Which 4k/8mp color night vision cameras???

Of course the Colorvu 8MP easily outperforms the 2MP for the same field of view because the ColorVu is on a 1/1.2" sensor while the 2MP and the 3MP you replaced is on the 1/2.8" sensor....
The 4mm 8MP G2 Colorvu outperforms the Hik 3MP varifocal DS-2CD2H35FWD-IZS, even with the 3MP zoomed in with a much narrower FOV. Both for light sensitivity and clarity (detail) e.g. reading number plates at the same distance. You're looking at specs for colour at 0.005 lux compared to 0.0005 lux.
 
The 4mm 8MP G2 Colorvu outperforms the Hik 3MP varifocal DS-2CD2H35FWD-IZS, even with the 3MP zoomed in with a much narrower FOV. Both for light sensitivity and clarity (detail) e.g. reading number plates at the same distance. You're looking at specs for colour at 0.005 lux compared to 0.0005 lux.

What part do you not understand or follow?

I said a 2MP or 3MP and an 8MP on the same 1/2.8" sensor size - the 2MP will win at night. And an 8MP on a 1/1.2" sensor will win the battle against the 2MP or 3MP on a 1/2.8" sensor. Where did I ever say different. I didn't.

Do your test with a budget cam 8MP on a 1/2.8" sensor and prove to me at night that it would be better than a 2MP on the same 1/2.8" sensor - it won't and the 2MP will win. That is the point I am making.

You are confusing sensor sizes and not making the same comparisons and missing the point. I am talking about different MPs on the same sensor size....and you want to keep arguing that your G2 on the much larger 1/1.2" sensor produces a better image than your 3MP on a 1/2.8" sensor. Of course if freaking does....I never said it wouldn't.

Now you want to argue that I said same field of view, but there is a little give and take in a field of view. Of course the DORI is better on the 8MP than the 3MP so the field of view can be a little wider on the 8MP versus 3MP (although the DORI numbers are exaggerated to begin with). But there does come a distance that a 2MP will win based on what is currently on the market. I will take my 2MP Z12E 64mm focal length at 90 feet over my 8MP 2.8mm fixed lens on the 1/1.2" sensor trying to get an IDENTIFY at 90 feet every night of the week.

Show me an object moving at 90 feet away at night with a 8MP 4mm focal length. You will not be able to use for IDENTIFY purposes....but a 2MP with enough optical zoom will be able to do so.

Your are comparing a 4mm focal length 8MP on a much larger 1/1.2" sensor to a 3MP on the 1/2.8" sensor for a focal length of 12mm. Of course the freaking 1/1.2" sensor will win because the 3MP isn't on an ideal MP/sensor ratio to begin with (plus the DORI is better), so zooming it in needs even more light. I am not saying it won't.

I do not pay attention to the color specs because every manufacturer plays games. Is that 0.005 lux at 1/3s shutter that is useless for motion or is it a 1/60s shutter that the lux is spec'ed at....

Again, my point is on not chasing MP and go for sensor size and if the sensor size is the same, at night the lower MP will generally beat the higher MP. But if the higher MP is on a larger sensor, then it wins. But also recognize the market also doesn't currently have 8MP on a large enough focal length for some applications and there are distances that a 2MP will be the better option due to a larger focal length that is available.
 
Last edited:
Remember, too, that sensor sizes are fractions and therefor inverse of what they may seem to be.

Disclaimer - These sizes are what the manufacturers advertise and may, or may not, be the true size of the sensor in the camera.
1/3" = .333" Great for 720P
1/2.8" = .357" (think a .38 caliber bullet) Great for 2MP
1/1.8" = .555" (bigger than a .50 caliber bullet or ball) Great for 4MP
1/1.2" = .833" (bigger than a 20mm chain gun round) Great for 8MP
 
But there does come a distance that a 2MP will win based on what is currently on the market. I will take my 2MP Z12E 64mm focal length at 90 feet over my 8MP 2.8mm fixed lens on the 1/1.2" sensor trying to get an IDENTIFY at 90 feet every night of the week.

