Dahua cams...VBR vs CBR: Would you expect a higher amount of motion blur?

[Sybertiger]

Honestly, I don't feel like I've made great strides in setting up the cam settings. Rather, I use the factory settings for the most part and try to tweak them. I'm starting to digest the excellent thread on common issues which says to dump the factory settings. It's been helpful but thought I'd get more specific info from you all. Most likely this isn't a VBR vs CBR issue but all the exposure settings.

Part of my problem in getting the settings adjusted right is I live in such a quiet neighborhood there isn't enough action for me to test and re-test the settings. Probably most of you have the same issue, it's just you so you have to make setting changes then go outside then come inside then review then do it again and again and again. Maybe that's the fun in it? Anyhow, I was playing around with VBR on my Dahua IPC-HDBW4231F-E2-M (aka Boobie Cam) and you can see the result below. Looks a bit blurry in the facial shot. It was 6:02pm and the sun sets at 6:22pm but the sun was very low in the sky and below the tree line so it's actually a lot darker than what the camera is showing thanks to the Starlight CMOS sensor that makes it look brighter than it really is. I was testing just the boobie cam with VBR because the camera's view(s) is primarily static all the time, i.e. no constant foot or vehicle traffic.

The Dahua setting Exposure settings are for only "Day" or "Night" but really I feel like I need Day, Night AND Twilight settings and while this wasn't a part of twilight it was getting dark. Curious for those who have similar Dahua 2MP Starlight cams what are you using for your frame-rate plus exposure settings in the Day and at Night? And, if night is forced color what exposure settings are you using? Anyone care to share their settings.



The shutter speed is too low in this example...it's mostly to show what settings you can make. Curious to know the mode, shutter speed, gain, 3D NR grade, etc. I can't remember what it was set to at the time of the pics but it was in VBR.


This is a Dahua IPC-T5442TM-AS 4MP in CBR...seems decent.


This is a Dahua IPC-HDBW4231F-E2-M bobbie cam with VBR on...fuzzy face


Same boobie cam but closer to the cam...who says "no one ever looks up"?....fuzzy face


This is a Dahua IPC-HDBW4231F-AS wedge cam in CBR...great shot


This is the second cam that's a part of the same boobie cam in VBR. Fuzzy head but maybe he moved very quickly causing blur.

 

[bigredfish]

I bribed the wife to run around the house a lot when i was testing. Cost me a bloody fortune.. Sometimes the cats will cooperate.

I use Daytime on 2MP/4MP Starlights
30FPS
Iframe matches FPS
Exposure 8192 CBR
1/60
DNR usually 30-40

Nighttime same except 1/120 on my forced color cams, I stay with 1/60 mostly on my B&W/IR cams

VBR in theory produces better quality because it allows the camera to go to a much higher bitrate and lower when no activity, but my experience is the transition as it "ramps up" can be problematic when i just get a few frames of a fast moving car for LPR or just a quick glimpse, say 2-3 steps of someone walking quickly.

 

[Sybertiger]

I'd have the girlfriend do the running around but it's either like herding cats or trying to direct a bengal tiger what to do...you just "don't go there". Thanks for the tips, I'll try them out.

 

[wittaj]

I bribed the wife to run around the house a lot when i was testing. Cost me a bloody fortune.. Sometimes the cats will cooperate.

I use Daytime on 2MP/4MP Starlights
30FPS
Iframe matches FPS
Exposure 8192 CBR
1/60
DNR usually 30-40

Nighttime same except 1/120 on my forced color cams, I stay with 1/60 mostly on my B&W/IR cams

VBR in theory produces better quality because it allows the camera to go to a much higher bitrate and lower when no activity, but my experience is the transition as it "ramps up" can be problematic when i just get a few frames of a fast moving car for LPR or just a quick glimps, say 2-3 steps of someone walking quickly.

Do you get decent pics at 1/60 in daytime or is a number missing?

 

[Tom S]

Part of what strikes me is in the grand scheme of things is that overall all you have good overall capture of what occurred. You have good overall views of the subject and some good face shots as well. I too am experiencing times when I am impressed with the cameras and the times when I am not. These cams obviously do not present 30 single images of the same quality as 30 single snap shots. The video stream makes people much more recognizable then single shots. I like that you posted this.

 

[Sybertiger]

"3D-DNR technology provides better noise reduction, however it tends to blur the moving objects."

I think I had the 3D DNR cranked up at the time because I didn't have any test subject and the static image looked better. Possibly explains why, when I had a moving object I got facial blur.....or the dude's a warlock and something about interfering with camera systems...yeah, I watch too much Sci-Fi or did I see that in the X-files?

 

[DsineR]

Here's a good overview describing VBR & CBR, from IPVM.

 

[bigredfish]

VBR for LPR at 120ft when you get 2 maybe 3 frames of the plate at 30FPS, does not do as well as CBR for me.

 

[bigredfish]

Do you get decent pics at 1/60 in daytime or is a number missing?

Sure do. Not sure how the compression here messes with the image, but this is clear
as can be when frame by frame on my laptop
View attachment DVR_ch17_20190725112951_20190725113018.mp4

 

[wittaj]

Sure do. Not sure how the compression here messes with the image, but this is clar as can be when frame by frame on my laptop

WOW - I guess I have to adjust my other settings then because that is not a good speed on mine.

What does the 1/120 at night look like? Mine is way too dark at that rate.

I recognize some tire shine evidence in the wet driveway LOL - glad I'm not the only one that does that lol.

 

[bigredfish]

It’s all a matter of how much white light you have to start out with. I have coach lights that have 1000 lumens on each side of the garage.

