packet capture tools
RTP Health Analyzer
TraceRoo's RTP Health Analyzer inspects RTP packets inside a local PCAP and compares each stream side by side so directional voice quality trouble is easier to spot.
What It Checks
- Source and destination for each RTP stream.
- Codec or RTP payload type, SSRC, packet count, and expected packet count.
- RTP sequence gaps that suggest packet loss.
- High jitter and arrival-time spikes that can cause choppy audio.
- Duplicate RTP packets.
- Out-of-order packets.
When To Use It
Use this tool after capturing a call on an SBC, endpoint, firewall span port, or packet broker. It is helpful for one-way audio, robotic audio, clipping, intermittent silence, and calls where SIP signaling looks clean but media quality is poor.
Example Workflow
- Capture the call traffic as PCAP or PCAPNG.
- Open the RTP Health Analyzer and upload the capture.
- Choose the RTP clock if needed.
- Compare streams side by side by source, destination, codec, SSRC, packet count, loss, jitter, and gaps.
- Review the strongest highlighted streams first, then compare directionality.
How To Read Results
- Packet loss means RTP sequence numbers skipped. Sustained loss usually matches clipped or missing audio.
- Jitter means packet arrival timing varied. High jitter can sound robotic even when packet loss is low.
- Out-of-order packets can be corrected by jitter buffers, but frequent reordering can still degrade quality.
- Duplicate packets may be capture artifacts or network duplication. Compare capture points if possible.
- If only one direction is unhealthy, check NAT, firewall policy, routing, and media anchoring for that direction.
How To Use RTP Health Analyzer
- Upload a PCAP or PCAPNG captured close to the reported call problem.
- Review each RTP stream by SSRC, source, destination, packet count, loss, jitter, duplicates, and reordering.
- Start with the strongest highlighted streams, then compare directionality to see whether one side of the call is worse.
- Open SIP Call Flow Analyzer when you need to compare media symptoms with the negotiated SDP.
Troubleshooting Flow
Example Result
Packet loss: 2.8%
Jitter: 38 ms
Out-of-order packets: 14
Health: needs review
- Sustained loss can sound like clipped or missing audio.
- High jitter can sound robotic even when loss is low.
- If only one direction is unhealthy, check NAT, firewall rules, routing, and media anchoring.
Caller -> callee: 0.1% loss, 4 ms jitter
Callee -> caller: 6.4% loss, 52 ms jitter
Health: one direction needs review
- Directional problems often point to NAT, firewall policy, asymmetric routing, or a capture point seeing only one side.
- Compare source and destination IPs with SDP media addresses.
- Use the SIP Call Flow Analyzer if codec or media address negotiation looks suspicious.
What Good And Bad Results Look Like
- Both directions show low loss and stable jitter.
- Packet counts make sense for the call duration.
- Sequence behavior matches what users reported hearing.
- One direction has sustained loss or high jitter.
- Large sequence gaps line up with clipped audio.
- RTP appears only one way after answer.
Common Mistakes
- Judging voice quality from SIP signaling alone.
- Reading both call directions as one combined stream.
- Ignoring capture location when only one side of media is visible.
Practical Troubleshooting Workflow
- Confirm the capture includes the problem call window.
- Review RTP streams by direction and SSRC.
- Start with streams marked for review, then compare directionality.
- Compare media results with SIP SDP when the direction looks wrong.
FAQ
What is high jitter for RTP?
There is no single universal number, but sustained jitter above common voice buffer ranges is a warning sign. TraceRoo highlights streams where timing variation is likely to affect audio quality.
Can RTP packet loss cause one-way audio?
Packet loss usually causes choppy or missing audio, but complete one-way audio is often caused by routing, NAT, firewall, codec, or media negotiation issues.
Can encrypted SRTP be analyzed directly?
Timing and packet sequence behavior can still be inspected, but audio payload extraction requires the right SRTP key material and authorization to decrypt the media.