No video or interrupted video
The endpoint can answer and allow access while its video still fails. Glassknock separates received data from decoded frames.
The device sends video data. Glassknock receives the data. FFmpeg turns H.264 data into readable video frames. Receiving data alone does not prove that a frame can be read.
Follow the observed failure
| Observation | What it means | Next useful action |
|---|---|---|
| FFmpeg is unavailable. | The receiving computer cannot complete the decode check. | Install or fix the local dependency, then repeat. |
The local decoder stops before producing a frame (decoder_failed). | This sample did not establish readable video. The cause is unconfirmed. | Check that FFmpeg 5.1 or newer is installed with the required decoder and options, then repeat. |
| The feed uses an unsupported codec. | This release cannot decode it with its supported method. | Confirm whether an existing H.264 feed is available. |
| Access works; no media arrives. | The request worked, but this sample received no useful video data. | Confirm the selected path and the endpoint’s streaming state. |
| Media arrives; no frame decodes. | Receiving bytes did not establish usable video. | Review the supported codec and decode evidence. |
| Media or frames have long gaps. | Delivery did not meet the configured limit during this sample. | Inspect the path and server load with the responsible person. |
Do not call an unsupported codec a defective camera. Do not call a still scene a frozen feed. Do not claim a gap proves inadequate bandwidth; several causes can interrupt delivery.
Check the exact feed
Confirm which existing camera or recorder feed the monitoring system needs. A low-bandwidth secondary feed can work while the primary feed fails. Testing one does not validate the other.
This release uses RTSP or RTSPS over TCP and H.264. It does not test UDP delivery or H.265. Ask the installer whether there is an already configured compatible endpoint. A setting change must follow a supported equipment procedure and its own authorization; the diagnostic tool does not make that change.
Check test length and conditions
A short test can miss an occasional problem. Record sample duration and actual observation time. If the sample was interrupted or cancelled, repeat it before accepting reception.
Run from the intended receiving computer. Testing near the camera can produce a different result from testing across a VPN. Testing one feed at a time does not establish capacity for all site feeds together.
Correct and repeat
Save a report before a confirmed correction. Repeat the same test settings afterward and compare the results. A decoded frame proves that this sample produced readable video. It does not prove day/night image quality, intrusion detection, recording, or alert delivery.