Your IPC-HFW5231E-Z12E (I assume that's the model you mention above) has a " 1/2.8" 2MP Progressive Scan Starvis CMOS " sensor- apparently the same sensor/ size as the previously mentioned older 2MP Dahua IPC-HDW5231R-Z? But with a much more powerful 64mm max f1.4 lens (horizontal FOV 59 to only 5 degrees.) So at 64mm you've only got 5 degrees horizontal FOV, with no way to remotely move the camera to cover a different area? Minimum illumination is given as 0.05Lux/F1.4 ( Color,1/30s,30IRE), the same as the older 5231R-Z 0.05Lux/F1.6 ( Color,1/30s,30IRE) .

I took down a 5231R-Z and replaced it with a 8MP G2 Colorvu. Even with the 5231R-Z zoomed in, it's going to struggle as the low light performance isn't as good.
 
It is not worth debating with you anymore as you are missing the point and you are clearly one of the "more MP is better" fanboys and does not understand the concept of optical zoom LOL :lmao: Do you run your cameras on default/auto settings as well LOL.

But to make it clear for others that find this thread, I will leave you with this.

You are correct that the more optical zoom (or focal length) one has, the tighter the field of view. That is why we have more than 4 wide angle cameras and use varifocal cameras so that we can optically zoom in to pinch points and other areas of interest to get the clean IDENTIFY captures of someone at a distance beyond what our 2.8 to 4mm fixed lens cameras can capture. While the varifocals are great at helping to identify at a distance, they come at a cost of a reduced field of view, just like the wide-angles are great at seeing a wide area, but they come at the expense of IDENTIFY at distance.

For roughly the same distance to be covered to realistically IDENTIFY, yes an 8MP on a 1/1.2" sensor will beat a 2MP or 3MP on a 1/2.8" sensor. And especially during the day, the additional MP can bring some additional clarity for some digital zoom that the 2MP cannot. Nobody is questioning that. But what about the distances that are beyond the realistic IDENTIFY of a particular focal length. Unfortunately outside of PTZs, there are not varifocal options yet for 8MP on the 1/1.2" sensor to optically zoom for longer distances.

Don't get me wrong, I am not knocking an 8MP on the 1/1.2" sensor. It has its place, just like every camera does. I am simply stating that someone needs to get the correct camera for the area trying to be covered. A wide angle 2.8mm to IDENTIFY someone 50 feet away is the wrong camera regardless of how good the camera is. I am not going to use my Dahua 4K/X 8MP on the 1/1.2" sensor for IDENTIFY purposes 60 feet away - that is the wrong camera for that application. A 2.8mm camera to IDENTIFY someone within 10 feet is a good choice OR it is an overview camera to see something happened but not be able to identify who.

One camera cannot be the be all, see all. Each one is selected for covering a specific area. Most of us here have different brands and types, from fixed cams, to varifocals, to PTZs, each one selected for it's primary purpose and to utilize the strength of that particular camera.

You show your field of view earlier in the thread (which looks like way too much WDR btw and some purple chromatic aberration)...and you are proud that you can read the plates of the red car that you said is about 15m (50 feet) away. Good for you, with a static object you can read the plates of a car at that distance that you probably couldn't with the 2MP. But can you make out the facial features of someone 50 feet away with that camera? (the answer is no). And at night, forget it.

1656993933828.png




Our long time resident camera expert Wildcat ran the Dahua 4K/X 8MP 1/1.2" sensor equivalent of your G2 thru the paces. Keep in mind this 4K/X camera is incredible and has much fewer complaints here than the G2. Folks here have both and prefer the 4K/X.

He had the 3.6mm version and here is the screenshot from 40 feet in the ideal daylight, which based on DORI numbers is the supposed IDENTIFY distance for this camera with the 3.6mm lens and I think most of would agree that this is not IDENTIFY quality, even if digitally zoomed in (thus showing similar to LUX numbers that DORI numbers are estimates and a manufacturer can make it whatever they want and personally I found they need to be cut in half or more and even more for nighttime):

1656994122348.png

While I have an 8MP on the 1/1.2" sensor for an overview camera and to provide IDENTIFY within 10-15 feet of the camera, I also use varifocals zoomed in to pinch points and other areas of interest at a further distance out. Here is a snapshot from my 2MP set to the appropriate focal length to actually IDENTIFY at 40 feet.

guyAT40feet.jpg

Which one of the two images gives you the best option to IDENTIFY the person? I think most will agree the 2MP optically zoomed...


