Dahua HFW-1831E in-depth review

Discussion in 'Dahua' started by wopi82, Sep 14, 2018.

Share This Page

  1. wopi82

    wopi82 Getting comfortable

    Joined:
    Feb 9, 2017
    Messages:
    84
    Likes Received:
    386
    Location:
    Poland
    I present you a review of HFW-1831E camera. This is a h265 Lite series model, most likely equipped with fairly new IMX334 Sony sensor. This is a 1/1.8 inch, 8 megapixel CMOS with a pixel size of 2 micrometers. I got a model with 4mm lens, giving a horizontal FOV of around 87 degrees. I will oppose it to 5831R-ZE model, which is equipped with IMX274 CMOS (1/2.5 inch, 8MP) and also 4231EM-AS Stalight camera, equipped with very popular IMX290 fullHD sensor. I got the camera from Andy in exchange for a in-depth review. Thank you Andy :)

    Before I start the review, I would like to mention a few things. For this review I decided to relay both on video examples with some motion, as well as snaps from recorded clips. Previous reviews were based mostly on static snaps from cameras, but as some users noticed, the real quality of the image is when some motion appears. Also I changed a method of taking stills. Previously these were mostly snapshots from Web Interface. I've noticed however, that this function severely reduces amount of data in the picture. It cuts off shadows and highlights. See an example below.

    snap.jpg

    On the left is a bitmap file taken as snapshot with Web Interface function. On the right is a paused h264 movie of the same scene. At first glance not a big difference. But when I squeeze the histogram (brighten the image) it turns out, the left image has no data in the background grass part. The same goes for highlights. I've already mentioned this in HDW5831R-ZE review. Most of WDR data might be lost, when using snapshot button in Web Interface. So be careful when using it, especially in vulnerable situations, when every detail of the image is necessary, for recognition. In this review as well as HFW4239T review, I used only frames grabbed straight from DAV files.

    The movie files in this review were also untouched during montage. I used DAV files, converted to AVI (no recompression) and then converted to mp4 (also no recompression). The only additional compression you see is introduced by Youtube (please, watch all Youtube material in 4K resolution!). All this, together with a few other obstacles I encountered during testing, made me testing this camera for over a month. Sorry for such delay.

    The first thing to note, is that this model is very compact. It is way smaller than 4239T. Build quality is good. The only plastic part is top and bottom white cover. All the rest seems to be metal. A very important difference between this model and its varifocal brother, is that it has NO SD CARD SLOT. I wasn’t aware of it. Before I checked the specs, I was searching a place to disassemble the camera and put an SD card in. It took me at least 10 minutes to realize that something is wrong :) Here is a link to a Thread where Burton_Flooring shows how to disassemble it.

    1831E-photos.jpg

    The camera ISO ranges from around 50 to 10000.

    both cams.jpg

    Top gain is 1EV lower than most of the other cams I’ve tested. There is however a fairly bright lens onboard with f1.6 aperture, so final brightness scale from 0% to 100% gain should be almost identical to HDW4231EM-AS. In comparison to 5831R-ZE the starting gain will be 1EV brighter, while top gain will be 1EV lower.

    diagram-1.jpg

    diagram-2.jpg

    Because of significantly larger pixel size in comparison to 5831R-ZE – 2.0um vs 1.62um - we should expect a better image quality from this camera. The recorded material will be also more fluent, because it supports 25fps (also in WDR mode) while 5831R-ZE supports only 15fps (14fps in WDR mode). I wasn’t expecting any difference during daylight but it turned out that the image is slightly sharper. It’s actually sharp from edge to edge, while with 5831 I have a problem with out-of-focus areas of the image at certain zoom positions. I’ve mentioned this in its review. To compare both cams I’ve set 5831 zoom to match FOV of 1831. Unfortunately at this position there is blurriness in top-right corner. So when comparing both cams, please do not take this part of the image as representative of camera resolution.



