There have been several posts about using RTSP for capturing video vs a capture card. I am hoping this post can open a more detailed dialog about how to handle latency. One significant question - is it effective to run RTSP with a mixture of devices.

I ran across a device (I purchased on Amazon - see below) that seemed reasonably priced compared to other devices similar to this. At $159 the FBE200 will take an HDMI feed and turn it into h.264 accessible via RTSP. I also have a PTZOptics 20X PTZ camera that supports h.264. The PTZ optics has about a 0.5 sec delay and the FBE200 has about a 1.2 second delay.

If you had to use these in the same stream (and/or recording) it would be very difficult to do voice synch. However, if I didn't have to worry about this synch, i.e. they were different camera angles or an audience shot, then perhaps it may be viable. The benifit then would be you could add any HDMI device into a mix as long as audio synch was not an issue.

Is there a way to sync video inputs via a "delay" that I am missing? Any thoughts?

FMUSER H.264 1080p HD HDMI Video Encoder, IPTV Encoder,Live Stream Broadcast,RTMP LAN Encoder
https://www.amazon.com/gp/product/B01N4PJDX2/ref=as_li_tl?ie=UTF8&camp=1789&creative=9325&creativeASIN=B01N4PJDX2&linkCode=as2&tag=n05284-20&linkId=b4191f0362a3950ea0e0289f90995cbb

bottom = actual time
left = PTZ optics
right = FMUSER

Knowing the amount of latency involved with each device can be useful. However, I would not bother trying to sync such sources. They simply won't be stable. They're not locked to common reference, so they will drift. It's just a question of how much.

I completely agree with DWAM that RTMP/RTSP are not intended for production purposes.

That said, there are circumstances where latency is not an issue and they can be used in inventive ways to enhance a production. For example, add a very long exterior shot of a venue or stage.