full-log BP algorithm.

4. Images transmission sub-module:

   • Coding and Decoding Grayscale and RGB Images.

   • Pixel by pixel coding & decoding (small matrices)

   • Row by row coding & decoding (large sparse matrices)

   • BER: Bit Error Rate function.

5. Sound transmission sub-module:

   • Coding and Decoding audio files.

   • BER_audio: Bit Error Rate function.

What’s new:

• Compatibility of scipy.sparse.csr objects (CSR format) and numpy arrays.

• Row by row image decoding (More efficient than pixel coding) using large matrices.

• 4 times faster coding.

• 5 to 10 times faster decoding.

In the upcoming versions:

• Use of large matrices (csr) in sound transmission sub-module.

• Library of ready-to-use large matrices (csr).

• Text Transmission functions.

Contact:

Please contact [e-mail] for any bug encountered / any further information.