    Studying carefully the image I’ve noticed that despite it is very sharp, there is some kind of softness creeping in. It looks like the color (chroma) is slightly soft in comparison to objects outlines (luma). It resembles a leaking color out of object. This can be a 4:2:0 chroma subsampling issue here, but I’m only guessing. https://en.wikipedia.org/wiki/Chroma_subsampling This is mostly visible when motion appears. Object's edges become soft.

    It might sound strange, but the codec efficiency seems to be worse in comparison to 5831. In shadowed areas and on moving objects, blocky artifacts appear. Luckily Dahua has raised the top bitrate in 1831 to over 15mbit/s. At 25fps this is an equivalent for 9mbit/s @ 15fps in 5831. Before firmware update I had an option to set over 50mbit/s in custom mode. However at 30mbit/s I had lot of dropped frames even in live view. Over 40mbit/s there was no image, only black window. After rising the bitrate to 15000kbps the blocky shadows are much reduced but still present. Even 30mbit/s do not resolve the problem, while 5831R-ZE is almost free of such artifacts at 9mbit/s. See the below example.

    artefacts.jpg

    Checkout motorcycle detail:

    artefacts-2.jpg

    On 5831 turning smart codec on, makes a significant improvement over image. There is no pulsation of I-frames and overall quality is improved. On 1831 the difference is minor. I can still notice pulsation of I-frames and some blocks of compression on moving objects. Only after reducing resolution to 4MP while keeping 15mbit/s bitrate, the blocks of compression are mostly gone.

    One way to reduce compression blocks in dark areas of the image is to turn on WDR. This will brighten dark parts and give a codec a chance to encode in more efficient way. The WDR on 1831 works much different than on other cameras I’ve tested. It has only 2 modes through the scale. From 1% to 44% it affects the brightness curve. It brightens shadows and reduces highlights without introducing WDR artifacts. It is based on single exposure and I have to admit, it looks very good. From 45% to 100% it turns on true WDR. In this mode the camera favors shadows rather than highlights. Midtones become over-sharpened, shadows are boosted up so much, that there is virtually no place to hide. It much resembles, what in AXIS cameras is called “forensic WDR”.

    WDR.jpg

    From a security point of view is seems to be perfect. From esthetic side it is slightly distracting. I much prefer the first WDR mode. At night WDR works well. It does not darken the image as in case of 5831 but introduces noise in shadows. The best setting I found for night is to set WDR for 20% and bring gamma all the way down. This will bring shadows up, but keep noise at acceptable level. Here is an example with such settings for 1/25s exposure:

    20180911_213629.jpg

    There is also a downside of true WDR mode. Can you see my face in the video example? It’s completely flat in true WDR mode. Notice also the quality of a text on my t-shirt.

    I think, there is no need for commenting nighttime footage. The 1831’s image is better. The noise pattern is a pleasant grain instead of blocks and specks as in 5831. Even at 100% gain, the image is acceptable. Noise reduction should be more efficient on 1831 as it has more frames to stack in a given period of time, due to higher framerate.

    When opposed to 4231EM-AS the 1831 gives similar if not slightly better results at night.



    Noise reduction seems more aggressive on 1831. Even during daytime - see the stacking artifacts on the car:
    vlcsnap-2018-09-14-23h19m09s230.png

    It seems to be most flare resistant of all cams I own. See the WDR comparison between 1831 and 5831. The sun makes 5831's image washed out in corners. This might be due to the fact 1831 is a fixed lens camera with less glass through the lens. The varifocal version might introduce more flares.

    At daytime the 1831 is a definite winner. The only drawback is its compression. 4231EM-AS has a very clean, blockless image, while 1831 despite the 15000kbps and smart codec is still introducing lot of compression artifacts. Especially when objects are moving fast.

