Asterisk PJSIP: SIP and Audio Codecs (Latest Versions)

SIP controls call signaling, but the actual voice quality in Asterisk depends on audio codecs.

Many real-world call problems — poor audio, one-way sound, robotic voice, or failed calls — are caused not by SIP itself, but by codec negotiation and media handling.

This guide explains how SIP and audio codecs work together in modern Asterisk (PJSIP), including practical troubleshooting knowledge used in production systems.

What Is an Audio Codec?

A codec is the method used to encode and decode voice audio for transmission over IP networks.

• Microphone captures raw audio
• Codec compresses it into digital packets
• Packets travel via RTP
• Receiver decodes and plays sound

Different codecs balance: quality, bandwidth usage, and CPU load.

How SIP Chooses a Codec

Codec negotiation happens inside the SDP section of SIP messages (INVITE and 200 OK).

• Caller sends a list of supported codecs
• Receiver selects one common codec
• RTP audio begins using the chosen codec

If there is no common codec, the call may fail or connect without audio.

Most Common Codecs in Asterisk Systems

G.711 (PCMU / PCMA)

• Standard telephony codec
• Very low compression → high quality
• Uses more bandwidth (~64 kbps per direction)
• Required by many SIP trunk providers

G.729

• Highly compressed (~8 kbps)
• Lower bandwidth usage
• May require commercial licensing
• Slightly reduced audio quality

Opus

• Modern high-quality codec
• Adaptive bitrate and excellent clarity
• Common in WebRTC/browser calling
• Not always supported by SIP trunks

GSM / iLBC / Speex

• Legacy or niche codecs
• Used in specific environments
• Rare in modern carrier trunks

Bandwidth Comparison (Approximate)

• G.711 → ~80–90 kbps with RTP overhead
• G.729 → ~24–32 kbps
• Opus → ~16–64 kbps (adaptive)

Call centers with many simultaneous calls must consider total bandwidth consumption.

Where Codecs Are Configured in PJSIP

Codecs are usually defined per endpoint or trunk:

allow=opus,ulaw,alaw
disallow=all

Best practice:

• Disallow everything first
• Explicitly allow only required codecs

Codec Mismatch Problems in Real Systems

Call Connects but No Audio

• No shared codec between endpoints
• Transcoding failure

Robotic or Distorted Voice

• Network packet loss or jitter
• Low-bitrate codec under poor conditions

High CPU Usage on Asterisk

• Heavy transcoding between codecs
• Multiple simultaneous conversions

What Is Transcoding (and Why It Matters)

Transcoding happens when:

• Caller uses one codec (e.g., Opus)
• Provider requires another (e.g., G.711)
• Asterisk converts audio between them

This increases:

• CPU usage
• Latency
• Risk of audio degradation

Best practice: avoid transcoding when possible.

Recommended Codec Strategy for Modern Deployments

Internal Phones / WebRTC

• Use Opus for best quality
• Allow G.711 as fallback

SIP Trunks / PSTN

• Primary → G.711 (ulaw/alaw)
• Avoid unnecessary codecs

Large Call Centers

• Minimize transcoding paths
• Keep codec plan consistent across system

Debugging Codec Negotiation in Asterisk

asterisk -rvvv
pjsip set logger on
core show channel

Look for:

• Selected codec in SDP
• Transcoding messages
• RTP packet flow

Common Beginner Mistakes

• Allowing too many codecs “just in case”
• Forgetting trunk codec limitations
• Ignoring CPU impact of transcoding
• Assuming SIP failure when problem is RTP/codec

Key Takeaway

SIP establishes the call, but audio codecs determine real voice quality and performance.

Stable Asterisk systems follow three rules:

• Use common codecs like G.711 for trunks
• Use modern codecs like Opus internally
• Avoid unnecessary transcoding

Mastering codec strategy is essential for clear audio, low CPU usage, and scalable telecom systems.