1. Topic 3: Data
Compression

2. Example: Full HD Image (1920x1080), with a 32-bit colour depth
Compression
Saving high-quality data in their raw format can be quite memory costly
Saving an image with lots of pixels/high colour depth
Saving an audio file with a high bitrate
So, we need to work out a way of reducing memory size, but keeping most of the data’s quality
Example: Full HD Image (1920x1080), with a 32-bit colour depth
Memory Size: 8.294MB

3. Compression This is where compression comes in
Compression tries to reduce memory size while keeping data quality
There are lots of compression techniques, but they fall into two categories
Lossy: some quality is lost
Lossless: no quality is lost

4. Compression Lossless Compression
Maintains data quality in its entirety
So no data is lost at all
Can still reduce overall memory size
Though doesn’t have the biggest impact

5. Compression Lossless Compression
Best used on data that we want to maintain entirely
Especially if the data is used on a computer
Computers can use pictures to train themselves to detect objects and people!
For something like this, lossless compression would work best (as the computer will need all the original data).

6. Compression Lossless Compression
There are a few lossless compression techniques
Run-Length Encoding: Used in PDFs
Huffman Coding: Used for lots of things (including PNGs)
Lempel-Ziv-Welch: Used for GIFs
You only need to know Run-Length Encoding, not the others

7. Compression Lossy Compression Will reduce or destroy data quality
Meaning data can be lost
Reduces overall memory size quite a lot
Far better at it than lossless compression

8. Compression Lossy Compression Best used for any user-consumed data
Images people will look at
Audio that people will listen to
Humans don’t tend to pick up on fine detail loss that lossy compression can result in
It’s often preferred to use lossy compression on images, audio, and video as humans won’t see the loss in detail, and the gain in decreased memory size is very worth it

9. Compression Lossy Compression
There are a few lossy compression techniques as well
MP3: Used in MP3 files (audio)
JPEG: Used in JPG files (images)
MPEG-4: Used in MP4 files (video)
Lossy compression techniques tend to be tied to file formats

10. Storing an image with the following colours:
Run-Length Encoding
There are different versions of Run-Length Encoding
Simplest method does the following
Storing an image with the following colours:
rrrrbbgg
bbbggggg
rrrrrrrr
We can compress that to:
(4, r) (2, b) (2, g)
(3, b) (5, g)
(8, r)

11. Storing an image with the following colours:
Run-Length Encoding
For every series of repeating values, we
Write the number of repeats first
Then write the value that is being repeated
Storing an image with the following colours:
rrrrbbgg
bbbggggg
rrrrrrrr
We can compress that to:
(4, r) (2, b) (2, g)
(3, b) (5, g)
(8, r)

12. Storing an image with the following colours:
Run-Length Encoding
If a colour takes 3 bytes, we can compress:
10 repeating colours (30 bytes)
Into 1 number (1 byte) and 1 colour (3 bytes)
Storing an image with the following colours:
rrrrbbgg
bbbggggg
rrrrrrrr
We can compress that to:
(4, r) (2, b) (2, g)
(3, b) (5, g)
(8, r)

13. Perform the Run-Length Compression technique on the following strings
aaaaaaaaaabbbbbbbbbbccccccbbbbb
bbbbbbbbbbbccccccaaaacccaaacccccc
bbbcccccbbbbaaaaaaccccbbbbaaacccc
ANSWERS
From top-left to bottom-right:
(10, a) (10, b) (6, c) (5, b)
(11, b) (6, c) (4, a) (3, c) (3, a) (6, c)
(3, b) (5, c) (4, b) (6, a) (4, c) (4, b) (3, a) (4, c)