    The scaling algorithm of the image has been vastly improved on 1831. Now it works exactly how it should. 2655x1520 / 1920x1080 / 1280x720 resolutions are full of detail and introduce less noise. FullHD image at night looks very good.
    fullHD-comparison.jpg

    Overall, in terms of image quality, this camera is my favorite. Not only you get a sharp, clear image at daytime, but also a decent image quality at night, even in color. There are almost no flares at night. 25fps gives you a pleasant, fluent video both for WDR and nonWDR mode. Scaling to lower resolutions works like charm now. You also get higher than on other cams bitrate option. WDR curves correction boosts the image without introducing any artifacts.

    From the drawback list I can point out a less efficient compression algorithm, which introduces blocks of compression on moving objects. The smart codec slightly improves this situation, but 5831 seems more efficient here. There is a problem with resetting date stamp. Even after firmware update. This model has no SD card slot. For me that's the worst thing, as I have no NVR and rely only on SD cards. Its varifocal brother has a SD slot.
     
    Last edited: Sep 14, 2018
  2. bp2008

    bp2008 Staff Member

    Joined:
    Mar 10, 2014
    Messages:
    8,745
    Likes Received:
    5,804
    Excellent work, as always. You really go the extra mile to test and compare these things. Thank you so much for doing this!
     
    mat200 likes this.
  3. LoLight

    LoLight Getting the hang of it

    Joined:
    Jul 9, 2018
    Messages:
    25
    Likes Received:
    32
    Location:
    WA
    Thank you Wopi82!
    Your super-comprehensive reviews are exceptionally done especially from a technical point of view.
     
    mat200 likes this.
  4. eman

    eman n3wb

    Joined:
    Aug 21, 2017
    Messages:
    14
    Likes Received:
    2
    Thanks. Excellent review as usual. If the sensor is the Sony IMX334 it is Starvis.
     
    mat200 likes this.
  5. flynreelow

    flynreelow Pulling my weight

    Joined:
    Dec 12, 2016
    Messages:
    322
    Likes Received:
    106
    these reviews are truly phenomenal.. keep them up...
     
    mikeTRON and mat200 like this.
  6. bug99

    bug99 Pulling my weight

    Joined:
    Dec 27, 2016
    Messages:
    328
    Likes Received:
    108
    Excellent review! Thanks. One small note. the data sheet calls out 4mm as having 102 deg FOV, not 87 deg as the smaller sensors have. they claim that 6mm jumps all the way down to 64 deg.
     
    mat200 likes this.
  7. wopi82

    wopi82 Getting comfortable

    Joined:
    Feb 9, 2017
    Messages:
    84
    Likes Received:
    386
    Location:
    Poland
    Untitled-1.jpg
     
    mat200 likes this.
  8. bug99

    bug99 Pulling my weight

    Joined:
    Dec 27, 2016
    Messages:
    328
    Likes Received:
    108
  9. wopi82

    wopi82 Getting comfortable

    Joined:
    Feb 9, 2017
    Messages:
    84
    Likes Received:
    386
    Location:
    Poland
    Indeed, something is wrong in these specs. A 4mm lens gives identical FOV as my 4231EM-AS with 3.6mm lens.
     
    mat200 likes this.
  10. Burton_Flooring

    Burton_Flooring Pulling my weight

    Joined:
    Apr 20, 2018
    Messages:
    125
    Likes Received:
    157
    @wopi82 Awesome review as always!

    THANK YOU, THANK YOU, THANK YOU! For the tip about the WDR changing modes above 44%, another thing that gives you that same detail "pop" during the day is HLC and cranking the Gamma UP during DAYTIME. (I had to do that after trying out your WDR discovery, you're right, above 44% makes the aesthetics of the image go down the tube in favor of preserving all details).

    Just remember that none of the Dahua cameras hold their Customized bitrate when power cycled. In fact the 1831E drops to 8192kbps and won't let you go back to 15104kbps, the only option it gives you is 13568kbps (after a reboot w/ a customized bitrate). You have to change to 8192kbps, reboot, then set the customized bitrate again, in order to get the 15104kbps option back. But, like I said, the other Dahua cameras do this too.

