# How does a Huffman encoded message gets unified before channel encoding?

I aim to understand the signal entire transmission pipeline. Currenlty, I am facing a problem regarding the handling of a Huffman encoded message.

Example:

• Message $X$ contains symbols $a(50\%)$, $b(30\%)$ and $c(20\%)$.
• Huffman Code with code word length $2$:

\begin{align} cc &= 11 \\ bc &= 101\\ ac &= 011\\ \vdots &=\vdots\\ aa &= 0000 \end{align} There are code words with different lengths. When I go further down the signal transmission pipeline i. e. modulation or channel coding, both requisite bit packages of equal length.

Question: Where and how do the Huffman encoded code words get unified in its length?

Suggestion which doesn't work:

I always take 8 bits regardless of the encoded symbol.
$$11111111 = cccccccc$$
$$101101101 (9\text{bits})\implies\text{cutoff 1 bit} \implies 10110110 (8\text{bit}) \implies bcbc?? \\\implies\text{doesn't work}$$

• A question regarding the edit. Why did you changed "Bits" to "bits". In my understanding bits is the unit for information and Bits are binary symbols. Why aren't they binary symbols in this case?
– siva
Commented Jan 5, 2017 at 1:34

So, the output of the Huffman encoder can be regarded as a stream of 0s and 1s. The channel coder will subdivide the stream in groups of $k$ bits and convert them to groups of $n$ bits. Again, the channel coder output is a stream of bits that is processed by the modulator.