T54PRO-ZE Field-of-View (and change in aperture/exposure) as focal length changes

[alekk]

A while back, I read this post from @Perimeter ... so I setup a T54PRO-ZE 10' away from my garage along with some various "measurement props" and then took a dozen snapshots as I zoomed from 0 to the maximum of 2426 - recording the zoom/focus numbers.

I then pulled those images into Photoshop and using my "every two feet tape marks", measured the horizontal field-of-view in feet ... and then using some simple trigonometry, calculated the horizontal FOV in degrees.

It was a cloudless early afternoon day, so while this was in the shade, the exposure was not changing. The settings were fixed 1/1000s shutter, 100 iris, 0 gain, and no NR. So ideally(!) the exposure was fixed ... with any difference being because as I zoomed in, the aperture closes up. And hopefully(!) the camera's processing wasn't "adjusting" the exposure ... or at least too much! So again using Photoshop, I did the (semi) same rectangle cut-out from the middle of the sheet ... which wasn't a "perfect" white, but not shabby.

I generated histogram values from ZOOM=0 ... and then as the camera zooms in (and the exposure dropped as the aperture closes up), I used that "same" rectangle (obviously more pixels) to generate a (darker) histogram ... and then I use the Photoshop Exposure took to push those values to as close as possible to the ZOOM=0 values.

So while the FOV numbers are probably fairly accurate (I got 49-113 degrees at 10' focus distance versus specs of 48-114 degrees at unknown focus distance), I'm not sure about the aperture vs. focal length results ... but there is no doubt from my real-world use that it's brightest wide-open (specs says f/1.2) and I saw (basically) a full stop loss of light (which would be about F/1.8) ... but again, there may be some processing trickery that is preventing me from truly measuring that ... as it seems a bit optimistic over that zoom range ... but could be.

BTW, Dahua clearly has better methods to measure this data than me - they should publish it! ;-)

A few images in-line ... with the complete set attached ... along with the spreadsheet of results.

 

[alekk]

Reserved for possible future updates ...

 

[CamCrazy]

This most useful, appreciate the time you put in, not a quick task! Looks to my eyes that light falls away quicker past 0382 and 0582. Would very much like to know at what point (potentially) the f stop changes from f1.2 to f1.x

Indeed this is likely data that Dahua can publish, for now I will pull my ZE back to zoom below 0400 and see how it looks. Did some brief testing with this and got better results at night a week or so back, it makes sense.

 

[alekk]

It's actually a slower fall-off that I thought it would be going into it ... so again, I'm not sure if there is some processing trickery that is interfering with my measurements.

And yes, you can definitely get a brighter image at night "going wide" ... and in fact, since I'm OK with a 1920x1080HD image, I used to zoom in to the FOV I wanted and have the camera output 1920x1080. But now I zoom OUT, grab the entire 2688x1520 image, and then crop that to 1920x1080.

Which means I'm giving up a little bit of resolution, but getting a brighter image for the same shutter speed and gain ... since the aperture is wider.

 

[CamCrazy]

It's actually a slower fall-off that I thought it would be going into it ... so again, I'm not sure if there is some processing trickery that is interfering with my measurements.

And yes, you can definitely get a brighter image at night "going wide" ... and in fact, since I'm OK with a 1920x1080HD image, I used to zoom in to the FOV I wanted and have the camera output 1920x1080. But now I zoom OUT, grab the entire 2688x1520 image, and then crop that to 1920x1080.

Which means I'm giving up a little bit of resolution, but getting a brighter image for the same shutter speed and gain ... since the aperture is wider.

Agreed, I was almost hoping it would fall off sooner at about 30% zoom, which does make you wonder about processing more.

I think that I've already crossed over into the 'annoyed they released this too soon' camp, with what appears to be a potentially (my opinion only!) sub par sensor. I have held my tongue pretty well so far, but if you stick 'Pro' on the end of a camera you best make damn sure it can live up to the name!

Having dual illumination is nice, speaker onboard along with mic also nice. But I would give up any of those for better image quality on this and any other camera. Fingers crossed with firmware they can get this nearer to the target. Side note, is anyone seeing issues with white LED's staying on all day when set to auto, I assume that is not expected behaviour, I am on firmware 1028 now. If that is expected then excuse my ignorance.

 

[alekk]

Yea, I think there was a regression with LED light behavior with 1028FW ... and Andy has a 1031FW (which is 10/15 from Dahua) which I think addresses this ... but pretty sure it doesn't include the image improvements ... which I like.

LOL a bit confusing ... hopefully there is a roll-up release coming soon.

 

[CamCrazy]

Yea, I think there was a regression with LED light behavior with 1028FW ... and Andy has a 1031FW (which is 10/15 from Dahua) which I think addresses this ... but pretty sure it doesn't include the image improvements ... which I like.

LOL a bit confusing ... hopefully there is a roll-up release coming soon.

Ah OK, that makes sense, I realised the latest release was features only. Yes, we need a roll up with some serious reductions in 'auto bot' adjustments!!