Voice SDK for Speech & Telephony Audio¶

The AVBlocks Voice SDK is a cross-platform library for encoding, decoding and converting speech and voice audio. It is available for C++, .NET (C#, VB, F#) and Python, and runs on Windows, macOS and Linux.

It is designed for voice-recording, VoIP, telephony, call-logging, dictation and IVR applications that need compact, efficient speech codecs rather than full-fidelity music codecs.

Voice & speech codecs¶

Codec	Encode	Decode	Typical use
Opus	✓	✓	Modern VoIP, voice chat, WebRTC, low-latency speech
G.711 (A-Law / μ-Law)	✓	✓	Traditional telephony, PSTN, call recording
G.726 / ADPCM	✓	✓	Legacy telephony, voicemail, voice storage
PCM / LPCM / WAV	✓	✓	Raw / uncompressed voice capture and interchange

What you can build with the Voice SDK¶

Call recording / logging — capture and store telephony audio as G.711, G.726 or Opus.
VoIP and voice chat — encode low-latency speech with Opus.
IVR and voicemail systems — decode and play back G.711 / G.726 prompts and messages.
Dictation and transcription pipelines — record voice and convert to PCM for speech-to-text.
Format conversion — convert between Opus, G.711, G.726 and WAV.

Voice audio processing¶

Audio resampling (e.g. 8 kHz telephony to 16 kHz / 44.1 kHz)
Mono to stereo and stereo to mono
Change bitrate to control quality vs. size

Get started¶

Part of the Audio SDK. For music and general-purpose audio, see the Audio SDK; for video, see the Video SDK.