Worlds First Review - Dahua - IPC-Color4K-X / DH-IPC-HFW5849T1-ASE-LED - Full Color 4K Camera

Just to set a quick expectation. I’m going to take a day or 2 off to enjoy the holiday as well so will continue editing the night video together snd look to post early next week. Stay tuned and for those celebrating the 4th July have a happy and safe one with your family, friends and loved ones !

You deserve the time off. You have put a lot of man hours into this test along with the others that you have done. ;)
 
Enjoy the 4th, remember what it represents and remember it is possible only because of citizens with guns and cannons.
 
There absolutely would be a trade off on both cams for DOF but that goes with the territory on any lens build and there is absolutely room for a middle ground there in my opinion where both cams could benefit from improvement. Will be interesting to see their varifocal counterparts when released albeit have to watch for chromatic aberation.

I agree and I'd rather have the focus sharp close to the camera where you can get a good facial recognition shot than further away where the subject is less recognisable. This is where I agree out of the box, the Dahua may have better reconition although big caveat - it's kind of ironic that the Dahua does such a great job close up to the lens but then loses so much focus just a few feet further back where the Hik picks up the mantle. The closer pic always wins for facial recognition of course, but only if your subject walks that close and looks at the camera. Maybe if they don't see it and walk up to it looking at the structure to break in. If not that 10ft back focus as per the Hik is arguably more useful. To me an ideal would be to have a sharp focus from the camera all the way back to the wall. Beyond that a little drop off is acceptable as you're not going to ID anyone coming across the grass and anyone who comes closer walks into the ID zone. If they turn back, then no harm no foul.

Although it goes to movies raher than CCTV I like seeing the detail in a picture and the loss of grass and tree leaf detail spoils the quality appearance. There's good arguments for saying superfluos and only the face matters, but who doesn't like a quality image? Sometimes with wildlife or family other shots it's nice to capture a quality picture overall.

As for sharpening, probably something we're going to have to agree to disagree over. I'm guessing your background is broadcast cameraman or engineer sounding from you knowledge and I've had many a discussion with them in the past about TV and film pictures on tv forums. Many tend to like the old film 24fps look with blur and soft but detailed pictured. As someone more from the digital age, I like the opposite, detailied but super sharp razor edged life like pictures where the detail and shaprness jump out at you. My main suspicion with the Dahua is some kind of compression decision (I don't know much about the subject) as the leaves are splodges and the grass similar which to me suggests it's not sharpening that's affecting these areas but rather they've been replaced by textures as a part of the compression process. Relevence to CCTV? Not much as said before. But when you see the Hik pic, nice to have the detail retained.
 
  • Like
Reactions: vidmo and Parley
To get rid of the pulsing noise pattern mentioned you can increase the DNR TIME "Expert" mode has been added under "Image Enhancement" Digital Noise Reduction. Time DNR level. I believe this to be 3D noise cancelling setting and not sure what impact it will have on the picture as in Motion or Blur as not tested for this as yet but certainly fixes the pulsating picture as WC mentions as a problem.
 
To get rid of the pulsing noise pattern mentioned you can increase the DNR TIME "Expert" mode has been added under "Image Enhancement" Digital Noise Reduction. Time DNR level. I believe this to be 3D noise cancelling setting and not sure what impact it will have on the picture as in Motion or Blur as not tested for this as yet but certainly fixes the pulsating picture as WC mentions as a problem.

Already tested it. Unfortunately does impact motion blur and you have to increase quite a fair amount (over 50) to just reduce the pulsing very slightly. Doesn’t get rid of it because it’s HW/SOC based. This is definitely something needing to be fixed by Hik. Dahua had the same issue with v1 of the 2831 4Ks. So neither company (nor others to be fair) has been immune to this but either way should be taken care of these days by these manufacturers in design and build. Increasing NR is not something I advocate for day or night due to its impact. Keep it as low as possible so you don’t lose detail. Certainly can be used very sparingly but again this needs to be fixed as the issue will still be present and you don’t want to sacrifice video quality.
 
Got my hands on the ‘release’ version of the FW for the Dahua so when back from the holiday break will give it a run. Trying to make sure I actually take some time off :) However to set the right expectations this one does not have the final fixes I put forward incorporated yet or object monitoring re-enabled yet. It does however have the initial bug fixes etc I raised. My understanding is that the new fixes I recommended and Object Monitoring will be in the next release planned for end of July. As soon as I see a test release of that one will take a look too and let you know my thoughts. Just wanted to make sure people had the latest info on that.

Happy 4th !
 
All

Hope everyone had a safe and happy 4th !

Apologies for the delay in thread update and in DM response, I've not been slacking off :) just re-shooting the night comparison as I wanted to test the new 'shipping' version of the FW that I was provided on the Dahua. For clarity this is NOT the one that includes the final codebase fixes and tweaks I've put forward to (and have been accepted by) Dahua engineering but instead is based on the original bugs I identified (now fixed) and some tweaks that Dahua also completed themselves. However after testing this 'shipping' version, I saw some issues with those tweaks they made to AGC and underlying performance and so have recommended that when they ship the cam they instead use the original one* I provided feedback on (the bug fix version) and then look to finalize the one with my new tweaks so I can re-test ASAP. I'm told that the ETA on that 'Wildcat' tweaked version will be end of July to mid August.

Therefore, re-shot the night video for a 3rd time (I know, I know but hey I want you all to have the most accurate video in this review that represents current FW) so you would have a real comparison to what will be the new 'shipping version' :) Working on the edit to combine them and final video will be coming soon.

