https://github.com/friendlymatthew/jpeg

Implements jpeg algorithms using portable SIMD.
https://github.com/friendlymatthew/jpeg

jfif jpeg simd

Last synced: about 1 year ago
JSON representation

Implements jpeg algorithms using portable SIMD.

• Host: GitHub
• URL: https://github.com/friendlymatthew/jpeg
• Owner: friendlymatthew
• Created: 2024-05-22T21:36:11.000Z (about 2 years ago)
• Default Branch: master
• Last Pushed: 2024-12-27T15:24:33.000Z (over 1 year ago)
• Last Synced: 2025-04-20T22:32:36.643Z (about 1 year ago)
• Topics: jfif, jpeg, simd
• Language: Rust
• Homepage:
• Size: 895 KB
• Stars: 7
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# jpeg-decoder

A decoder following the [ITU-T.81 Specification](https://www.w3.org/Graphics/JPEG/itu-t81.pdf).






    

    



```rust

use jpeg::prelude::*;

fn main() -> anyhow::Result<()> {

    let mut decoder = Decoder {

        mmap: unsafe { Mmap::map(&File::open("mike.jpg")?)? },

        cursor: 0,

        encoding: CodingProcess::BaselineDCT,

    };

    let ycbcrs: Vec = decoder.decode()?;

    let rgbs = ColorSpace::convert_ycbcr_to_rgb(ycbcrs);

    Ok(())

}

```

## Coding Process

The decoder uses baseline sequential as its coding process.

For other coding processes, see `CODING_PROCESSES.md`.

## Portable SIMD

`jpeg` makes use of portable SIMD throughout various steps of the decoding process, as well as color space conversion

and grayscaling. For example, consider dequantization:



This can be easily vectorized, simply cast the quantization table and coefficients into `Simd` and perform a

lane-multiplication with each other.

Such trivial cases also exist during coefficient level-shifting and marker segment parsing.