Hi everyone. This time I present you a quick review of Dahua HDW4431C-A. This is an official Chinese market model, not available in Europe and presumably nor in US. My friend bought two of these cams on Polish auction service for $80 each, so I decided to take one for quick testing.
The box is all Chinese language. Interestingly, the web address is not www.dahuasecurity.com but www.dahuatech.com
C-A model is lacking SD card slot, so the only option is to send recorded material to some remote storage like NVR or FTP. It is also less vandal proof, with plastic mounting. Only the camera ball is metal. There is no screw for securing the hold but instead we have a "click" button. Honestly the whole mounting construction looks so unreliable, that I had a feeling the camera will drop off due to its own weight. Luckily it didn't happen.
This camera is equipped with OmniVision's OV4689 4 megapixel 1/3 inch CMOS sensor and 2.8mm f2.0 lens. I've read quite a bit of negative opinion about its night time performance so was curious, how it will perform, compared to Dahua HDW4231EM-AS with Sony's IMX290 sensor and HDW5831R-ZE with IMX274.
As always, first I check GAIN scale and convert it to ISO. Similarly to my two other cams, this camera is reaching ISO20000. It seems to me that all Dahua cams - at least from Eco Savvy 3.0 line - have similar GAIN levels. Some will start at lower ISO, some at higher, but they all end up at ISO20000. That's the case with HDW4231, HDW5831, HDW4431 and HDW5631 (I got some samples from Aristobrat). And so, all of these cams will achieve same image brightness at top GAIN yet with different noise level. Of course this is true if all cams have same lens brightness. If the camera is equipped with f1.4 lens, the image will be brighter than f2.0 camera. That's what you can see on below comparison. The HDW5831 with f1.4 lens becomes brighter from middle gain upward, while both 4231 and 4431 - both equipped with f2.0 lens - remain almost the same till the end.
Here I've cropped the images to show the amount of detail/noise at 100% scale. To do that, I've cropped a bright part of the image and a dark part to see how noise reduction is behaving in differently lit regions.
It's clearly visible that 4431 is more noisy than 4231, however it is not as bad as I thought it will be. Until GAIN 80 the noise reduction is moderate and there is plenty of detail. At GAIN 90 and higher, the noise reduction becomes very aggressive and the image becomes really soft. The same happens to 5831 in which GAIN 70 is already very soft. Just remember that 4431 has a 2.0um pixel size while 5831 has only 1.62um pixels. If you compare BW mode in bright part of a scene, it seems the 4431 is comparable with 4231 until almost GAIN 100. It's the dark places, where 4431 can't keep up.
To show that even better, I set both 4431and 4231 to 1/100s exposure to reduce overexposure of bright area under the street lamp. Here you can compare whole images from both cameras, set to GAIN 100. Compare details of the car or bushes under the lamp.
4431 suffers from so called "white orbs effect". It creates a white circle in a place of strong light source, like street lamp or headlights. In my night time examples, notice the street lamp shape. Even distant lamps are circle shaped at lower gain.
The BW mode noise from 4231 resembles a fine analog film grain, while 4431 noise is more like a bulky digital noise with lot of big, dark and bright specks flickering all over the image.
And the same comparison in color mode:
The ISO/GAIN diagram looks like this:
If we add lens brightness to the equation, the diagram of image brightness will look as follows:
I've added to the diagram a new Dahua HFW4239T-ASE Full color Starlight camera. It is equipped with f1.0 lens and IMX290 sensor (finally they came to this idea!). Thus I was able to simulate its performance, moving 4231 curve 2 stops up (the difference between f2.0 and f1.0 is exactly 2EV). I'm eagerly waiting to get this camera. Maybe Empiretecandy will provide it for testing
The daytime performance is good. There is no huge difference between 2 megapixel and 4 megapixel resolution, but remember that my 4231 has a 3.6mm lens while 4431 has 2.8mm.
Differences show up when it comes to color/brightness setup as well as WDR. When I tested 5831 opposed to 4231 - both based on SONY sensors - the differences in image look were minor. They had similar color saturation and image brightness. 5831 had better WDR range, but that was all.
In case of 4431 the image looks way different (all parameters set to 50). It is more saturated than two other cams. The camera seems to calculate the exposure lower than 4231 and it favors greens and blues, similar to Nikon digital cameras. So the sky and vegetation are slightly oversaturated. It's a matter of taste weather you like it or not. Personally I like the way it looks. To match colors in my 4231, I had to tweak the image settings heavily. But it worked only for certain lighting conditions. In different lighting, the settings would have to be different. So no matter what, you will get different look from this camera, than from SONY sensor based cams.
