Hi everyone. Today I present you a review of full color starlight+ camera, Dahua IPC-T5442TM-AS-LED. I got this one from Andy in exchange for in-depth review. Thank you Andy 
This is a 4 megapixel, turret type camera with white light illumination. It is equipped with 1/1.8-inch type CMOS sensor, which is supposed to give good image quality in low light conditions. This model has no ICR, which means it won’t be able to see any IR illumination. The housing is very well done, with all elements made of metal. Additionally, unlike other turret models I’ve tested, it has a direct access to microSD card slot as well as reset button. Oddly, there are regular cross type screws instead of star one. In comparison to HDW4231EM-AS the camera has the same size.
The sensor itself is a bit of mystery for me. At the moment of announcement of Eco-savvy 4.0 AI series, there was only one sensor I found, that - to some extend - matched Dahua’s specs. It was SmartSens SC4210. However after detailed reading, I found it is not a 4MP but 3,9MP sensor, and its resolution doesn’t match any of 4MP Dahua’s cams. Miraculously, a week after I got the camera for testing, Sony announces its 4MP Starvis 1/1.8-inch IMX347 which ideally matches camera specs. Dahua however is not stating that this is a Starvis sensor, while they did it in every other camera based on Sony sensor. Another two weeks later, OmniVision announces its 4MP OS04A10, also matching the specs. So finally I have no clue which one is inside. After playing with it for a while, I can say, that the noise pattern is different from any Sony sensor based camera I own. It more resembles the one from HDW4431C-A I’ve tested, which was based on OmniVision’s sensor (I do not mean noise level, only its “nature” or “type of grain”). Both Sony’s and OV’s sensors have a 2,9um pixel size, which suggests that image quality should match IMX290 (used for example in HDW5231R-ZE or HDW4231EM-AS) which also has a pixel size of 2.9um. It is of course also dependent on lens brightness used in certain camera.
As always, first I asses the GAIN scale, translating it to ISO scale. As this is a Full Color model, I would expect it to reach much higher sensitivity, than regular models with ICR. At least this was a case with HFW4239T, which reached a stunning ISO200000.
This model stretches its sensitivity from ISO50 to around ISO40000. This is 1EV brighter than non-Full Color Dahua cam. This is not much in exchange for losing ICR.
At daylight, the camera provides a very nice looking, sharp image. The firmware lets us set maximum bitrate to 20000Kbps, which is a lot. At 11000Kbps there is already no compression visible, nor any pulsation from key frames. I have found no problems with codec artifacts around moving objects as in case of HFW1831.
Comparing this camera to any other in terms of image detail might be tricky because I got a model with 2.8mm lens, while all other cams I own are equipped with 3.6mm lenses. The difference in field of view is significant. The only way to asses detail level is to set two cameras in a distance, one after another (wide angle cam in front), so that an object of interest on both will be of the same size in respect to whole image. There is only a certain region in the image, where objects appear to be more less the same size in frame in both cams. I've roughly marked this region in the image below:
Everything further than this region works in favor of 4231, while everything closer favors T5442 (dolly zoom). So here is a side by side comparison with HDW4231, where T5442 was about 3 meters in front of 4231. Most recorded objects appear beyond the "region of same scale", so this comparison as well as night footage is slightly unfair for T5442.
The WDR function works very well. As always true WDR activates from 45% up. It turns on instantaneously, with no flickering. With a little bit of tweaking the image looks very pleasing with lot of details in shadow and highlight areas. There are no distortions on edges of bright-to-dark regions. Below 45% only the image curve is affected, bumping up shadows. There is no exposure stacking. This however causes some detail degradation in flat looking areas like grass or walls. So be careful. WDR artifacts due to image stacking are at the same level as in other cameras I’ve tested. So only a fast moving objects in certain conditions will exhibit "tearing". At 110 degrees field of view even a really fast moving car is hardly showing any WDR distortion.