    @EMPIRETECANDY Why would Dahua let you set the bitrate to a value higher than it's rated value, but then not make the setting stick upon a reboot? That's the only really odd thing about their firmware (along w/ defaulting to SmartIR over manual settings as well).

    I will try your recommendation on the Gamma being dropped and the WDR on max 20. (This thing has crazy WDR compared to 4231EM-AS). When I changed locations and mounted the camera I had a whole other ball game to play w/ settings, lol.
     
  11. EMPIRETECANDY

    EMPIRETECANDY IPCT Vendor

    Joined:
    Nov 8, 2016
    Messages:
    2,757
    Likes Received:
    5,117
    Location:
    HONGKONG
    Very nice, thank you!
     
    mat200 likes this.
  12. vidmo

    vidmo n3wb

    Joined:
    Jun 23, 2014
    Messages:
    19
    Likes Received:
    12
    Thanks for the review @wopi82. One small constructive criticism though when doing your review, can you make sure that the primary camera you're reviewing stays on either the left or right in both the stills and videos? I kept having to look at the headers to determine which camera was which since they vary a bit.

    As @Burton_Flooring has suggested I'm curious to see if the firmware improves.
     
    mat200 and CCTVCam like this.
  13. M1CHA3L

    M1CHA3L n3wb

    Joined:
    Nov 21, 2016
    Messages:
    24
    Likes Received:
    1
    mat200 likes this.
  14. vidmo

    vidmo n3wb

    Joined:
    Jun 23, 2014
    Messages:
    19
    Likes Received:
    12
    @wopi82 Just to clarify, does this mean you think that the 102 or the 87 FOV is the correct number?
     
  15. wopi82

    wopi82 Getting comfortable

    Joined:
    Feb 9, 2017
    Messages:
    84
    Likes Received:
    386
    Location:
    Poland
    @vidmo I know, I messed up with sides :) But it was too late to correct it, when I realized that. 87 degrees is correct. It has identical FOV as 4231EM-AS with 3.6mm lens.
    @M1CHA3L Varifocal version has slightly darker lens at wide side - f1.9 (1831E has f1.6 lens) which means it won't be as sensitive as 1831E. But this difference is small. Of course it will get even darker when zoomed. It might also be more prone to flares.
    @Burton_Flooring I'm glad you like the review and the tip :) Every Dahua camera I own and tested, has this "jump" between 44% and 45% WDR.
     
    mat200 likes this.
  16. trevorf

    trevorf n3wb

    Joined:
    Sep 18, 2018
    Messages:
    14
    Likes Received:
    8
    Location:
    Brisbane
  17. redfive

    redfive Pulling my weight

    Joined:
    Apr 13, 2016
    Messages:
    268
    Likes Received:
    112
    @trevorf If the IPC-HFW2831T-ZS is installed outdoor, can you post a 4k snapshot, taken during the daytime ?
     
  18. trevorf

    trevorf n3wb

    Joined:
    Sep 18, 2018
    Messages:
    14
    Likes Received:
    8
    Location:
    Brisbane
  19. trevorf

    trevorf n3wb

    Joined:
    Sep 18, 2018
    Messages:
    14
    Likes Received:
    8
    Location:
    Brisbane
    In full sun.
    192.168.1.171_IPC_20180929103457314@1.jpg
     
    Last edited: Sep 29, 2018
    redfive likes this.
  20. trevorf

    trevorf n3wb

    Joined:
    Sep 18, 2018
    Messages:
    14
    Likes Received:
    8
    Location:
    Brisbane
    In case the the screenshot is dependent on the bitrate, I took this whilst at 26144 - the shots above taken at 6144.

    Cloudy today, sorry.
    192.168.1.171_IPC_main_20180930070853_@1.jpg
     
    Last edited: Sep 29, 2018
    redfive likes this.