Normen Hansen

Heres a full log, 16MB.. [ndilog.txt](https://github.com/teltek/gst-plugin-ndi/files/7366011/ndilog.txt)

Correct, I didn't see or hear anything - these logs are from the exact `gst-launch` pipeline in the README using the official "NDI HX Camera" app on iPhone/Pad as a...

Funnily, in a different context ([obs-gstreamer plugin](https://github.com/fzwoch/obs-gstreamer)) the queue seems to work when using `timestamp-mode=2` - no additional buffers, no different settings, even when the appsink is synced... However I...

> That's a bad idea unless you know that the pipeline's latency is going to be lower than what you're configuring on it. This API should only be used if...

Heres a log with `max-size-time` at 10s, only `GST_DEBUG=ndi*:9` as the log gets too big with a full debug: [ndilog.txt](https://github.com/teltek/gst-plugin-ndi/files/7373909/ndilog.txt)

Heres a full log with `GST_DEBUG=6`, hope thats enough. [ndilog.txt](https://github.com/teltek/gst-plugin-ndi/files/7374010/ndilog.txt)

> If the `appsink` in the end is `sync=true`, check in the debug logs when the buffers are arriving, what latency is actually configured on the `appsink` and what it's...

> > Except for the NDI pipeline which even with 10 seconds set as latency doesn't change its "real world" latency in the pipe. > > As said, check the...

But wheres that time difference coming from? Its _by far_ not coming in 7 seconds late in reality. I can see the video, it's not 7 seconds late.

Okay, so based on the fact that I only get video from the "NDI HX Camera" when using `timestamp-mode=2` (timecode) and that the video timecode then is 7 seconds off...