When it gets dark, things at first glance doesn't look so great. With no motion in frame, the camera delivers a nice image, with plenty details. At factory settings, noise reduction is set to 50%. At this level noise is well reduced on static image with one exception - blue color. Every piece of the image with blue color, like late evening sky, is filled with nasty glittering grain. It’s not a subtle analog looking grain like in case of HFW1831, but rather an unpleasant digital mess, with bright sparkles. Rising noise reduction to 60% reduces this effect, but of course also softens whole image. When something starts to move in frame, it seems the noise reduction can’t keep up with amount of noise. My first impression was that noise reduction is better than in other cameras, because even at higher values there is virtually no ghosting – which is a common problem in other camera models. But instead of ghosting we get an odd pixilation effect. It looks like if a moving object was disintegrating into a bunch of tiny pixels. There is however a solution to this problem, which I will describe in a moment.
Generally, when compared to HDW4231 the image looks much better. A WDR mode at night works very well. Not only it slightly reduces overall noise, but also brings up shadow and reduces blown areas like street lamps. In case of 4231 or 5831 WDR in most cases has a negative impact on image quality.
HDW-T5442 has no ICR, so switching to black&white mode isn't rising the IR filter, and so the camera is unable to see IR light. The only other model I have with no ICR is HFW4239T. I decided to compare them. While in case of 4239T switching to BW does nothing but color reduction, the T 5442 loses brightness by about 0.5EV. It also improves noise appearance, reducing chroma (color) noise and making it less distracting. I switched both cams to MJPG compression and turned off noise reduction, to compare noise pattern.
Now, here is a nifty little trick to get rid of excess of noise in T5442. I came to this by playing with different settings. It turned out, that the main reason for high noise and pixilation effect is... sharpness setting. Normally, i.e. in all other cameras I’ve tested, sharpness increases contrast on the edge of objects in the image, having little effect on grain level. I suppose it looks like this: first the noise reduction is applied, and then sharpness. In case of T5442 it seems that sharpness is applied straight to raw image before noise reduction. This causes a significant noise increase if you apply values higher than 40-50%. Look at the example below.
At 50% sharpness and 50% NR the noise is already visible. At 75% sharpness it looks terrible. But if you lower sharpness to 25%, bright sparkles disappear. The image is still quite sharp. Now it’s enough to set NR to 20-30%.
There are a few new features inside Web Interface that are worth mentioning.
First one is a possibility to set only a portion of an image to be displayed in Sub Stream2. I used it to asses noise detail while recording with MJPEG codec. It’s a nice feature when you have a wide angle lens.
Another new feature is Day/Night profile. Now you can set 2 different profiles for day and night, while having Day & Night mode set to “Auto”. The only drawback is that during the day, you won’t be able to see W/B mode. You can set all the other parameters, but W/B mode will turn on automatically, when it gets dark. I’ve tested it just once and it seemed to work fine.
There is now a new feature called LDC, which I think stands for Lens Distortion Correction. When turned on, it straightens a barrel distortion of the lens. It will however cut of a considerable portion of the image.
In playback section there are two new features. One lets you sort recorded material (video or snapshot) according to a fairly long list of parameters.
Another one is a very handy Batch Download window. Finally you can select a time span and download all videos or images at once. This method is also much faster than standard one by one download. In my case a 295MB mp4 file took 2min and 1s to download. The same file downloaded through Batch Download window is on my PC in 33 seconds.
The T5442 is much better than its predecessor, the 4431. It is also clearly better than HDW4231EM-AS, and is capable of delivering a very nice image when used wisely. Very often I see threads on this forum, where users complain about blurred, grainy image from camera that was supposed to be very good. It than turns out, they set everything to AUTO mode. It’s the same as buying an expensive DSLR camera and wishing to get nice pictures just by setting an AUTO mode. It doesn’t work. I always say - know your gear, learn how to get as much as possible from it, by learning how it works. The cameras we discuss here are also sophisticated equipment. It’s not enough to buy it and wish it will just work. If you pay a little bit attention to your image settings, you'll get a nice image. If you leave it on default settings, you might be slightly disappointed. And it actually concerns every camera you buy.
