SpeechSDK and MusicSDK lack audio file save capability

[NianJiuZst]

Summary

SpeechSDK.synthesize() and MusicSDK.generate() return hex-encoded audio in response.data.audio, but provide no method to decode and save that audio to disk. Users must manually implement hex validation, decoding, directory creation, and default filename generation.

The CLI already handles all of this via saveAudioOutput() in src/output/audio.ts.

Proposed solution

Add save(response, outPath?, ext?) methods to both SDKs that:

• Validate hex format and length
• Decode via Buffer.from(hex, 'hex')
• Create intermediate directories
• Generate timestamp-based default filenames
• Throw on missing audio data or disk-full errors

See linked PR.