Sensor size and why dont they use bigger ones.

[Kentg]

After reading a few links about sensor size and pixel rates and how they affect quality and light levels its some very enlightening eaceful: information.
I do not know how expensive it is to go to the next bigger size sensor from 1/3 to 1/1.7 to 2/3 and so on. Looking at the hikvision website for comparison their 2 line and a lot of the 4 line cams are all still1/3 and untill you get to the 6mp and 4K cams do they go to a 1/1.7 or 1/1.8 sensor.
The difference from a 1/3 to 1/1.7 is quite a lot it seems from this chart I found in Wiki in an entry "image sensor format".

https://en.wikipedia.org/wiki/File:Digital_camera_sensor_area.svg

It would seem that the 1/3 size is the most common now in home system cams and the move up to the larger sensors is pretty expensive I would figure. Also with that much more sensor area how does that affect the processor needs on a cpu ?

Anyone here care to enlighten me on the issue of sensor size, cost and whats in the pipeline for equipment in the home user price range?

 

[CamFan]

Pixel size has nothing to do with processing power, pixel count does. Bigger pixels can be exposed to more photons, thus can be better in low light.

Bigger sensors need bigger lenses and will have shorter depth of field. For security, you want long depth of field so everything is in focus. Think pinhole camera.

So it's a trade off with security cameras. Get the sensor small for depth of field, lower cost sensors, smaller and lower cost lenses.

Go too small and the laws of physics will have a say with photons hitting the sensor, affecting low light performance.

 

[Kentg]

Pixel size has nothing to do with processing power, pixel count does. Bigger pixels can be exposed to more photons, thus can be better in low light.

Bigger sensors need bigger lenses and will have shorter depth of field. For security, you want long depth of field so everything is in focus. Think pinhole camera.

So it's a trade off with security cameras. Get the sensor small for depth of field, lower cost sensors, smaller and lower cost lenses.

Go too small and the laws of physics will have a say with photons hitting the sensor, affecting low light performance.

A bigger lens can still have the same magnafication cant it as a smaller lens? A larger sensor with 4mp than a smaller lens with 4mp would obviously have better low light properties naturally.

Can you point me to some info explaining the sensor size and how it along with the larger lens affects field of depth?

 

[Kentg]

I found this among others using google "cmos sensor size and depth of field"
https://photographylife.com/sensor-size-perspective-and-depth-of-field

It is confusing but it seems that a sensor of say 2/3 would be fine as long as the lens was configured for this use. Im would think the guys who design and make these cameras are trying to increase the sensor size for the image quality and there must be other reasons as well they are the size they are.
Id like to see some info from these folks on why they dont from their own perspective.

Im not saying your not right but it seems like something that could be overcome.

 

[wxman]

Bigger sensors need bigger lenses and will have shorter depth of field. For security, you want long depth of field so everything is in focus. Think pinhole camera.

But wouldn't a lens with smaller aperture easily solve this?

 

[CamFan]

I think 1/2" will be as big as they go. S Mount/M12 lenses seem to max out at 1/2". Going to 2/3" and larger will be a c-mount lens. It has to do with image circle coming out the back of the lens being able to cover the entire sensor.

It's all about economics of scale. The small sensors they use now are driven by cell phones. Smaller sensors means your optics must be manufactured to tighter tolerances. The M12 lenses are amazing in terms of image quality, surpassing c Mount lenses long ago. They didn't have to make a 5MP c-mount lens. Not saying they can't, its just that they have huge investment in M12 lenses to get the high image quality for really low cost. They would need a really compelling reason to abandon M12.

I think we will see more 1/2" sensors being used in security. But not any bigger.

 

[fenderman]

Hikvision already uses 1/1.8 lenses in their darkfighter series cameras http://www.bhphotovideo.com/c/product/1162153-REG/hikvision_darkfighter_series_ds_2cd4526fwd_izh_2mp.html
The pricing keeps getting better...that 603 for a legitimate (authorized seller) darkfigher with motorized varifocal is cheap....

 

[AlpineWatch]

FWIW, coming from a place that manufactures things on wafers, the smaller the die size the more you can fit on a wafer, the lower the per part price. In the camera sensor world, smaller pixels - smaller die size.

