(Hikvision) camera physics/electronics: further advice sought

[ElectronicBattle]

Hello all

Is an older generation Hikvision camera with a larger sensor better at the lowest light level margins that a newer generation with a smaller sensor? I think the answer is "yes".

I have been trying (and failing) to buy a pair of Hikvision DS-2CD2T87G2-L or -LSU/SL "ColorVu" cameras which are now "end of life" and are not being manufactured. In the UK, although several reputable online sellers claim to have limited stock, what I was sold and invoiced for was a G2-L but what I was actually supplied was the G2H model. I pointed out the error and was told "but Hikvision recommended we supply the G2H s a replacement for the G2-L".

I'm getting a refund for the following reasons:

older G2-L is a 1/1.2" 8 Mp sensor with an F1.0 lens and claims a low light sensitivity of 0.0005 Lux @ (F1.0, AGC ON)
newer G2H is a 1/1.8" 8 Mp sensor with an F1.0 lens and also claims a low light sensitivity of 0.0005 Lux @ (F1.0, AGC ON)

However, the light gathering of the larger (1/1.2") G2-L sensor is 2.25x greater than the smaller G2H and hence, in true low light conditions the G2-L should outperform the G2H based on the photon capture and conversion efficiency and hence the signal to noise ratio at the sensor face and then presented to the image processor.

I accept that the AI image enhancement algorithms (de-noising, stacking) of the newer camera model might perform well, down to some arbitrary threshold and might be "better" in medium light conditions than the G2-L, but my calculations show that the G2-L will provide a better signal to noise ratio in the lower-still light conditions which I want the camera for.

The various tests or reviews of the newer ColorVu cameras show them being used in urban street scenes where there is a degree of light contribution from artificial street lights. For my one camera application this is also the case but for my other, the only light will be starlight or moonlight. I know that when there's none, I won't see anything with the G2-L because there's no IR and I would have to enable to "gentle" onboard white LEDs. The newer G2H does have a IR floodlight as well as the same gentle white LED illumination.

Therefore I say that the G2H is not a replacement for the G2-L, but - has anyone got any actual real-world experience which either supports what I am saying or contradicts me?

[NickM-27]

They both have the same 0.0005 Lux so they are both claiming the same light sensitivity. There are definitely some newer technologies that are making results better with smaller sensor sizes, so there is no reason to not believe them. But if the older one is more cost effective too then it makes more sense.

[stowaway-ovr]

It might also be worth doing a search / posting your question here - a lot of technical views on Hik cameras.

https://ipcamtalk.com/forums/hikvision.5/

I have had the same issue with the 86G2 -> 87G2 -> 87G2H models and the varying approaches to IR illumination. I would recommend trying to find a decent review where someone has mounted the two cameras side by side, and looked at different images under varying light conditions (including "none").

I've not read it all back but some good quality input: https://www.use-ip.co.uk/forum/threads/pro-series-with-smart-hybrid-light-review-and-first-impressions.10173/

[ElectronicBattle]

Ahh yes ...ipcamtalk... I have tried to join their forum several times during the last couple of years but I can't get the forum or the admins to respond to me! I tried two different email addresses (one of which is gmail): I just get told "you're pending" and it persists like that for years. Trying to contact their admin results in my message being cancelled because I am told I haven't completed a non-existent "prove you're a human" puzzle - different browsers, no popup blockers, given up. I conclude they don't like my ISP's assigned IP block (but I can at least read their forum posts - thanks for the URL!).

But - I agree, it's a good place to read up. I'm confident that for wildlife observation, the overall sensitivity (lens apertures x sensor size / pixel count) is a major contributing factor to make the correct choice. I have an area which is 50m distant and with no artificial light it is very dark indeed (tree cover partially shadows the area too from moon/stars) which is not typical for the usual security camera scene processing.

More modern cameras (2024/2025/2026) seem to have settled on a 1/1.8" sensor and they use a lot of AI processing and, ultimately, longer exposure times: OK for street scenes but not OK for wildlife in the dark distance. I think the G2-L is better than the G2H and I haven't even looked at the G3 because they too settle on the smaller sensor (and Annke don't do a 4mm lens and Uniview don't seem to have quite what I want - I have looked at so many specs I can't remember my own name at the moment).

None of the options (including my original T87G2-L or 387G2-L choice) do dual-illumination - which means being able to choose "no illumination" for dark scenes (relying on ColorVu low light technology alone) , then switching to "soft white LED" to augment in very low/no light or being able to instead select IR for stealthy (wildlife friendly) scene lighting. The lighting will be important to make sure I can get faster shutter speeds and hence reduced ghosting and motion blur.

I have chatted to the people at use-IP pand they were very helpful and I will have a good read of the forum and that post in particular; thanks for pointing it out to me.

86G2 -> 87G2 -> 87G2H did you decide what to do? I know in my case I want the 8Mp versions for highest quality daytime observations and hence I want the 1/1.2" to make up for it during the night.

All quite complicated!