About

Converting images to true black and white (1-bit monochrome) requires more than desaturation. A grayscale image contains 256 shades; a binary black-and-white image contains exactly 2. The conversion quality depends entirely on the threshold algorithm and how mid-tones are handled. Poor threshold selection destroys detail in shadows or highlights - critical for document scanning, stencil creation, or print preparation where ink coverage must be predictable.

This tool implements three conversion methods: direct luminosity thresholding using the ITU-R BT.601 standard (L = 0.299R + 0.587G + 0.114B), Floyd-Steinberg error diffusion dithering for preserving tonal gradients, and ordered Bayer dithering for stylized halftone effects. The threshold slider provides 256 levels of control. Output is a lossless PNG with transparency preserved where applicable.

Formulas

The luminosity-weighted grayscale conversion follows the ITU-R BT.601 standard, which accounts for human eye sensitivity to green light:

L = 0.299R + 0.587G + 0.114B

Where L is the resulting luminance value (0 - 255), and R, G, B are the red, green, and blue channel values respectively. For binary thresholding, the output pixel value P is determined by:

P =

{

255 if L ≥ T0 if L < T

Where T is the user-defined threshold (0 - 255). Floyd-Steinberg dithering distributes the quantization error e = L − P to neighboring pixels using the diffusion matrix:

*716316516116

The asterisk marks the current pixel position. Ordered dithering uses a 4×4 Bayer matrix to create a repeating threshold pattern, comparing each pixel's luminance against a spatially varying threshold derived from the matrix position.

Reference Data

Conversion Mode	Best For	Preserves Gradients	File Size	Processing Speed
Binary Threshold	Documents, logos, line art	No	Smallest	Fastest
Grayscale	Photos needing tone retention	Yes (256 levels)	Medium	Fast
Floyd-Steinberg Dither	Photos, artwork with gradients	Yes (perceptual)	Small	Medium
Ordered (Bayer) Dither	Retro/comic style, screen print	Patterned	Small	Fast
Threshold Reference Values
Threshold 64	Dark images, shadow detail	-	-	-
Threshold 128	Balanced (default)	-	-	-
Threshold 192	Bright images, highlight detail	-	-	-
Color Space Standards
ITU-R BT.601	SD video, legacy images	R: 0.299	G: 0.587	B: 0.114
ITU-R BT.709	HD video, modern cameras	R: 0.2126	G: 0.7152	B: 0.0722
Common Use Cases
Document Scanning	Threshold 140 - 180	Binary mode	OCR-ready	-
Screen Printing	Threshold 100 - 130	Binary or Bayer	High contrast	-
Laser Engraving	Threshold 110 - 140	Floyd-Steinberg	Halftone effect	-
Stencil Art	Threshold 90 - 120	Binary mode	Clean edges	-
Fax Preparation	Threshold 128	Binary mode	1-bit TIFF compatible	-

Frequently Asked Questions

For text documents with white backgrounds, use threshold values between 140-180. Lower values (140-155) work better for documents with light gray backgrounds or faded text. Higher values (165-180) are optimal for clean, high-contrast printed text. Always use Binary Threshold mode rather than dithering for OCR compatibility.

Floyd-Steinberg dithering intentionally creates a grain pattern by distributing quantization error to neighboring pixels. This preserves the perception of mid-tones in a 1-bit image. The grain is more visible at low resolutions. For images intended for print at 300 DPI or higher, the dots blend visually. If you need clean edges, use Binary Threshold mode instead.

Simple averaging (R+G+B)/3 treats all color channels equally, but human vision is most sensitive to green light and least sensitive to blue. The ITU-R BT.601 formula weights green at 58.7%, red at 29.9%, and blue at 11.4%. This produces grayscale values that match perceived brightness. Using simple averaging often makes blues appear too light and greens too dark.

Yes. This tool preserves the alpha channel during conversion. Transparent pixels remain transparent in the output PNG. Semi-transparent pixels are processed normally for their RGB values while maintaining their original alpha value. For images with anti-aliased edges, consider using a higher threshold to prevent gray edge artifacts from becoming visible black pixels.

Floyd-Steinberg uses error diffusion, creating organic, noise-like patterns that closely approximate continuous tones. Ordered (Bayer) dithering uses a fixed 4×4 threshold matrix, producing a regular, grid-like halftone pattern. Floyd-Steinberg is better for photographic images. Ordered dithering creates a retro, comic-book, or screen-print aesthetic and processes faster since it doesn't require error propagation between pixels.