Wide Dynamic Range Myth

DashCamMan

DashCamMan

Administrator
Joined
Jun 30, 2012
Messages
2,150
Reaction score
1,352
Country
Canada
I know that niko has recently created some posts regarding how many cameras who claim to have Wide Dynamic Range, actually do not have it. This is specific to cameras with the Novatek NT96650 Processor and Aptina AR0330 CMOS sensor which have become very popular this year. I have done some research on this topic as well and he is correct. I have updated the FAQs on these cameras with the following info:

Dynamic Range is measured in decibels (dB) and most cameras have a dynamic range of 60 dB to 80 dB. The human eye has a dynamic range of approximately 100 dB. Anything higher than 100 dB is often referred to as "Wide Dynamic Range" (to be confirmed).

While the Novatek NT96650 Processor supports Wide Dynamic Range, the CMOS sensor used, the Aptina AR0330, has a dynamic range of 72.4 dB (see datasheet). Hence from a pure technical perspective, this camera is not providing true Wide Dynamic Range. Note that a high dynamic range number (>100 dB) does not guarantee great night performance if other aspects of the camera are poor (processor, lens, firmware).

There is a lot of technical information behind this which I'm sure niko or others can elaborate on. I am not an expert in this area by any means so I am open to suggestions on this topic. Also note that even if a CMOS sensor claims a high dynamic range number on a spec sheet, the actual measured number might be less and apparently there are standard ways to measure this.
 
Last edited:
Even in the professional video cameras, it's about a 10 to 13 F-Stop range, with 6dB per F-Stop, that comes to about 10 F-Stops x 6dB = 60 dB ---> 13 F-Stops x 6dB = 78dB. That's the current limit of the imagery technology. In contrast to the older image sensor chips we do have a much wider dynamic range. In a retaliative sense, we do have a "wide dynamic range" compared to the older technology. So marketing can stretch the meaning of the term WDR as a selling point.

I've read about the idea of making image sensor chips with both large and small sensor elements. The large areas would read the dark areas and the small would read the bright highlights.

For daytime shots, the current technology seems good enough, but it might help with night shots where the dynamic range is much greater. If I get a chance, I could simulate this with my Nikon Camera. It take 2 different exposures and puts them together to obtain HDR (High Dynamic Range).
 
Thanks, DashCamMan.

I had all this info, but did not know how correctly to put it into words, due to my not that good technical English. Good that you took initiative into and also did some research.
In simple words to explain, these days commonly used "WDR" you see in the specs of mid price-range dash cameras ( mostly NT966500 + AR0330 based ) is achieved by FIRMWARE ENHANCEMENT, where by default manufacturer uses higher ( + ) EV value. This way we can see video sensitivity "improvement" at night, but of course it's not fully true WDR and should be called HDR( high dynamic range ) as jokiin and BobDiaz mentioned, or fair would be even: Firmware Enhanced WDR.

It's just a marketing trick about "WDR myth" ( as tittle says ), which mostly started from DOD LS300W, where they started advertising "magics" of WDR. Of course to stay in business, all other manufacturers also started using this trick, but without informing people that this is just a trick. I can compare this "mania" to so called "FullHD"-mania where they use 720p to 1080p interpolation, which we had 1-2 years ago, maybe still have a little bit where people are not educated. It's become like a trend these days, where if dash camera do not have word WDR in the specs, - then most people think it's a bad camera, - don't buy it, move on and find other dash camera with words WDR in the specs. - It's not true ! Thats why I tried to start debates ( and thanks DashCamMan picked up ), to educate people. By testing dozens of high quality Korean dash cams, where there were no "wdr" word mentioned in the specs, but still night video quality was quite good ( comparable with ls300w type dash cameras ), maybe a little bit more background noise due to low bit-rate, but still night video quality sensitivity was comparable to so called fake-wdr ones ( example: Itronics ITB-100HD, Livue-100LB, and so on ).

Best way to determine if dash camera can have firmware enchanced WDR or true WDR, is to look into CMOS sensor specs. But of course it will not guarantee full potenital of specs used. Aloso very important is how DSP, CPU and firmware is tuned-up to get most of the CMOS sensor. However it is more possible to get true WDR if you see from CMOS specs its wide dynamic range is 100db or more.

Here are some samples of different CMOS sensors specs how to read WDR.

Very popular OV2710 ( mostly used with A2S60 CPU ), it has 69db.
High End OmniVision OV10633 can produce up to 115Db wide dynamic range.
Please note that SNR on both stays 39Db, dispite huge class differences.

OV CMOS.jpg

Most popular CMOS used on so called "wdr" dash cameras is AR0330. Strangely that there are inconsistency in their specs.
If you download Flyer, it shows 72.4Db but Specs sheet shows 69.5Db. , ???
For such a big name as Aptina this kind of inconsistency should not be allowed. Maybe they change value to correct ones, once ( hopefully ) they read my post.

