CallMeter Docs

Jitter Buffer Packets Pushed

Understand jitter buffer packets pushed — the count of packets successfully delivered from the buffer to the decoder for playback.

Jitter buffer packets pushed is a cumulative count of packets that the jitter buffer has successfully delivered to the audio or video decoder for playback. Each pushed packet represents a unit of media that made it through the entire receive pipeline — from the network, through the buffer, and into the decoder.

Think of it as the number of items that successfully made it through quality control on a production line. The buffer is the quality control checkpoint, and pushed packets are the ones that passed and moved to the next stage.

How It Works

The jitter buffer receives incoming RTP packets, reorders them if necessary, waits for the appropriate playback time, and then pushes them to the decoder in sequence. The packets pushed counter increments each time a packet is released to the decoder.

Not every received packet becomes a pushed packet. Several things reduce the count:

  • Lost packets — Never arrived, so nothing to push
  • Late packets — Arrived after their playback deadline, discarded
  • Duplicate packets — Second copy discarded, only the first counts

The relationship is: Packets Pushed = Packets Received - Losses - Late Packets - Duplicates

This counter is cumulative and increases throughout the stream.

Effective throughput

Packets pushed measures the effective media throughput — the actual data that reaches the decoder. Comparing this with total packets received gives you the buffer's delivery efficiency. A ratio close to 100% means the buffer is successfully handling network impairments. A lower ratio means packets are being lost or arriving too late.

Why It Matters

Packets pushed is the positive counterpart to loss metrics. While late packets and buffer losses tell you what went wrong, packets pushed tells you what went right. It measures the actual volume of media delivered for playback.

For SIP load testing, packets pushed helps you calculate buffer efficiency:

Buffer Efficiency = Packets Pushed / Packets Received x 100%

This single number summarizes how well the jitter buffer is performing its job. An efficiency of 99.5% means only 0.5% of received media was lost to buffer-level issues. An efficiency of 95% means 5% of the media never reached the decoder — a significant quality problem.

Tracking packets pushed over time also reveals throughput stability. A steady push rate indicates consistent media delivery. A rate that drops or becomes erratic signals buffer problems or upstream issues.

Common Scenarios

ScenarioPackets Pushed BehaviorWhat It Means
Clean network, no lossMatches packets received closelyIdeal — buffer is fully effective
Moderate loss, good RTXSlightly below received countBuffer losing some packets but RTX recovering most
High jitter, many lateNoticeably below received countBuffer deadline too tight for network conditions
Severe lossSignificantly below received countNetwork unable to deliver consistent media

How to Interpret It

  1. Calculate efficiency — Divide packets pushed by packets received. Target above 99% for good quality.
  2. Compare with losses and late packets — Packets pushed + losses + late packets + duplicates should roughly equal packets received. If the math does not add up, investigate.
  3. Watch the push rate over time — A declining push rate during a test indicates worsening conditions. A stable rate indicates consistent delivery.
  4. Compare across endpoints — Endpoints with lower push rates are experiencing worse buffer-level delivery, even if network-level metrics look similar.
  • Packet Losses — Packets the buffer declared lost, reducing the pushed count
  • Late Packets — Packets discarded as too late, reducing the pushed count
  • Duplicates — Duplicate packets discarded, not counted in pushed
  • Current Delay — The buffer delay that determines how long packets wait before being pushed

On this page