July 5, 2016

Huffman Tutorial Part 5

  1. Quick Run-Down of How it Works
  2. Testing
  3. Assignment for This Part
  4. Final Words

If you have arrived at this point, it is assumed that you have slaved through the parts previous to this one in the series. You may have spent a lot of time - ranging from a week to half a month - on this (rather large) endeavor. And if you have made it through, :confetti_ball: CONGRATULATIONS :confetti_ball:! Good work. Now, here is a way to test that your encoder works: by decoding it!

Quick Run-Down of How it Works

Should be rather straightforward:

  1. Read the binary
  2. Take up the header information
  3. Using the header information, (re)construct the Huffman tree
  4. Using the Huffman tree, construct a map mapping the binary digits to a character
  5. Using said map, iterate through the leftover binary data, and write the characters to stream (whether it be output or file)
  6. Close the streams

Because of DRY principles, you should reuse the code you wrote in your Huffman compression program (in C/C++'s case, use #include or something - try to avoid copying and pasting). This goes for (most likely) steps #3-4.


Should also be quite self-explanatory: you want to see if your Huffman compressor/decompresser works, you run a plain-text file through the compressor, producing a binary file. You run said binary file through the decompresser, producing a plain-text file. You then proceed to check if the plain-text file you produced is the same with the original file.

valgrind is your friend.

Assignment for This Part

  • :star: create Huffman decompresser with code reuse
  • :star: finish Huffman compressor/decompresser

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

Final Words

Wasn't that fun? This project teaches you about the applications of binary trees, the recursion that comes with it, and a little bit of bit manipulation.

For some reason, I like it. In fact, I like doing things that involve (low-level) bit manipulations (like Huffman and Chip-8 emulators). Ha, the only thing I seem to dislike about it is when I found out about endianness - the differences between big and little endian. Some systems (like mine) used little endian, meaning that some data-types (such as words) are stored with the least significant byte in the smallest address:

Original Data: 0xABCDEF

Address: 0000 | 0001 | 0002
Data:      EF |   CD |   AB

Address: 0000 | 0001 | 0002
Data:      AB |   CD |   EF

Hmmm... this sounds interesting. It is something that I know about, that I have a little experience in. Looks like an idea for the next tutorial/walkthrough....

Anyways, I hope you had as much fun making this project as I did! Hopefully, your teacher would award you properly, or, if not applicable, you would reward yourself with a nice healthy burst of dopamine. ~Niiiiiicce.

Tags: Huffman c++ decompression decoder