Base-64 Encoding Makes Binary Data Safe
The base-64 encoding converts a series of arbitrary bytes into a longer sequence of common text characters that are all legal header field values. Base-64 encoding lets us take user input or binary data, pack it into a safe format, and ship it as HTTP header field values without fear of them containing colons, newlines, or binary values that would break HTTP parsers.
Base-64 encoding was developed as part of the MIME multimedia electronic mail standard, so MIME could transport rich text and arbitrary binary data between different legacy email gateways. Base-64 encoding is similar in spirit, but more efficient in space, to the uuencode and BinHex standards for textifying binary data. Section 6.8 of MIME RFC 2045 details the base-64 algorithm.
Some mail gateways would silently strip many "non-printing" characters with ASCII values between 0 and 31. Other programs would interpret some bytes as flow control characters or other special control characters, or convert carriage returns to line feeds and the like. Some programs would experience fatal errors upon receiving international characters with a value above 127 because the software was not "8-bit clean."