About

Every raster image is a grid of pixels. Each pixel stores color as numeric channel values in the range 0 - 255. This tool reads those values and converts each one to its 8-bit binary equivalent. A red channel value of 200 becomes 11001000. A threshold mode reduces each pixel to a single bit: 1 or 0, using the ITU-R BT.709 luminance model Y = 0.2126R + 0.7152G + 0.0722B. Misinterpreting channel order (RGB vs BGR) or ignoring gamma correction leads to corrupted binary representations. This converter handles standard sRGB images and lets you downsample before conversion to keep output manageable.

The tool processes images entirely in your browser using the Canvas API. No data leaves your device. Output formats include full RGBA binary (32 bits per pixel), grayscale binary (8 bits per pixel), and 1-bit threshold. For an image of w × h pixels in full RGBA mode, the output contains w × h × 32 binary digits. Downsample aggressively for large images. A 100×100 image at full RGBA already produces 320,000 bits of output.

Formulas

Each pixel channel value c in the range [0, 255] is converted to an 8-bit binary string:

b = bin(c), padded to 8 digits

For grayscale mode, the luminance Y is computed using the ITU-R BT.709 standard coefficients before binary conversion:

Y = 0.2126 ⋅ R + 0.7152 ⋅ G + 0.0722 ⋅ B

For 1-bit threshold mode, each pixel is reduced to a single binary digit:

{

1 if Y ≥ T0 if Y < T

Where T is the user-defined threshold (default 128). Full RGBA mode concatenates four 8-bit values per pixel: bin(R) + bin(G) + bin(B) + bin(A), yielding 32 bits per pixel. The total output length for an image of dimensions w × h is:

L = w × h × bpp

Where bpp = bits per pixel (32 for RGBA, 24 for RGB, 8 for grayscale, 1 for threshold).

Reference Data

Output Mode	Bits per Pixel	Channels Used	Typical Use Case
Full RGBA	32	R, G, B, A	Complete pixel data preservation
RGB Only	24	R, G, B	Color data without transparency
Grayscale	8	Luminance (Y)	Monochrome binary representation
1-Bit Threshold	1	Luminance vs threshold	Binary art, ASCII art source
Common Image Dimensions & Output Sizes (1-Bit Mode)
8×8	64 bits		Sprite/icon tiles
16×16	256 bits		Favicon, retro sprites
32×32	1,024 bits		Small icons
64×64	4,096 bits		Thumbnail analysis
100×100	10,000 bits		Small bitmap study
256×256	65,536 bits		Texture analysis
512×512	262,144 bits		Medium bitmap
1024×1024	1,048,576 bits		High-res analysis (slow output)
Bit Depth Reference
1-bit	2 values		Black & white only
4-bit	16 values		Early CGA/EGA graphics
8-bit	256 values per channel		Standard sRGB (this tool)
16-bit	65,536 values per channel		HDR / medical imaging
24-bit	16,777,216 colors total		True Color (RGB)
32-bit	4,294,967,296 values		True Color + Alpha (RGBA)

Frequently Asked Questions

A simple average treats all channels equally: (R + G + B) ÷ 3. The ITU-R BT.709 formula weights green at 0.7152 because human vision is most sensitive to green wavelengths. A pure green pixel (0, 255, 0) yields luminance 182 with BT.709 but 85 with simple averaging. This difference changes threshold results significantly.

Threshold mode ignores the alpha channel entirely. It computes luminance from RGB values only. Fully transparent pixels (alpha = 0) still have RGB data stored in the bitmap, and those values are used. If you need transparency-aware conversion, use RGBA mode and inspect the alpha binary separately.

A 512×512 image in RGBA mode produces 8,388,608 binary digits. Rendering that much text in the DOM causes browser memory issues and potential tab crashes. The downsample option lets you reduce any image to a manageable size while preserving its visual structure through nearest-neighbor or bilinear interpolation.

This tool reads raw sRGB channel values as stored in the image file (gamma-encoded). It does not apply inverse gamma correction before binary conversion. The getImageData() Canvas API returns values in sRGB space. If you need linear-light binary values, you would need to apply the sRGB transfer function externally: c_linear = ((c_sRGB + 0.055) ÷ 1.055)^2.4 for values > 0.04045.

Yes, if you use RGB or RGBA mode without downsampling. Each pixel's channels map to sequential 8-bit groups. Parse the binary string row by row, split into 8-bit chunks, convert each to decimal (0-255), and reconstruct the pixel array. The original image dimensions are included in the output header. Threshold mode is lossy and cannot reconstruct the original.

This tool outputs standard big-endian binary (most significant bit first). A channel value of 200 outputs as 11001000, where the leftmost bit represents 2⁷ = 128. Little-endian would reverse this to 00010011. Most binary data standards and network protocols use big-endian (network byte order).