Do we have IR illuminated/very dark images of movement comparison between Sony IMX415 and the 4M 1/1.8 sensor used in WizMind-S 5442-S3?

[luk8899]

As in the subject, I've been searching and browsing and I'm having hard time locating any comparison of images in Black & White mode with IR illumination and movement (or even without movement) of any mid range consumer camera that can do 30fps @ 4K and h265 (so we exclude ancient ones) with Sony's IMX415 and the forum's recommendation for pitch black + IR conditions WizMind-S 5442-S3?

I'm trying to figure out the magnitude of the difference and when is it most visible.

 

[The Automation Guy]

First, I think there are plenty of examples on this forum if you look closely. I don't have any direct links handy this second, but hopefully someone will post some examples.

Second, you mentioned "30 FPS", but that is a worthless spec most of the time. Keep in mind that FPS is not the same thing as shutter speed. So the fact that you run 30 FPS does not mean there is any difference in quality from someone running 15 FPS. In other words, 30 fps does not mean 1/30th of a second shutter speed, nor does 15 fps mean a 1/15th shutter speed. The only difference between 30 fps and 15fps is the fact that you are just recording twice the amount of data and therefore using twice the harddrive space at 30 fps. If you are recording a highway where cars are passing by at 70 mph, then a 30 fps(or faster) recording might be appropriate. However in 99% of the home use cases, there is no reason to record at 30 fps because the subjects we are capturing are traveling at slower speeds and therefore there is nothing lost by only recording 15 frames per second vs 30. It all depends on the purpose of the camera. For example, a license plate camera capturing vehicles on road with a 45 mph speed limit will need a much higher fps rate than your "driveway" camera is going to need. Your driveway camera can easily get away with a 10-12 fps capture rate and still capture more than enough information, while a camera at your front door can record even fewer fps and still be effective.

Third, as far as which cameras provide the best low/no light performance, it really is just a physics problem. You can look up the sensor size (at least reputable companies but the sensor size in the specs. There are unreputable companies that still try to hide this information) and resolution of a camera. Divide the area of the sensor by the number of pixels (ie the actual resolution of the camera) and that give you the area of each individual pixel. The larger the pixel, the more light it is able to absorb in any given period of time. Cameras with the largest individual pixel size will generally give you the best low/no light performance.

Far too often people think higher resolution means better cameras. Normally it is the opposite. Most consumer grade 4k cameras have relatively small sensors. This, combined with the high resolution/number of pixels ensures poor low/no light performance. A 1080p camera using the same sensor will perform much better in low/no light than the 4k camera will.

 

[luk8899]

First, I think there are plenty of examples on this forum if you look closely. I don't have any direct links handy this second, but hopefully someone will post some examples.

I'm not going to be surprised if I post the first ones once my cameras arrive

Second, you mentioned "30 FPS", but that is a worthless spec most of the time. Keep in mind that FPS is not the same thing as shutter speed. So the fact that you run 30 FPS does not mean there is any difference in quality from someone running 15 FPS. In other words, 30 fps does not mean 1/30th of a second shutter speed, nor does 15 fps mean a 1/15th shutter speed. The only difference between 30 fps and 15fps is the fact that you are just recording twice the amount of data and therefore using twice the harddrive space at 30 fps. If you are recording a highway where cars are passing by at 70 mph, then a 30 fps(or faster) recording might be appropriate. However in 99% of the home use cases, there is no reason to record at 30 fps because the subjects we are capturing are traveling at slower speeds and therefore there is nothing lost by only recording 15 frames per second vs 30. It all depends on the purpose of the camera. For example, a license plate camera capturing vehicles on road with a 45 mph speed limit will need a much higher fps rate than your "driveway" camera is going to need. Your driveway camera can easily get away with a 10-12 fps capture rate and still capture more than enough information, while a camera at your front door can record even fewer fps and still be effective.

Oh no, I'm not running at 30fps. But by the camera being 30fps capable at 4K means it has a modern CPU. This often means it is not an older model, therefore the software is often a bit better(if all other specs are roughly equal). So for example if I'm buying a camera and the only difference is FPS I'll go for the higher fps although I'll not be running it at that speed.

Third, as far as which cameras provide the best low/no light performance, it really is just a physics problem. You can look up the sensor size (at least reputable companies but the sensor size in the specs. There are unreputable companies that still try to hide this information) and resolution of a camera. Divide the area of the sensor by the number of pixels (ie the actual resolution of the camera) and that give you the area of each individual pixel. The larger the pixel, the more light it is able to absorb in any given period of time. Cameras with the largest individual pixel size will generally give you the best low/no light performance.

Far too often people think higher resolution means better cameras. Normally it is the opposite. Most consumer grade 4k cameras have relatively small sensors. This, combined with the high resolution/number of pixels ensures poor low/no light performance. A 1080p camera using the same sensor will perform much better in low/no light than the 4k camera will.

It is all true, but there is some nuance. The pixel size on its own can be sometimes misleading. You have to take into account who made that sensor and when(did they have access to latest tech). And the software quality. The software plays a huge role in image processing and can break a nice sensor or elevate a mediocre one (only to certain extent of course)

For example some years ago I bought some XM branded cameras from Aliexpress with 1/1.8 4K sensors (I believe they have SmartSens SC850SL). The pixel size is larger than my other cameras with imx415. Even the date of first manufacture suggests SmartSens would be better with it coming out in 2021 vs Sony in 2019. But they are not(in the dark with IR - they are better in low light if you want color). All my imx415 cameras produce better IR pictures. Why? Well, I can think of two reasons. First SmartSens does not publish their testing methodology for their sensitivity numbers(that are claimed 40% better than imx415 for visible light). Why not? We can only guess. Perhaps the sensor is indeed better in visible light but not in IR? Or maybe the software in that camera is letting it down in low light conditions... who knows. (I did try locking the exposure it made no difference)

So my point is yes, sensor size is important, but not the only thing to consider.