Avigilon WDR Shadowing Or Compression?

Hi.

I have often noticed a "shadow" (not shade, more like a ghost), or several shadows, following people across the scene. It is easiest to see when the background is in a uniform colour, and the person is walking from one side to the other. I have only noticed it on Avigilon cameras, but I have only really looked for it on those cameras, so it may be a normal problem. I thought it was wdr ghosting, from the extra exposures.

But the other day I tried to remove the shadows by turning off wdr, and they were still there. And I noticed that they sort of cleared on the I frames. So I am starting to suspect that it is the prediction algorithms in the codec that is the culprit, and not the wdr. I have not done any extensive testing, so I'm not really sure, but I am thinking that the shadows are there for all moving parts of the image, corrupting other stuff, like the face. So removing the shadows would maybe produce better images.

Have anybody noticed this for other cameras, and could anybody confirm or deny any of my hypotheses? Or come up with another one? If it is the compression, then I guess there is little to do about it myself, but if there is some kind of tinkering with the image settings that will remove them, then I would appreciate a tip:)

Login to read this IPVM discussion.
Why do I need to log in?
IPVM conducts unique testing and research funded by member's payments enabling us to offer the most independent, accurate and in-depth information.

***** *** **** ** a ******* **** ** image ******* *** *****?  I **** *** ******** of ******** ******* ** I ***** ** ********** to *** ** * can *** *** ***** your ******. 

*****. * ** *** allowed ** **** *** images **** **** ******. Next **** * *** hold ** ** ******** camera, * **** **** an ******* *****, *** post ** ****. ***** be * **** **** though.

* ******* * **** seen ** ** *** 1 *** * ** dome *******. **** ** and **. *** **** behind *** ****** ****** is ******* **** ****, when * *** ** most *******.

* **** **** **** on ******** *************, *** just ********.  ** ***** I ******* ** *** H.264 ** *.***, *** I ****** *.*** *** it ***** ******* ** H.264.

* **** *** ****/******* theory **** ***** "***** codecs" *** **** * part. ** ** ***** also ** ******* **** have ***** *** ***** sensitivity *** *** ****** is ****** *** **** to **** ** *****.  To ** *********...

*** **** **** * saw **, ** *** inside, *** ********, **** the ******* ****** *** camera. ****** ** ****** light. * *** ** on * ** ** well, *** **** *** a *** ***** *****. Did ** **** *** "smart *****"? ***** ** disappears **** *****? ** you **** * ****** to **** ****? * don't **** * ****** available.

*** **** ** **** I-frame ********? ** ****** like *********** ********* ** me. **** ****** ** a ***** ** ***** and ***** ****** ****** the ******. *** *******, behind **** ***** *** from ****.*** * ** * Frame **********, *** *** *** a *****, **** *** line ** *** **** of *** *** ********.

** **** ***'** ****** is *******, *'* ********* adjusting **** *-***** ******** to ***** **** *********. Avigilon ***** ** ******** Interval, *** **'* ** the *********** *** ***** Rate ****. ** ** you're ***** ** *** your ******** ******** ****** be **.

* ******* *** *** gop ** ***** * second. * *** *** verify *** **** **** is *** **** *** the **** ******, *** I **** ***** **** I *** *** ***********.

* ** *** **** if ** ** *********, like *** **** ** your ******* *******. **** I ***, ***** ** me **** ************ ****** of *** ****** ******, with * **** *** opacity.

*****'* ************** ** ********* we *********** ** **** older ** ***** ***** we **** ****** *** "ghosting" *****.  ******** *** keyframe **** ** *** image **** **** * big **********.  ** *** not ******* ********* *** problem, *** ** **** it * *** **** noticeable.  ********'* ******* ******** was *** ** ** turn *** *** . . . *** ***** that ******* *** *** we ********* ***** ** domes, **** *** *** a ****-**** ******** *** us.  ********* *** *** make * ********** ** the ******** *******, *** the *-***** ********** *** the ****** ***. 

** **** ***** **** a ********** ******** ****** appears ** **** ******* the ***** **** ****.  Seems **** ** *** have **** **** *** upgrade **** **** ** ACC6, *** * ****** recall *** ****.

*** ********** *** **** that *** ***** *** present ** * *** H3 *** *** ******, but ** **** *** seen ** **** *** of *** ** *******.  We *** ********* ** 6.14.2.12.

*** ****** **** ** use * ******** *** second **** ********.  ******** when *** ***** ***** it **** *** *********. So * ******** *** second ** ****** ***********. 

... *** *** *********, our ******* ****** ******** are ******** ******** *** most *****, *** **** 1/60s *** ********, ** fps, ******** ********=** (** once *** ******), *******=* (Avigilon *******).