The others above have already pointed out that the optics get cheaper as the lenses can get smaller. So the chip manufacturers really don't want to keep making the expensive parts (large die, less per wafer) and the camera guys like that they can used cheaper optics (smaller elements).

 

[Kentg]

FWIW, coming from a place that manufactures things on wafers, the smaller the die size the more you can fit on a wafer, the lower the per part price. In the camera sensor world, smaller pixels - smaller die size.

The others above have already pointed out that the optics get cheaper as the lenses can get smaller. So the chip manufacturers really don't want to keep making the expensive parts (large die, less per wafer) and the camera guys like that they can used cheaper optics (smaller elements).

Yeah but the camera market seems to be growing and the mfg who makes a better one be it by sensor size or some manipulation of smaller ones will make money and that's what drives the market.
Its why users like me made the jump to IP from analog and the next step depends on low light/night image quality for a reasonable price.

The first mfg to offer a cam that does quality night images/video at a reasonable price will make millions and I cant believe that they wont explore avenues.

 

[fenderman]

Yeah but the camera market seems to be growing and the mfg who makes a better one be it by sensor size or some manipulation of smaller ones will make money and that's what drives the market.
Its why users like me made the jump to IP from analog and the next step depends on low light/night image quality for a reasonable price.

The first mfg to offer a cam that does quality night images/video at a reasonable price will make millions and I cant believe that they wont explore avenues.

The large manufactures know the market better than any of us..they are not stupid, they will do everything they need to do to maximize profits...

 

[Ookie]

With respect, just until someone else comes along and eats their lunch.

Others (like Mobotix) are being forced to adapt or go the way of the dodo. As soon as one manufacturer makes a big jump in low light performance and/or image quality, the others will be forced to follow. Eventually someone will do it.

What Hikvision has been doing with quality is likely behind the massive drop in MSRP pricing by Mobotix.

That being said, I respectfully disagree that the most important next step is an increase in low light/image quality. If the current Hikvision models are having trouble with the lighting conditions on your premises, I respectfully suggest lighting is the bigger problem with your security plan.

Having a chance at deterrence is far more important that getting video of a crime that was allowed to happen. Anyone who thinks otherwise has never been to court in support of footage obtained by one of their camera installs. After EIGHT court appearances in a single case involving a stolen cell phone (the perp's attorney kept requesting adjournments in the hope that we did not show), I learned this the hard way.

Where I think that the industry needs to go is more accurate motion detection. As Carl (NCC--RIP) capably mentioned, Mobotix is now WAY behind on image quality. They were the best in the business five years ago. Their tech has basically stayed in the same place with regard to image quality, while others have progressed. That being said, the accuracy of MxActivity sensor (their motion detection system) is MINDBLOWING.

This video is representative of my personal experience. These cameras only record (and notify you) when a person or a vehicle is involved.

It's actually more accurate than PIR sensors, without the need for additional expense (both for the sensors, DC power, and wiring).

Too bad the image quality of Mobotix is outclassed by $100 Hik 2032s.

The real question is who will develop the complete system first. Will the Chinese figure out motion detection, or will the Germans upgrade image quality?

 

[alastairstevenson]

As soon as one manufacturer makes a big jump in low light performance and/or image quality, the others will be forced to follow.

Worth remembering too is that neither Hikvision nor Dahua or almost any others of the well-known names in surveillance manufacture the image sensor.
These are developed and sold by just a few specialist manufacturers such as Aptina, Omnivision, Sony, Panasonic etc and integrated into products by Hikvsion Dahua etc.
The boundaries of performance are largely determined by the current state of the art of semiconductor manufacturing, with the occasional structural innovation giving one manufacturer an edge over their competitors for a short while.

Discussions on this forum mostly revolve around what are really the low-end of surveillance cameras - it's easy to not realise there are amazing analytics capabilities in the devices further up the market where they are no longer just a video source but an intelligence source. Heat map, people counting, facial recognition, dropped object detection - something that only a few years back was a high-end high-cost defence capability, now a low-cost commercial commodity item.
We are already seeing some of those capabilities trickle down the product hierarchy, and it will be fascinating to see how that development continues. It's facilitated by software algorithms and processing power after all.