The WDR on 4431 works better in my opinion, than on 4231 or 5831. Similarly to my cams, it activates from level 45. But it behaves differently. It turns on and off much faster than on IMX sensors. When I turn on WDR on my cams, they go through a 3-4 seconds process of stacking multiple exposures. Then the image brightens up and stabilizes. On 4431 it's just one simple change from regular image to WDR mode. It takes less than a second. On IMX sensor based cams the WDR looks fine in a range from 45 to say 50. Higher values introduce a surrealistic, over-processed image. On 4431 WDR looks natural from level 45 till 100. There is not as much difference between these two values as on 4231 or 5831 cams. Lastly, WDR on 4231 creates some artifacts with moving as well as static objects. When enabled, the bright pixels are shifted one or two pixels to the left in reference to darker areas on the image. Below is an example of such case. Notice the pole under street lamp. It looks jagged. On 4431 I haven't noticed such problem.
Moving objects create ghosting effect when WDR is on. That's because two (or three in some cameras) differently exposured frames are captured one after another, and so there is some time lag between them. This is well visible on example below. The only idea that came to my mind, to trigger this effect, was to spin an iron cable in front of both cams.
I wonder what my neighbors thought, seeing me standing on a terrace with iron in one hand and spinning the cable with another ;D Anyway, as you can see, ghosting on 4231 is very strong. This camera can shoot 50 (PAL) fps. But with WDR it drops to 25fps. The 5831 can do 15 fps, but with WDR this drops to 14 fps. The 4431 is not losing any frame rate with WDR on. Furthermore, the ghosting is very small, as if the two exposures were captured in a very short interval of time.
At first I thought 4231 is losing half of its frame rate with WDR on, to use half of frames as dark exposures. But after stacking a few frames from my "cable test" it turned out it's not the case. In such scenario, the cable and its "ghost" would be evenly spread across its path, with equal distance between cables. As you can see below, this is not true.
The distance between cable and "ghost" is around 1/8 of distance between two footage frames. So because I recorded at 25fps, the time between two different WDR exposures is around 1/200s. In case of 4431 this time must be much shorter as you can barely see the ghosting effect.
Overall I have to say, this camera is not so bad. Especially when considering its price. Night time performance is poor and there is no SD card slot. A plastic mounting is crappy. But for a porpoise of weather cam (which in fact it is) it will serve well.
The box is all Chinese language. Interestingly, the web address is not www.dahuasecurity.com but www.dahuatech.com
C-A model is lacking SD card slot, so the only option is to send recorded material to some remote storage like NVR or FTP. It is also less vandal proof, with plastic mounting. Only the camera ball is metal. There is no screw for securing the hold but instead we have a "click" button. Honestly the whole mounting construction looks so unreliable, that I had a feeling the camera will drop off due to its own weight. Luckily it didn't happen.
This camera is equipped with OmniVision's OV4689 4 megapixel 1/3 inch CMOS sensor and 2.8mm f2.0 lens. I've read quite a bit of negative opinion about its night time performance so was curious, how it will perform, compared to Dahua HDW4231EM-AS with Sony's IMX290 sensor and HDW5831R-ZE with IMX274.
As always, first I check GAIN scale and convert it to ISO. Similarly to my two other cams, this camera is reaching ISO20000. It seems to me that all Dahua cams - at least from Eco Savvy 3.0 line - have similar GAIN levels. Some will start at lower ISO, some at higher, but they all end up at ISO20000. That's the case with HDW4231, HDW5831, HDW4431 and HDW5631 (I got some samples from Aristobrat). And so, all of these cams will achieve same image brightness at top GAIN yet with different noise level. Of course this is true if all cams have same lens brightness. If the camera is equipped with f1.4 lens, the image will be brighter than f2.0 camera. That's what you can see on below comparison. The HDW5831 with f1.4 lens becomes brighter from middle gain upward, while both 4231 and 4431 - both equipped with f2.0 lens - remain almost the same till the end.
Here I've cropped the images to show the amount of detail/noise at 100% scale. To do that, I've cropped a bright part of the image and a dark part to see how noise reduction is behaving in differently lit regions.
It's clearly visible that 4431 is more noisy than 4231, however it is not as bad as I thought it will be. Until GAIN 80 the noise reduction is moderate and there is plenty of detail. At GAIN 90 and higher, the noise reduction becomes very aggressive and the image becomes really soft. The same happens to 5831 in which GAIN 70 is already very soft. Just remember that 4431 has a 2.0um pixel size while 5831 has only 1.62um pixels. If you compare BW mode in bright part of a scene, it seems the 4431 is comparable with 4231 until almost GAIN 100. It's the dark places, where 4431 can't keep up.