AR0330.jpg

Higher Class AR0331 has up to 100db wide dynamic range.


ar0331.jpg
 
Last edited:
Anyway, the dashboard cameras have very good performance, both, daylight, in strong backlight and night time, in low light. I'm not sure if only the sensor, in single exposures for each frame is used to get WDR videos or a technology similar to that used in Magic Lantern firmware of Canon dSLRs, when for each HDR frame are taken two frames, one overexposed, one underexposed, then they are combined in a HDR frame... In that way the technology could surpass the sensor specifications.
Sorry for my poor english... but I hope you can understand.


Look at this video, with strong backlight:
 
BobDiaz said:
I've read about the idea of making image sensor chips with both large and small sensor elements. The large areas would read the dark areas and the small would read the bright highlights.

For daytime shots, the current technology seems good enough, but it might help with night shots where the dynamic range is much greater. If I get a chance, I could simulate this with my Nikon Camera. It take 2 different exposures and puts them together to obtain HDR (High Dynamic Range).
Click to expand...

Not bad idea, but there will be a lot of challanges need to be faced.
Lets take best closest to 1080p resolution cmos size of 1/3" is 2.1Mpix ( 1080 x 1920 = 2.073.600 pix ). Most popular well known to this figure is Sony Exmor IMX122 with 2.4Mpix ( 1/2.8" ). Each pixel size is 2.8 micron.
If somehow adding to same cmos area of 1/2.8" second pixel ( doubleing pixels, one for large exposure, second pixel for short exposure ) and still want to have true 1080p, then each pixel size will be dramatically decreased resulting in "absorbing / catching " less light.
The only way to keep each pixel same size and doubling amount of pixels ( 2.1mpix for long-, 2.1mpix for short exposure ) is to have bigger cmos size, lets say instead of 1/3" to use 1/2" cmos, but this will result in camera size and of course such technology price would be out or reach for consumers.
 
No need 1 pair of pixels (ie 2 pixels) for getting 2 exposures. It's enough the electronic shutter, who can do it. I explained in my previous message. Like I said, these cameras don't have mechanical shutter, as dSLRs, but so-called electronic shutter, which is easier to work with.
The problem would be, in this case, that will require more powerfull processors to combine those different exposed frames, for the final result.
 
Last edited:
ursamajor said:
No need 1 pair of pixels (ie 2 pixels) for getting 2 exposures. It's enough the electronic shutter, who can do it. I explained in my previous message. Like I said, these cameras don't have mechanical shutter, as dSLRs, but so-called electronic shutter, which is easier to work with.
The problem would be, in this case, that will require more powerfull processors to combine those different exposed frames, for the final result.
Click to expand...

Ok. I see your meaning comparing as its done on DSLRs.
 
Usually there is a lot of marketing behind these "inventions".
 
niko said:
Ok. I see your meaning comparing as its done on DSLRs.
Click to expand...
It wasn't my intention to compare these toys with a dSLRs, but to show that's much, much easier to make electronically a sensor who can capture HDR, than to invent a new more, more sensitive one.
 
Last edited:
niko said:
Yep, this might be the future of dash cams, but I am afraid we have to wait many years until that image sensor price drops as low, so can sell mass production dash cams to consumers at 100-150usd.
Click to expand...
Indeed, for example in astronomy imaging there are already sensitive sensors, but their prices are... prohibitive.
 
jokiin said:
Many of them are all spin and no substance, I'm always suspicious of anyone that doesn't want to tell you what they use, or worse, lies about what they use
Click to expand...

;) ? @ dod ls300w "A8" ( nt96650 ) CPU and their " real WDR technology", which lead to this thread here ?! ;)

Maybe can also add here TF pictures of Mobius CPU where he blurred for some reasons NT96650 in his review?
 
Mobius I can understand, small developer wanting to maintain some competitive advantage while they're working on their product

Large marketing company pretending to make their own chipset, very poor form
 
jokiin said:
Mobius I can understand, small developer wanting to maintain some competitive advantage while they're working on their product

Large marketing company pretending to make their own chipset, very poor form
Click to expand...

yes, I have mentioned that in mobius thread, totally understand that. Many people dont understand why its so and making wrong conclusions, so there should be a note " we dont want to reveille hardware at this stage for obviouse reasons of competition ".
 
As i undestood it is only a matter of using graphene some where in the normal process of Building a imaging chip, so it shouldent be a major retooling to make them.

Dont know why this invention dident come from Denmark as we are in the frontline when it come to cameras for Space / sattelite use.
 
You must log in or register to reply here.
Thread starter Similar threads Forum Replies Date
T Recommendations: Truck Wide View Rear Window Dash Cam Discussion 4
D 140* FOV on A129 Duo rear-cam not wide enough Dash Cam Discussion 6
C Dual cameras for wider FOV? Dash Cam Discussion 13
D The downside of a wide angle lens Dash Cam Discussion 65
gapi WDR (Wide Dynamic Range) When to Enable or Disable? Dash Cam Discussion 2
Similar threads
Recommendations: Truck Wide View Rear Window
140* FOV on A129 Duo rear-cam not wide enough
Dual cameras for wider FOV?
The downside of a wide angle lens
WDR (Wide Dynamic Range) When to Enable or Disable?
Back
Top