This is a 4 megapixel, turret type camera with white light illumination. It is equipped with 1/1.8-inch type CMOS sensor, which is supposed to give good image quality in low light conditions. This model has no ICR, which means it won’t be able to see any IR illumination. The housing is very well done, with all elements made of metal. Additionally, unlike other turret models I’ve tested, it has a direct access to microSD card slot as well as reset button. Oddly, there are regular cross type screws instead of star one. In comparison to HDW4231EM-AS the camera has the same size.
The sensor itself is a bit of mystery for me. At the moment of announcement of Eco-savvy 4.0 AI series, there was only one sensor I found, that - to some extend - matched Dahua’s specs. It was SmartSens SC4210. However after detailed reading, I found it is not a 4MP but 3,9MP sensor, and its resolution doesn’t match any of 4MP Dahua’s cams. Miraculously, a week after I got the camera for testing, Sony announces its 4MP Starvis 1/1.8-inch IMX347 which ideally matches camera specs. Dahua however is not stating that this is a Starvis sensor, while they did it in every other camera based on Sony sensor. Another two weeks later, OmniVision announces its 4MP OS04A10, also matching the specs. So finally I have no clue which one is inside. After playing with it for a while, I can say, that the noise pattern is different from any Sony sensor based camera I own. It more resembles the one from HDW4431C-A I’ve tested, which was based on OmniVision’s sensor (I do not mean noise level, only its “nature” or “type of grain”). Both Sony’s and OV’s sensors have a 2,9um pixel size, which suggests that image quality should match IMX290 (used for example in HDW5231R-ZE or HDW4231EM-AS) which also has a pixel size of 2.9um. It is of course also dependent on lens brightness used in certain camera.
As always, first I asses the GAIN scale, translating it to ISO scale. As this is a Full Color model, I would expect it to reach much higher sensitivity, than regular models with ICR. At least this was a case with HFW4239T, which reached a stunning ISO200000.
This model stretches its sensitivity from ISO50 to around ISO40000. This is 1EV brighter than non-Full Color Dahua cam. This is not much in exchange for losing ICR.
At daylight, the camera provides a very nice looking, sharp image. The firmware lets us set maximum bitrate to 20000Kbps, which is a lot. At 11000Kbps there is already no compression visible, nor any pulsation from key frames. I have found no problems with codec artifacts around moving objects as in case of HFW1831.
Comparing this camera to any other in terms of image detail might be tricky because I got a model with 2.8mm lens, while all other cams I own are equipped with 3.6mm lenses. The difference in field of view is significant. The only way to asses detail level is to set two cameras in a distance, one after another (wide angle cam in front), so that an object of interest on both will be of the same size in respect to whole image. There is only a certain region in the image, where objects appear to be more less the same size in frame in both cams. I've roughly marked this region in the image below:
Everything further than this region works in favor of 4231, while everything closer favors T5442 (dolly zoom). So here is a side by side comparison with HDW4231, where T5442 was about 3 meters in front of 4231. Most recorded objects appear beyond the "region of same scale", so this comparison as well as night footage is slightly unfair for T5442.
The WDR function works very well. As always true WDR activates from 45% up. It turns on instantaneously, with no flickering. With a little bit of tweaking the image looks very pleasing with lot of details in shadow and highlight areas. There are no distortions on edges of bright-to-dark regions. Below 45% only the image curve is affected, bumping up shadows. There is no exposure stacking. This however causes some detail degradation in flat looking areas like grass or walls. So be careful. WDR artifacts due to image stacking are at the same level as in other cameras I’ve tested. So only a fast moving objects in certain conditions will exhibit "tearing". At 110 degrees field of view even a really fast moving car is hardly showing any WDR distortion.
