Garbled text after Base64 decoding: UTF-8 troubleshooting

Follow café through the byte representation

The UTF-8 bytes for café are 63 61 66 c3 a9 in hexadecimal. Their Base64 representation is Y2Fmw6k=. The final character occupies two bytes; interpreting each byte as a character produces the visible mojibake café.

Use the following values as a controlled test. They let you distinguish a damaged Base64 string from a correct byte sequence decoded with the wrong character encoding. A screenshot of garbled text alone cannot establish which stage failed.

Decode bytes before interpreting text

The browser atob function produces a binary string. Its character codes represent byte values; the result is not automatically a correctly decoded Unicode string. Feed those values into TextDecoder when the producer used UTF-8.

This snippet handles the known standard Base64 example. Production code should validate the expected alphabet and data contract. It does not provide a general file decoder or a way to infer the encoding of arbitrary bytes.

const binary = atob("Y2Fmw6k=");
const bytes = Uint8Array.from(binary, ch => ch.charCodeAt(0));
const text = new TextDecoder("utf-8", { fatal: true }).decode(bytes);
console.log(text); // café

// Encoding the same text:
const input = new TextEncoder().encode("café");
btoa(Array.from(input, b => String.fromCharCode(b)).join(""));
// Y2Fmw6k=

Preserve the original bytes while investigating

If PageUtil reports that decoded bytes are not UTF-8, the producer may have used another encoding, supplied binary content, or truncated the data. Do not force replacement characters and assume the result is fixed. Compare a known-good sample with the producing application’s charset settings.

If text was already corrupted before it was Base64-encoded, decoding correctly will preserve that corruption. Find the earliest step where the expected UTF-8 bytes differ. Repeated “repair” conversions can make the original evidence unrecoverable.

For large files or non-text payloads, use a local byte-oriented tool appropriate to the format. PageUtil’s text conversion stays on your device, but it intentionally rejects invalid UTF-8 rather than guessing another encoding.

References

Use the linked specifications and product documentation to check behavior in your own environment.