Now the above images were in optimal daylight, so what about nighttime.

Prove me wrong that your 4mm G2 can produce a better IDENTIFY image of a person 90 feet away than this at night of a moving person than my 2MP Z12E that only has the coach lights from my house as a source of light as there are no street lights and the neighbors do not turn their lights on....


1656993235842.png

Heck I will be nice and give you one to try at 80 feet:

1656993435772.png

As I have said, someone needs to identify the areas they want to cover and pick a camera focal length designed to cover that distance. In some instances, it may be a 2MP or 4MP that is the right camera and not an 8MP. DO NOT CHASE MP!!!

I look forward to you proving me wrong with images of moving people at 40 feet during the day and 90 feet and 80 feet at night from your G2 4mm fixed lens that are better than these 2MP examples....
 
Last edited:
For anyone wanting to compare the Hikvision Colorvu G2 8MP cameras to one of the older low light favourites, the 2MP varifocal Dahua IPC-HDW5231R-Z (with a 1/2.8 sensor, the camera locked in colour mode.) Both were set to 1/150s with the gain on the 5231 adjusted to around 70. (Even that was probably too high- I adjusted it until there was no improvement in the scene detail/ brightness) Same lighting conditions, no WDR/ HLC/ BLC used.

Dahua 5231R-Z

ch03_20220713050045 - Copy edited.jpg

Hikvision Hikvision DS-2CD2387G2-LU 8MP 4mm

ch04_20220713050045 - Copy edited.jpg

Exported from the NVR they both came off as 1920x1080 resolution - so native for the 2MP Dahua, but reduced in resolution for the 8MP Hik. That means I can't do a crop on the 8MP pictures to see how good the detail is- I'll have to check if it's some setting on the NVR that can be changed.

There's very poor street lighting (only one street light- in the shot off to the right.) My house is lower than the road and the driveway is made of light grey pavers (they were wet in the photos.) With the houses across the road being unlit at night, it complicates the camera set up to get the scene exposure even. I picked the shot of a black car as that colour is the hardest for the cameras to display. At the other extreme, people close to my house can have their facial features over exposed due to the dark background and them being close to the sensor lights on my house. (I need to do some more experimenting with HLC to see if it is worthwhile using.)

Dahua 5231R-Z
ch03_20220716194633 edit - Copy.jpg

Hikvision 4mm 8MP G2 Colorvu
ch04_20220716194633 edit - Copy.jpg

I have blurred out the car plate number on my car apart from the first letter. The 5231 was zoomed in a bit to compensate for it having less pixels, but when there isn't enough light the zoom is useless.

The f1.0 lenses on the 8MP Colourvu can present issues with focus for close up subjects. The minimum focus distance is greater than 3.6m for the 4mm versions I have.
 
For anyone wanting to compare the Hikvision Colorvu G2 8MP cameras to one of the older low light favourites, the 2MP varifocal Dahua IPC-HDW5231R-Z (with a 1/2.8 sensor, the camera locked in colour mode.) Both were set to 1/150s with the gain on the 5231 adjusted to around 70. (Even that was probably too high- I adjusted it until there was no improvement in the scene detail/ brightness) Same lighting conditions, no WDR/ HLC/ BLC used.

Dahua 5231R-Z

View attachment 133998

Hikvision Hikvision DS-2CD2387G2-LU 8MP 4mm

View attachment 133999

Exported from the NVR they both came off as 1920x1080 resolution - so native for the 2MP Dahua, but reduced in resolution for the 8MP Hik. That means I can't do a crop on the 8MP pictures to see how good the detail is- I'll have to check if it's some setting on the NVR that can be changed.

There's very poor street lighting (only one street light- in the shot off to the right.) My house is lower than the road and the driveway is made of light grey pavers (they were wet in the photos.) With the houses across the road being unlit at night, it complicates the camera set up to get the scene exposure even. I picked the shot of a black car as that colour is the hardest for the cameras to display. At the other extreme, people close to my house can have their facial features over exposed due to the dark background and them being close to the sensor lights on my house. (I need to do some more experimenting with HLC to see if it is worthwhile using.)

Dahua 5231R-Z
View attachment 134000

Hikvision 4mm 8MP G2 Colorvu
View attachment 134001

I have blurred out the car plate number on my car apart from the first letter. The 5231 was zoomed in a bit to compensate for it having less pixels, but when there isn't enough light the zoom is useless.

The f1.0 lenses on the 8MP Colourvu can present issues with focus for close up subjects. The minimum focus distance is greater than 3.6m for the 4mm versions I have.


This is super helpful thank you. Now I gotta break it to the wife that all 5 of our cameras broke at the same time...

That black car in the background of the first shot is exactly what I see with my 5231 so it's very cool to see how much better things could be.

+1 for neighbor's Hearst
 
This is super helpful thank you. Now I gotta break it to the wife that all 5 of our cameras broke at the same time...

Yes, that can be an issue! I update a couple of cameras every 3 to 4 years.

The Colorvu's have definitely been a big step up in low light performance. I've updated two cameras in the backyard with a different model 8MP- with IR capabilities- as I don't want the backyard lit up with visible light all night. I do have automated lights on permanently at night out the front (plus the sensor lights), but people with decent street lighting probably wouldn't need it. The onboard white light the Colorvu has should be enough.

I've enabled HLC on the Colorvus and will see how they perform with it enabled (as far as the over exposed faces go when close to the bright sensor lights.)
 
Just keep in mind that "ColorVu", StarLight", "FullColor" and so on are all marketing terms and have no technical meaning whatsoever. They are dreamed up by marketing departments. The best way to determine night performance is with real world video that includes motion.
 
Just keep in mind that "ColorVu", StarLight", "FullColor" and so on are all marketing terms and have no technical meaning whatsoever. They are dreamed up by marketing departments. The best way to determine night performance is with real world video that includes motion.

Exactly why I posted the examples. Same lighting conditions and exposure times.
 
  • Like
Reactions: JDreaming
Identical settings for two completely different cameras isn't the best way to test. Each camera needs to be tuned independently to achieve the best video. Even two cameras of the same make and model can have differences, variations in component tolerances, that make that important.
 
For anyone wanting to compare the Hikvision Colorvu G2 8MP cameras to one of the older low light favourites, the 2MP varifocal Dahua IPC-HDW5231R-Z (with a 1/2.8 sensor, the camera locked in colour mode.) Both were set to 1/150s with the gain on the 5231 adjusted to around 70. (Even that was probably too high- I adjusted it until there was no improvement in the scene detail/ brightness) Same lighting conditions, no WDR/ HLC/ BLC used.

Dahua 5231R-Z

View attachment 133998

Hikvision Hikvision DS-2CD2387G2-LU 8MP 4mm

View attachment 133999

Exported from the NVR they both came off as 1920x1080 resolution - so native for the 2MP Dahua, but reduced in resolution for the 8MP Hik. That means I can't do a crop on the 8MP pictures to see how good the detail is- I'll have to check if it's some setting on the NVR that can be changed.

There's very poor street lighting (only one street light- in the shot off to the right.) My house is lower than the road and the driveway is made of light grey pavers (they were wet in the photos.) With the houses across the road being unlit at night, it complicates the camera set up to get the scene exposure even. I picked the shot of a black car as that colour is the hardest for the cameras to display. At the other extreme, people close to my house can have their facial features over exposed due to the dark background and them being close to the sensor lights on my house. (I need to do some more experimenting with HLC to see if it is worthwhile using.)