When it gets dark, things at first glance doesn't look so great. With no motion in frame, the camera delivers a nice image, with plenty details. At factory settings, noise reduction is set to 50%. At this level noise is well reduced on static image with one exception - blue color. Every piece of the image with blue color, like late evening sky, is filled with nasty glittering grain. It’s not a subtle analog looking grain like in case of HFW1831, but rather an unpleasant digital mess, with bright sparkles. Rising noise reduction to 60% reduces this effect, but of course also softens whole image. When something starts to move in frame, it seems the noise reduction can’t keep up with amount of noise. My first impression was that noise reduction is better than in other cameras, because even at higher values there is virtually no ghosting – which is a common problem in other camera models. But instead of ghosting we get an odd pixilation effect. It looks like if a moving object was disintegrating into a bunch of tiny pixels. There is however a solution to this problem, which I will describe in a moment.
Generally, when compared to HDW4231 the image looks much better. A WDR mode at night works very well. Not only it slightly reduces overall noise, but also brings up shadow and reduces blown areas like street lamps. In case of 4231 or 5831 WDR in most cases has a negative impact on image quality.
HDW-T5442 has no ICR, so switching to black&white mode isn't rising the IR filter, and so the camera is unable to see IR light. The only other model I have with no ICR is HFW4239T. I decided to compare them. While in case of 4239T switching to BW does nothing but color reduction, the T 5442 loses brightness by about 0.5EV. It also improves noise appearance, reducing chroma (color) noise and making it less distracting. I switched both cams to MJPG compression and turned off noise reduction, to compare noise pattern.
Now, here is a nifty little trick to get rid of excess of noise in T5442. I came to this by playing with different settings. It turned out, that the main reason for high noise and pixilation effect is... sharpness setting. Normally, i.e. in all other cameras I’ve tested, sharpness increases contrast on the edge of objects in the image, having little effect on grain level. I suppose it looks like this: first the noise reduction is applied, and then sharpness. In case of T5442 it seems that sharpness is applied straight to raw image before noise reduction. This causes a significant noise increase if you apply values higher than 40-50%. Look at the example below.
At 50% sharpness and 50% NR the noise is already visible. At 75% sharpness it looks terrible. But if you lower sharpness to 25%, bright sparkles disappear. The image is still quite sharp. Now it’s enough to set NR to 20-30%.
There are a few new features inside Web Interface that are worth mentioning.
First one is a possibility to set only a portion of an image to be displayed in Sub Stream2. I used it to asses noise detail while recording with MJPEG codec. It’s a nice feature when you have a wide angle lens.
Another new feature is Day/Night profile. Now you can set 2 different profiles for day and night, while having Day & Night mode set to “Auto”. The only drawback is that during the day, you won’t be able to see W/B mode. You can set all the other parameters, but W/B mode will turn on automatically, when it gets dark. I’ve tested it just once and it seemed to work fine.
There is now a new feature called LDC, which I think stands for Lens Distortion Correction. When turned on, it straightens a barrel distortion of the lens. It will however cut of a considerable portion of the image.
In playback section there are two new features. One lets you sort recorded material (video or snapshot) according to a fairly long list of parameters.
Another one is a very handy Batch Download window. Finally you can select a time span and download all videos or images at once. This method is also much faster than standard one by one download. In my case a 295MB mp4 file took 2min and 1s to download. The same file downloaded through Batch Download window is on my PC in 33 seconds.
The T5442 is much better than its predecessor, the 4431. It is also clearly better than HDW4231EM-AS, and is capable of delivering a very nice image when used wisely. Very often I see threads on this forum, where users complain about blurred, grainy image from camera that was supposed to be very good. It than turns out, they set everything to AUTO mode. It’s the same as buying an expensive DSLR camera and wishing to get nice pictures just by setting an AUTO mode. It doesn’t work. I always say - know your gear, learn how to get as much as possible from it, by learning how it works. The cameras we discuss here are also sophisticated equipment. It’s not enough to buy it and wish it will just work. If you pay a little bit attention to your image settings, you'll get a nice image. If you leave it on default settings, you might be slightly disappointed. And it actually concerns every camera you buy.
Last edited:
