Image Steganography

Image steganography: the LSB method in detail

The technical reference for the classic image-steganography technique this tool uses — what LSB changes, and the real trade-offs.

Method

LSB (least-significant-bit) encoding — the lowest bit of each pixel's red, green and blue value is replaced with one bit of your message. The alpha (transparency) channel is left untouched.

Data rate

3 bits per pixel (one per RGB channel), so roughly 1 byte for every 3 pixels, minus a 9-byte header.

Visible change

At most ±1 on a 0–255 colour scale per channel — far below what the eye can perceive, so the picture looks identical.

Per-pixel detail

Each pixel stores 3 message bits (R, G, B). Flipping a channel's lowest bit changes that colour by exactly 1 step out of 256 — e.g. a blue value of 144 becomes 145.

Output format

Always a lossless PNG. JPG and WebP re-compress the pixels, which would wipe out the hidden bits, so the tool never exports those.

How much fits

The capacity formula is floor(width × height × 3 ÷ 8) − 9 bytes. Worked through: an 800 × 600 image gives 179,991 bytes — about 180,000 ASCII characters. A 2000 × 1500 image gives about 1,099 KB.

Security: it hides, it doesn't lock

An honest note on LSB's limits. Steganography hides that a message exists; it does not, on its own, scramble what the message says. Anyone who knows to run this kind of image through an LSB extractor can read an un-passworded message. For real secrecy, set a password so the text is encrypted as well as hidden — and remember that LSB is not designed to survive forensic statistical analysis. Basic LSB also leaves statistical traces that dedicated steganalysis can flag, so treat it as "hidden from casual view," not "undetectable."

Tips that actually matter

• Choose a richly detailed colour photo as the cover. Smooth gradients and large flat areas are where LSB changes are easiest to spot under analysis.
• Keep and share the exact PNG the tool produces. Screenshotting it, re-saving it as JPG, or letting a chat app auto-compress it re-encodes the pixels and erases the hidden message.
• There's no benefit to enlarging an image just to gain capacity — upscaling adds pixels but not real detail, and the same bits could be flagged more easily.