How Base64 Decoding Works: Behind the 6-Bit Math

Decoding Base64 is the exact inverse of the encoding transformation. While encoding splits three 8-bit bytes (24 bits total) into four 6-bit indices, decoding takes groups of four characters, maps them back to their numeric values (0–63), and packs them into three standard 8-bit bytes.

Step-by-Step Bitwise Reassembly

1. Lookup: Each character in the quartet is matched against the 64-character table to obtain its 6-bit integer value (v1, v2, v3, v4).

2. Bit Shifting: The four 6-bit numbers are concatenated into a single 24-bit integer using bitwise operations: (v1 << 18) | (v2 << 12) | (v3 << 6) | v4.

3. Extraction: The three original bytes are extracted using bitmasks: Byte 1 is (bits >> 16) & 0xFF, Byte 2 is (bits >> 8) & 0xFF, and Byte 3 is bits & 0xFF.

Handling Padding Characters

When the input string ends with "=" or "==", the decoder recognizes that the original data stream ended on a 1-byte or 2-byte boundary. A single "=" indicates that the last quartet contains only two real bytes; "==" indicates that only one real byte was encoded.

Modern web browsers implement this natively in JavaScript using the `atob()` function or the more robust `TextDecoder` API for international UTF-8 strings.