How stream_id hash is generated from input name?

Sebastian Dröge sebastian at
Sun Aug 21 16:12:06 UTC 2022

On Sat, 2022-08-20 at 15:25 +0200, Tomy Elrond via gstreamer-devel
> These "hashes" must be related with input names, because for each
> input name it generates the same hash every time. For example:
> "
> input name: "file:///tmp/input_video_1.mp4" -> generated hash:
> "0b70cf574e1fef0b694a8a449a94732091cda1ce67bc4bf7c71c59d6c331f5b1/001
> "
> input name: "file:///tmp/input_video_2.mp4" -> generated hash:
> "7960608e8835f3eeaf8fba46351367dd76199a9e133117d4aeeccb862e8da9aa/001
> "
> "
> Is there any way how to "decode" these hashes to the original string
> (path of video file?

It's the SHA256 of the URI, currently. You can't decode that and you
must not rely on that staying the same forever.

The stream IDs are opaque IDs that only allow you to distinguish
different streams / find same streams.

> Or is there some other way to find out stream ids (for example first
> input -> 0, second input -> 1)?

Via the STREAM_START event.

Sebastian Dröge, Centricular Ltd ·

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the gstreamer-devel mailing list