Dahua 5231R-Z
View attachment 134000

Hikvision 4mm 8MP G2 Colorvu
View attachment 134001

I have blurred out the car plate number on my car apart from the first letter. The 5231 was zoomed in a bit to compensate for it having less pixels, but when there isn't enough light the zoom is useless.

The f1.0 lenses on the 8MP Colourvu can present issues with focus for close up subjects. The minimum focus distance is greater than 3.6m for the 4mm versions I have.

How far away are you from the camera?
 
Identical settings for two completely different cameras isn't the best way to test. Each camera needs to be tuned independently to achieve the best video.

For getting decent results (capturing moving subjects) the exposure time is critical- that's why both were set at 1/150 s. (1/150 for the Hik, 6.67mS for the Dahua.) I chose 1/150 as that is what I've been running since putting the Colorvu's up.

The Dahua has manual gain control- as I mentioned I set that to 70 (I had to push it that high for it to have any hope in those lighting conditions.) The Hikvision has AGC with no manual setting. The contrast on the Dahua had to be kept around 50% as below that it has the effect of washing out the whole scene. Noise control on the Dahua was around 50, the Hik a lot lower at ~30.

You're not going to be able to "tweak" the 5231 to the point where it matches the 8MP Colorvu in that situation at 1/150s - without lowering it's exposure time down to 1/25s, perhaps 1/50s. I was running the Dahua 5231R-Z and Hik 3MP 2335FWD-I (Powered by Dark Fighter) cameras prior to the Colorvu. Both of those models perform about the same in low light.

PS: The black car in the previous post was moving at around 35-40km/h and the person was walking.
 
For getting decent results (capturing moving subjects) the exposure time is critical- that's why both were set at 1/150 s. (1/150 for the Hik, 6.67mS for the Dahua.) I chose 1/150 as that is what I've been running since putting the Colorvu's up.

The Dahua has manual gain control- as I mentioned I set that to 70 (I had to push it that high for it to have any hope in those lighting conditions.) The Hikvision has AGC with no manual setting. The contrast on the Dahua had to be kept around 50% as below that it has the effect of washing out the whole scene. Noise control on the Dahua was around 50, the Hik a lot lower at ~30.

You're not going to be able to "tweak" the 5231 to the point where it matches the 8MP Colorvu in that situation at 1/150s - without lowering it's exposure time down to 1/25s, perhaps 1/50s. I was running the Dahua 5231R-Z and Hik 3MP 2335FWD-I (Powered by Dark Fighter) cameras prior to the Colorvu. Both of those models perform about the same in low light.

PS: The black car in the previous post was moving at around 35-40km/h and the person was walking.

While you are correct that exposure time is critical to catch moving objects, @sebastiantombs is more correct LOL. You have to dial in each camera to reach their full potential.

Plus, you have jacked up parameters like gain that it negates the fast shutter. To be honest, both of your sample images are not doing either camera any justice and you have turned them into looking like the consumer grade crap we see on Nextdoor LOL. Let's break it down....

And before I forget "Powered by Dark Fighter" is nothing but a meaningless marketing phrase.

You are suffering from the common mistake of trying to do too much with one camera!

Do you want this field of view to get clean captures of a perp near your car or a car moving on the street? You won't get both at night with one camera, and in the process you have screwed up the entire image to create mostly useless video in all conditions. Plus it is very difficult to get clean images of a moving car at night. Need a lot of things in your favor like great streetlights, great camera, and the right focal length to get it.

Gain works by amplifying the signal and trying to piece together missing parts to create an image. One can tell they have too much gain when there is too much noise in the static image and then ghost/blur in motion.

Look at all the black dot noise in your 5231 - why is that - too much freaking gain. And then with a car out in the street without much light, it has the ghost effect. This image is not representative of a 1/150 shutter due to the high gain values, and probably gamma and some others as well.

You mention you are using HLC at night. You should avoid that at night. That is part of the problem with your image.

You have way more light than I do. My neighbors don't have their lights on like yours do and I don't have a streetlight like you do and my nighttime images in Post #45 are better than yours...

1658455878335.png

And then the person that is what 15 feet from the camera, can anyone IDENTIFY or RECOGNIZE this person? This is as bad as or worse than a consumer grade Ring or Arlo LOL.