To show that even better, I set both 4431and 4231 to 1/100s exposure to reduce overexposure of bright area under the street lamp. Here you can compare whole images from both cameras, set to GAIN 100. Compare details of the car or bushes under the lamp.
4431 suffers from so called "white orbs effect". It creates a white circle in a place of strong light source, like street lamp or headlights. In my night time examples, notice the street lamp shape. Even distant lamps are circle shaped at lower gain.
The BW mode noise from 4231 resembles a fine analog film grain, while 4431 noise is more like a bulky digital noise with lot of big, dark and bright specks flickering all over the image.
And the same comparison in color mode:
The ISO/GAIN diagram looks like this:
If we add lens brightness to the equation, the diagram of image brightness will look as follows:
I've added to the diagram a new Dahua HFW4239T-ASE Full color Starlight camera. It is equipped with f1.0 lens and IMX290 sensor (finally they came to this idea!). Thus I was able to simulate its performance, moving 4231 curve 2 stops up (the difference between f2.0 and f1.0 is exactly 2EV). I'm eagerly waiting to get this camera. Maybe Empiretecandy will provide it for testing
The daytime performance is good. There is no huge difference between 2 megapixel and 4 megapixel resolution, but remember that my 4231 has a 3.6mm lens while 4431 has 2.8mm.
Differences show up when it comes to color/brightness setup as well as WDR. When I tested 5831 opposed to 4231 - both based on SONY sensors - the differences in image look were minor. They had similar color saturation and image brightness. 5831 had better WDR range, but that was all.
In case of 4431 the image looks way different (all parameters set to 50). It is more saturated than two other cams. The camera seems to calculate the exposure lower than 4231 and it favors greens and blues, similar to Nikon digital cameras. So the sky and vegetation are slightly oversaturated. It's a matter of taste weather you like it or not. Personally I like the way it looks. To match colors in my 4231, I had to tweak the image settings heavily. But it worked only for certain lighting conditions. In different lighting, the settings would have to be different. So no matter what, you will get different look from this camera, than from SONY sensor based cams.
The WDR on 4431 works better in my opinion, than on 4231 or 5831. Similarly to my cams, it activates from level 45. But it behaves differently. It turns on and off much faster than on IMX sensors. When I turn on WDR on my cams, they go through a 3-4 seconds process of stacking multiple exposures. Then the image brightens up and stabilizes. On 4431 it's just one simple change from regular image to WDR mode. It takes less than a second. On IMX sensor based cams the WDR looks fine in a range from 45 to say 50. Higher values introduce a surrealistic, over-processed image. On 4431 WDR looks natural from level 45 till 100. There is not as much difference between these two values as on 4231 or 5831 cams. Lastly, WDR on 4231 creates some artifacts with moving as well as static objects. When enabled, the bright pixels are shifted one or two pixels to the left in reference to darker areas on the image. Below is an example of such case. Notice the pole under street lamp. It looks jagged. On 4431 I haven't noticed such problem.
Moving objects create ghosting effect when WDR is on. That's because two (or three in some cameras) differently exposured frames are captured one after another, and so there is some time lag between them. This is well visible on example below. The only idea that came to my mind, to trigger this effect, was to spin an iron cable in front of both cams.
I wonder what my neighbors thought, seeing me standing on a terrace with iron in one hand and spinning the cable with another ;D Anyway, as you can see, ghosting on 4231 is very strong. This camera can shoot 50 (PAL) fps. But with WDR it drops to 25fps. The 5831 can do 15 fps, but with WDR this drops to 14 fps. The 4431 is not losing any frame rate with WDR on. Furthermore, the ghosting is very small, as if the two exposures were captured in a very short interval of time.
At first I thought 4231 is losing half of its frame rate with WDR on, to use half of frames as dark exposures. But after stacking a few frames from my "cable test" it turned out it's not the case. In such scenario, the cable and its "ghost" would be evenly spread across its path, with equal distance between cables. As you can see below, this is not true.
The distance between cable and "ghost" is around 1/8 of distance between two footage frames. So because I recorded at 25fps, the time between two different WDR exposures is around 1/200s. In case of 4431 this time must be much shorter as you can barely see the ghosting effect.
Overall I have to say, this camera is not so bad. Especially when considering its price. Night time performance is poor and there is no SD card slot. A plastic mounting is crappy. But for a porpoise of weather cam (which in fact it is) it will serve well.
