@hkjeffchan/imagemin-mozjpeg

imagemin-mozjpeg

Imagemin plugin for mozjpeg

Fork from https://github.com/imagemin/imagemin-mozjpeg Repackage to support cjs and upgrade package dependencies Removed binary from the package so that you can link to the most updated binary yourself !!! You should install the mozjpeg library manually

Install

You need to install mozjpeg in your system first !!!

$ npm install imagemin-mozjpeg

Usage

import imagemin from "imagemin";
import imageminMozjpeg from "imagemin-mozjpeg";

(async () => {
	await imagemin(["images/*.jpg"], {
		binaryPath: "/opt/homebrew/opt/mozjpeg/bin/cjpeg",
		destination: "build/images",
		plugins: [imageminMozjpeg()],
	});

	console.log("Images optimized");
})();

API

imageminMozjpeg(options?)(buffer)

Returns a Promise<Buffer>.

options

Type: object

binaryPath

Type: string

The full path to the mozjpeg executable

quality

Type: number

Compression quality, in range 0 (worst) to 100 (perfect).

progressive

Type: boolean
Default: true

false creates baseline JPEG file.

targa

Type: boolean
Default: false

Input file is Targa format (usually not needed).

revert

Type: boolean
Default: false

Revert to standard defaults instead of mozjpeg defaults.

fastCrush

Type: boolean
Default: false

Disable progressive scan optimization.

dcScanOpt

Type: number
Default: 1

Set DC scan optimization mode.

• 0 One scan for all components
• 1 One scan per component
• 2 Optimize between one scan for all components and one scan for 1st component plus one scan for remaining components

trellis

Type: boolean
Default: true

Trellis optimization.

trellisDC

Type: boolean
Default: true

Trellis optimization of DC coefficients.

tune

Type: string
Default: hvs-psnr

Set Trellis optimization method. Available methods: psnr, hvs-psnr, ssim, ms-ssim

overshoot

Type: boolean
Default: true

Black-on-white deringing via overshoot.

arithmetic

Type: boolean
Default: false

Use arithmetic coding.

dct

Type: string
Default: int

Set DCT method:

• int Use integer DCT
• fast Use fast integer DCT (less accurate)
• float Use floating-point DCT

quantBaseline

Type: boolean
Default: false

Use 8-bit quantization table entries for baseline JPEG compatibility.

quantTable

Type: number

Use predefined quantization table.

• 0 JPEG Annex K
• 1 Flat
• 2 Custom, tuned for MS-SSIM
• 3 ImageMagick table by N. Robidoux
• 4 Custom, tuned for PSNR-HVS
• 5 Table from paper by Klein, Silverstein and Carney

smooth

Type: number

Set the strength of smooth dithered input. (1...100)

maxMemory

Type: number

Set the maximum memory to use in kilobytes.

sample

Type: string[]

Set component sampling factors. Each item should be in the format HxV, for example 2x1.

buffer

Type: buffer

Buffer to optimize.