I'm trying to decompress AIFF audio data from an XFDF sound annotation, and play the audio in the browser. Any time I export a sound annotation from Adobe Acrobat and decompress the audio data it results in seemingly garbage data.
The code below works for compressed audio starting with 789C header, but a few compressed streams start with 4889 or 78DA and fail to play in the browser. After numerous tests with Acrobat, exported sound annotations in XFDF start with 4889.
// Two audio examples, one working and one failing.
// Can't paste it all due to char limit.
const compressedHexString_working = `789C3C9C059C13D7FBF5473293F55DBCB8BBBBBB53DC8ABB142FEE5AB4 ...`;
const compressedHexString_failing = `48892C97077C8FE716C7CFF9532EB7418D5E24768492DABB0862146DAF ...`;
// Convert hex string into ArrayBuffer.
const compressedByteArray = new Uint8Array(
compressedHexString_working
.match(/[a-fA-F0-9]{2}/g)
.map((str: string) => parseInt(str, 16))
);
const { default: zlib } = await import('pako');
const audioData = await zlib.inflate(compressedByteArray);
// Parse header bytes to determine audio file.
const fileHeader = new TextDecoder().decode(audioData).substring(0, 4);
console.log('Sound File Header:', fileHeader);
The console.log usually prints ID3, RIFF, or FORM when decompression is successful. But when decompressing the failing example, the result is a garbage value: �D�C.
Am I doing the decompression incorrectly? Or am I missing a step?
