I would really like to see depth of field also tested at various f-stops. I have found DoF to be a huge issue with these large sensor cameras. To compensate for the lack of day/night in the competing brands a low f-stop (at least f2.8 or not down to f1.4) is necessary but as a result you loose your DoF. Even better would be an IPVM DoF calculator:)
This is more of an issue with full frame sensors than APC-C sensors. I have had DoF issues when using the Avigilon 30MP cameras compared to the 8MP and 16MP. We have done shots where with the 30MP we could not get the whole image in focus but the issues went away with the 16MP and 8MP.
Sounds like a pretty significant load on the VMS. I'd be interested in understanding what the VMS impact is with multiple being viewed simultaneously on the client PC spec you have outlined above. How many can be viewed, etc.? What's the driving factor behind the high CPU uptick? Is that the load from server side motion processing or something else?
The CPU load measured is on the client side, not server. Exacq doesn't do server side motion, and it was turned off in Genetec.
On the server side, CPU usage didn't spike abnormally. Load on both VMSes was simply a function of throughput, as with lower resolution models. We did not test server side motion load, but that's a good thought. Usually that processing is done on a substream, regardless.
Was the Intel equipped client utilizing Intel's Quick sync on the Mobo? I know this has to be turned on in the BIOS when also utilizing with a discreet graphics card. I know Milestone's client is supposed to utilize it and I think Genetec does as well. This might bring down the CPU load.
We would like to see and know more about the user experience with this camera from an operational prospective when you have 50 plus cameras connected to a VMS and displayed on a video wall.
In an operational environment things go fast, zooming in and out, doing live and playback at the same time, exporting clips, panning left and right to track an individual, all while you have 50 plus cameras displayed on the video wall .
I also appreciate the bandwidth saving graphs you shared with us by using Zip stream, what type of activity was the camera recording for these tests? We would like to know if this can be applied in an environment where there is is constant activity: like stadium or airport, etc.. will be interesting to see how useful Zip stream will be there.
I didn't understand why to do a comparison between 4K, 20MP and 1080p cameras.. is almost like to compare bananas and watermelons, does not make sense. I think the right Axis camera to be compared with Dahua HF8835F is P1368, dont you think? Of course the 20MP sensor camera will perform worst in dark and will generate a biger net troughput.
I would like to see a Axis Q1659 vs Avigilon 24L-H4PRO-B super hight resolution cameras comparison. It would be interesting.