1.3 Compression

# Compression

- Compression is the process of reducing file size without a significant loss in quality which results in 
  - Reducing the time needed to search for data. 
  - Faster transfer of compressed files, which uses less bandwidth than uncompressed files. 
- Lossless Compression 
  - Type of compression that allows original data to perfectly reconstructed from compressed file when the file is opened by utilizing some form of replacement. 
  - E.g. bitmap (.bmp), vector graphic (.svg) and .png images, text file compression, database records 
  - Run-length Encoding (RLE) 
    - Form of lossless compression which is used for compressing text files and bitmap images. 
    - Reduces file size of a sequence of elements which has adjacent, identical elements (characters in text file and pixels in bitmap images). 
    - Repeating sequence of elements encoded in two values: The first value represents the number of identical data items (such as characters) in the run. The second value represents the code of the data item 
  -  E.g. RLE of bitmap image: 
    - We can represent the first row as a sequence of pixels: “W B B W W B B W” | W: white and B: black 
    - After applying RLE: “W 2B 2W 2B W”. 
    - In ‘2B’ 2 is the run count and B is the run value, which represents a run of two adjacent black pixels 
    - Process is repeated for other rows.

- Lossy Compression 
    - file compression method where parts of the original file cannot be recovered during decompression, so some of the original detail is lost. 
    - Type of compression which irreversibly eliminates unnecessary data 
    - File accuracy/quality lower than that of lossless but file size is lower (~10% of lossless). 
    - E.g. Sound files (.mp3), .jpeg images 
    - Sound files compression (.mp3) utilizes Perceptual Coding to remove certain parts of sound that are less audible/discernible to human hearing.

METHODS ON FILE COMPRESSION

1. Movie Files

- reduce the sampling rate used: fewer samples taken per second means less data is being stored 
  - Reduce the sampling resolution: fewer bits are used to represent each sample, so less data is stored 
  - Reduce frame rate

2. Image Files

- crop the image 
  - Decrease color/bit depth 
  - Reduce image resolution