In the meantime as I know all this mention of FW versions doesn't help anybody without context, let me give you the versions below

  1. Initial Version (Released with cam for test) - 21-05-26
  2. * Original bug fixed version - 21-06-11
  3. Dahua original planned 'shipping version' - 21-06-24

Therefore I am recommending to Dahua that the new 'shipping' version be based off *21-06-11 code for now when production orders start to ship.

HTH and my thanks to the Dahua engineering team for listening to the feedback and recommendations (in the security side of business a lot of people don't so really do appreciate it) and for @EMPIRETECANDY for helping drive the priority of getting my tweaked version out ASAP. Will update you as soon as I have that one in my hands for testing which should also include the re-addition of the new Object Monitoring rule.
 
Well, thanks to Youtube (and its known bug of taking a video upload then failing at point of their processing so it isn't made available), I've had to upload this 3 times BUT at least now their processing is finally doing what it should be. Will check back in tomorrow make sure that has completed then chapter it and post here. Got to love technology ! :)
 
All

Finally YouTube finished it’s re-processing after failing twice ! So here is the next video which is focused on night testing. Lets jump into how I set this test and the caveats :)

In this case I white balanced both cams using pro cards in the FOV. While this certainly ensures white colors are more accurately represented, remember these are not pro level DSLR or Cinema cams and as with any cam, sensors lean towards specific hues (therefore making the images look warmer, cooler etc due to their lean towards red, blue, green). This of course can be compensated further if you manually tweak RED/BLUE specifically for each lighting condition in FOV so not necessarily a big deal but mentioning it for completeness. Also worth noting that I advocate for the importance of this for color at night. I mention below why I take a slightly different approach for this testing (setting for worst case).

I also took the opportunity to dial the cams in further. Both cams were tweaked slightly and any settings that were touched as always have been added to the lower thirds in the video so you’ll see those. I also adjusted NR and gain to improve the image as much as possible without adding additional noise (reducing where possible) or inducing motion blur / frame smearing. These were dialed in for worst case (i.e. illuminator only testing) where there is generally little light, more noise and resulting images are often desaturated etc, then left these in place through the other tests. While this is great (and fair) when comparing side by side to keep consistency across multiple tests, if this were a final install I was doing or as part of a larger project I deploy globally, then I would be dialing in exposure/shutter + image settings specifically to the FOV, lighting conditions available and camera(s) used.

The Illuminator Only testing is completed with each manufacturers illuminators set to 100%. The reason for this is to show maximum reach of each set of onboard illuminators to give those that are using this without additional light, clear expectations on what is possible. We all know that AUTO mode on the illuminators will adjust as targets get closer to the cam (less light needed and to ensure washout is reduced) but again this test is to show power and light capabilities of each set of illuminators. Therefore any highlight blowouts are expected since these cams are NOT on AUTO in this test and due to the difference of AGC’s at code level I mentioned before that cannot be adjusted by end users. In the illuminator testing the FOV is only lit by these and individually so, i.e. when I call out Dahua illuminators then both cams are using only those listed NOT those + their own illuminators at the same time. This way critically you can see the image from each cam with only Dahua illuminators on then each cam with only Hik illuminators on. When getting to the additional light testing section illuminators are turned off. I also push the cams to 1/250 which I certainly would not use in this specific FOV with the lighting available BUT again show it so you understand what is possible.

As I’ve been asked about doing a versus approach (cams against each other) many times via DM for upcoming videos (on old and new cams), I wanted to address it here that any of those I entertain doing (if you’re interested) I would do under separate, non-review based threads. Therefore to be clear, this video here (as well as others in this thread) are NOT meant to be a ‘versus’ video but rather an opportunity to look at both these 4K, color at night cams together (+ the 5442 earlier) as I continue this Dahua cam review. This gives you the ability to see the performance of the Dahua camera under review here while understanding how it fits in this segment with others. My goal being to provide for a more informed understanding of extra features a camera like the Dahua may bring, quality of caps and what to expect in real-world deployments.

HTH

This is an 8.5 minute video with chapters included for easier navigation. Remember to view on YouTube and select 4K. As always, let me know with any questions.

Next Steps: I am going to look at publishing a re-shoot of the day footage, white balanced and then as this review is focused on the Dahua, will then do a wrap up post with my final thoughts on this specific cam.


VIDEO LINK
Compressor encoding issue so re-uploaded

 
Last edited:
Hikvision looks shit which I know is not the case with my 2.8mm version. Here is my back Yard which has no lighting except a couple of low power 2w solar lights and the built in Illuminator LED's of the Hikvision. I will turn off the solar lights later when dark but doubt they make much difference Parley and others have posted some clips also and they certainly look much better than the clips you have posted WC.

I have already posted clips from my 2.8mm Hikvision front camera with little motion blur with excellent night time clips in colour and the 4mm version should be better but not in your case they look very poor indeed.

Hikvision4kgate.20210709_004915493.58.jpg

Without the low power Solar Lights

Hikvision4kgate.20210714_224335755.jpg

This is the picture I get when I use your settings only the exposure I can only set to 1/50 or 1/75 so set to the higher 1/75 no option for me for 1/60.

Hikvision4kgate.20210714_225813973.jpg

Here is with additional 2x30w White LED lights which is Excellent if you want to keep them on at night or even have on a PIR.

Hikvision4kgate.20210714_231753310.jpg

Here is a picture taken from my Huawei P20 pro mobile phone

2021-07-14 23.03.48.jpg
 
Last edited:
I can see the improvements that Dahua has made to the camera since your first test. :) It is looking pretty darn good and I cannot wait to get one.

Andy, if you see this post, what I really need are some bullet varifocals with the 1/1.2" sensor and 4K. I know, I know, that I have mentioned this before. :)