1658456217816.png

Why is this image so bad? Because you jacked up the parameters to make it useless. In your goal of getting color at night, you defeated the main purpose of getting clean captures. You should not have a washed out face at night running in color or a washed out plate of a non-moving car running in color. If you do, that is a clue that you effed up a parameter somewhere because that is not enough light to blind these little sensors.

Based on that field of view and estimating the light I think you have available, I think you would get better results running that camera at 1/60 shutter or maybe 1/90 with gain no higher than 50 and gamma no higher than 50.
 
Last edited:
Can't be such bad on that colorvu. I test on our IPC-Color4K-X, it's sharper than the colorvu, but can't be such worse. The camera i test is 3.6mm lens.
Your place has good light, so you sure you are using the main stream? You can see this camera i tested pic is very good.

1658464672130.jpeg
 
Here is my Hikvision 8mp with moving image and I will have to disagree with Andy sorry but always said the hikvision was better than the Dahua in the 4k-x thread when Wildcat compared it to the Hikvision 8mp colourvu.

Here is a moving image on a 2.8mm lens.

View attachment Hikvision.20220610_020000_1.mp4
 
What i mean @ljw2k 4K Colorvu pic can't be such bad as @triumph202 shows, should be some setting is wrong. Just make a default to the settings.
Your place even use a very normal cam can get great color pic because lots of light. So use a 4K 1/1.2 or a 4K 1/2.8 no big difference. It's time to update your cam now, hikvision have a 180 4K full color now, will be good for street cornor watching. Make a plan and show us some new video would be great!
 
Last edited:
I would not say my street is well lit as I am just fortunate to have 1 streetlight outside my house and 4k 1/1.2" as above is very good but 4k with a 1/2.8" would be pretty poor to be honest like the 4k I have in now.

Picture snapshot only for comparison with the Hikvision 8mp 1/1.2" sensor.
I have seen that dual Hikvision camera the new model but the sensor not good 1/2.8" I think so wouldn't interest me, plus the size of it.


Here is a pic of a 4k with smaller 1/2.8 or might even be 1/3 sensor that old.

Front.20220723_012911464.28.jpg
 
What i mean @ljw2k 4K Colorvu pic can't be such bad as @triumph202 shows, should be some setting is wrong. Just make a default to the settings.

Andy, it's the lack of light out past the front of the car- on the street and in the yards of the houses across the street- causing the issues at my place. It's hard to see anything across the road with your eyesight at night- that's how dark it is. I've got multiple lights on the house, but the cameras judge their exposure on the lighting of the whole scene - which is mostly darker/ black in the back ground. When people get close to the lights their faces are very brightly lit and so they get over exposed (same for the back of the car.) HLC can be used to reduce the over exposure, but the shadow detail (past the front of the car) can then turn to rubbish. I need more even lighting over the whole scene and to play around more with the HLC. I'm also going to test leaving another mains powered sensor light off (that one is located behind the car/ where the person was walking to. It's about 7ft off the ground, so right near where people's faces are.) I've already got one mains powered sensor light left turned off because it was simply too much light from one direction.

Your example with the main road is very well lit overall by all the street lights and possibly the moon is helping too? You'll also never get a very brightly lit person close to that cam. :)

1658464672130-jpeg.134175


The Colorvu does have a wide 130dB WDR- most cameras are around 120. That does help with the varied levels of light at night and I have noticed a big improvement in that area. With the cameras being able to recognise Humans in a scene, it would be good if someone could program an exposure option that after recognising the human would use the human's lighting level to set the scene's exposure.

Here's another example of a face being overexposed from bright light- check out the areas of bare facial skin on the LHS Dahua (actually, face and the arms.) They're basically completely white and washed out. Once again, a person close to a bright light source in front of a dark background. There's no pink skin colour. The RHS Hik isn't as bad, but still overexposed. If the black face mask wasn't being worn I don't think the face would be easy to recognise or provide a usable still shot to ID someone.

1658629362927.png

(scene shot from )