Huffman Tutorial Part 0

05 Jul 2016

Tags: c++ Huffman encoder decoder compression


Are you curious in how compression systems work? Do you want to know how to compress and decompress files? Has your high school teacher assigned you the oh-so-complicated Huffman compression project? If you answer yes to any of these questions, good for you!

Through this tutorial series, you will learn the theory behind implementing a Huffman compressor/decompressor. Why only theory? That’s because my high school teacher asked me not to post assignments online; to stop students from ripping it off github, and using it as their own work.

The most code I’ll give would be snippets to demonstrate concepts and theory and examples; they are based off of my original project. If this disclaimer doesn’t discourage you, and you are still reading this, well, we’re getting to the good part.

What We’ll be Building

This tutorial is designed to guide you in the right direction - through my thought process in implementing this whole monstrosity that is the Huffman project. At the end of each part, there will be an assignments section for the reader to complete. It will involve building a working Huffman encoder. Obviously, it is completely optional, but it does help the process of understanding. And of course, you have to write one if your teacher gives this to you as an assignment.


An understanding of C++ syntax would be helpful in understanding the snippets of code presented. An understanding of linked lists and trees are also helpful. What this means is that you could probably get the gist of how trees work by reading the tutorial, but there is no guarantee. You can skip those parts.


Here is what parts of the project we’ll be focusing on in the next few parts:

Huffman Encoder

  1. basic setup, user stories, and file reading
  2. tree creation and character frequencies
  3. custom Huffman headers and data formats
  4. bit manipulation and file writing

Huffman Decoder

  1. code reuse and final words

How Huffman Compression Works

Here is a helpful wikipedia article to help you on that. Here is a helpful video that explains it nicely. If you are still stuck, well, I'll try my best to explain. We’ll just talk about important points that need to be solved.

What people have trouble in is how Huffman compression recognizes the end of file. When a file is compressed via Huffman compression, the output is a stream of bits. The problem comes when you try to write the bits to a file - it doesn’t work. Computers do memory in bytes. It would be really nice if the number of bytes you are trying to write to a file is a multiple of 8, but that isn’t usually the case. Thus, solutions have been proposed, and here are the 2 that I’ve heard of:

  1. Specify the number of extra bits so that the program ignores them
  2. Have specific “pseudo end-of-file” character that is appended onto every binary such that the program ignores any bits after said character

The 2nd choice seemed to make more sense; the 1st one would have me thinking about where I should put the number. I can just append the character to the character frequencies part of my header. It takes up slightly more space, but I wasn’t optimizing that much. We’ll go into this more in part 2.

Assignment for This Part

(:star: denotes a challenging task. :star2: denotes an even more challenging task.)