Video Decoder Errors

What It Measures

Video Decoder Errors counts the number of frames the video decoder failed to decode during a call. Each error represents a frame where the decoder encountered corrupted data, missing reference frames, or a codec-level inconsistency that prevented successful frame reconstruction.

A decoder error does not necessarily cause an immediate visible freeze — the decoder may conceal the error by repeating the previous frame or interpolating from adjacent frames. But each error increases the probability of a visible artifact, and errors above a certain threshold inevitably produce Video Freeze Events as the decoder's concealment capacity is exceeded.

Why It Matters

Decoder errors reveal corruption that Packet Loss Rate alone cannot capture. Video is particularly vulnerable because of inter-frame dependencies:

The reference frame cascade: Modern video codecs (H.264, VP8, H.265) use predictive coding where most frames (P-frames and B-frames) are encoded as differences relative to prior keyframes or reference frames. When a reference frame is corrupted or lost, every frame that depends on it is also undecodeble. A single lost packet in a keyframe can corrupt dozens of subsequent frames — producing many decoder errors from a single network event.

This means video decoder errors can dramatically exceed packet loss count. A 0.1% packet loss rate that happens to hit a keyframe can generate 30 decoder errors and 3 visible freeze events. The packet loss metric looks fine; the video experience is not.

For video infrastructure testing, decoder errors are the bridge between network-level loss statistics and viewer-visible quality:

• Zero errors: Network is delivering complete, decodable video streams. Codec, packetization, and network path are all working correctly.
• Low errors (1–10): Occasional network events are hitting the video stream. Monitor for increase under load. Check whether errors correlate with freeze events.
• High errors: Systematic issue. Could be packet loss hitting reference frames, RTP reordering causing out-of-sequence frame assembly, or a codec negotiation mismatch.

Decoder errors are also a quality assurance tool for codec configuration: if errors are high but packet loss is low, the encoder may be producing malformed bitstreams — a codec bug or misconfigured encoder parameter.

Thresholds

How to Fix It

1. Correlate with packet loss timing. Use the per-second timeline to check whether decoder errors spike when Packet Loss Rate increases. If they track together, the packet loss is hitting reference frames — reduce loss first.
2. Reduce the keyframe interval. A shorter Keyframe Interval limits the blast radius of reference frame corruption. After a decoder error, the stream recovers at the next keyframe — a shorter interval means faster recovery.
3. Enable NACK retransmission. If NACK Requests are being sent but not answered, the sender is not supporting retransmission. Decoder errors from recovered NACK packets should be lower than errors from dropped packets.
4. Check packetization mode. For H.264, verify that the packetization mode matches between encoder and decoder. A mode mismatch causes systematic decoder errors on every frame, not just on loss events.
5. Review codec configuration. If decoder errors are high but packet loss is zero, the encoder is producing a stream the decoder cannot handle. Check profile/level settings, reference frame count, and encoder parameter negotiation in the SDP.

Related Metrics

• Video Freeze Events — Freeze events triggered by decoder errors exceeding concealment capacity
• Discarded Frames — Frames discarded before decode; precedes decoder errors in the pipeline
• Keyframe Interval — Determines recovery time after decoder errors
• Packet Loss Rate — Primary network cause of decoder errors via reference frame loss