I recall @mech has some set to 1/250

see my driveway samples this thread. 5442 bullets at 1/120

Best Dahua camera for night vision?

What Dahua camera model under $200 should I look at for the best night time clarity out to 50ft away? I have 4000 lumen LED motion lights on the front of my garage where I plan to mount the camera. Should I be looking at the night color technology? What other specs should I look at to do the...

ipcamtalk.com

 

[wittaj]

It’s all a matter of how much white light you have to start out with. I have coach lights that have 1000 lumens on each side of the garage.

I recall @mech has some set to 1/250

see my driveway samples this thread. 5442 bullets at 1/120

Best Dahua camera for night vision?

What Dahua camera model under $200 should I look at for the best night time clarity out to 50ft away? I have 4000 lumen LED motion lights on the front of my garage where I plan to mount the camera. Should I be looking at the night color technology? What other specs should I look at to do the...

ipcamtalk.com

WOW - that is incredible. Have you posted screenshots of your other settings somewhere - my videos look nothing like those and I have a large amount of light in the proximity of the cams.

 

[bigredfish]

I was partially wrong, been I while since I checked, on one of the two 5442's I use Auto Exp during the day, the one in the video above.

 

[wittaj]

@bigredfish - thanks - I will adjust mine to that and see what I get. Are you picture values default (or close to it).

 

[bigredfish]

Yes, image settings default. (Note my experience is many Dahua cameras can use some added Contrast, the 5442's seem to be ok at 50)
WDR-18 during the day due to shadow from house across drive, and HLC50 at night which cuts glare when the flood kicks in

 

[Sybertiger]

I went back to CBR on all the cams...here's an example of a very cloudy day getting ready to rain. Just for grins, it's a contrast between the white guy no hat and black guy with hat plus hat guy is looking down all the time. But, note the door cam because of the overcast day with his skin tone loses detail even though it's 2pm. I got the guy dead to rights with his face about 30 inches from the cam in the middle of the day but......

Not a total loss but an example of how subtle changes make can make a big difference and sometimes it's out of your control to be able to account for everything when adjusting the cams.

 

[bigredfish]

And which frame you choose to capture. Movement may have stopped in that split second. When I choose a frame for a capture out of a video of my LPR cameras, some are better than others...

 

[area651]

View attachment 56114

Couple easy questions for you:
Why do you choose 264H instead of 264? (I do the same just b/c that's the default.)
you're "only" running 1080P resolution. If my camera can do "more" (like 3840x2160) then should I?
What is the "Quality" setting anyway? I have Amcrest cams and being rebranded Dahua, its pretty much the same options as yours. I just don't get why anyone would need to choose less than 6 which is best.
I notice you're running IFrame the same as your FPS. I noticed some of my cams run the I frame double of the FPS which I think would give you double the breaks to get a clear pic wouldnt it? Since the first thread you mentioned, I've just set all my cams equal like you have yours though.

 

[Sybertiger]

Couple easy questions for you:
Why do you choose 264H instead of 264? (I do the same just b/c that's the default.)
you're "only" running 1080P resolution. If my camera can do "more" (like 3840x2160) then should I?
What is the "Quality" setting anyway? I have Amcrest cams and being rebranded Dahua, its pretty much the same options as yours. I just don't get why anyone would need to choose less than 6 which is best.
I notice you're running IFrame the same as your FPS. I noticed some of my cams run the I frame double of the FPS which I think would give you double the breaks to get a clear pic wouldnt it? Since the first thread you mentioned, I've just set all my cams equal like you have yours though.

There are physically three mounted cams used in those shots. The side shot is from a 4MP 3.6mm IPC-T5442TM-AS, the doorbell shot is from a 2MP 2.8mm IPC-HDBW4231F-AS, the two downward shots are from a 2MP 3.6mm dual-cam IPC-HDBW4231F-E2-M. Only the side shot cam is 4MP the other two cams are 2MP which explains 1080P. You should run at the highest resolution of your cam.

H.264H will give you a little higher compression than regular H.264 saving bandwidth but more importantly save HD space.

6 is bigger than 5 so it must be better.

Set your FPS and iFrame to the same number. You got it backwards....the iFrame is a fixed frame interval that grabs a COMPLETE frame of info like snapping a picture. The FPS will set how many intermediary "partial frames" to be grabbed which are not really complete frames but rather a collection of "differences" from the fixed frame. iFrame set to 15 will grab twice as many fixed frames versus an iFrame set to 30.

 

[area651]

There are physically three mounted cams used in those shots. The side shot is from a 4MP 3.6mm IPC-T5442TM-AS, the doorbell shot is from a 2MP 2.8mm IPC-HDBW4231F-AS, the two downward shots are from a 2MP 3.6mm dual-cam IPC-HDBW4231F-E2-M. Only the side shot cam is 4MP the other two cams are 2MP which explains 1080P. You should run at the highest resolution of your cam.

H.264H will give you a little higher compression than regular H.264 saving bandwidth but more importantly HD space.

6 is bigger than 5 so it must be better.

Set your FPS and iFrame to the same number. You got it backwards....the iFrame is a fixed frame interval that grabs a COMPLETE frame of info like snapping a picture. The FPS will set how many intermediary "partial frames" to be grabbed which are not really complete frames but rather a collection of "differences" from the fixed frame. iFrame set to 15 will grab twice as many fixed frames versus an iFrame set to 30.

Thanks for the explanation. I had backed all my cams down to 1080P. I'll bump them back to max res.

Yeah, its just weird that quality option would even be there. I set all mine to "6/best" as well just b/c......I dunno. Just b/c that's best I guess.

And thanks for the iFrame clarification. keeping them set to the same is the plan. thanks again!