How does resolution affect field of view for cameras?

[TylerD321]

I am having a hard time wrapping my head around how exactly field of view works in terms of a camera.

 

From what makes sense to me, the field of view would be for the camera itself during a still image and then video would be taken from pieces of that.

 

For an 8 megapixel camera, the picture size will 3280x2464 pixels. Of course, the most common video resolutions are 1080p, 720p, or 480p. (1920x1080, 1280x720, 640x480)

 

I have attached a picture that describes how I imagine field of view works, is anyone more familiar on this topic that could shed some light for me? I have an arducam raspberry pi camera advertised at 175 degree field of view, the horizontal and vertical resolution is 96 dpi, and the bit depth is 24. I want to know what the field of view is of each video recording resolution.

 

Thank you!

[Electronics Wizardy]

This really depends on the camera, and Im not sure for the pi camera.

 

But most cameras will use pixel binning/line skipping to produce a low res video output with the full field of view of the camera. But really depends, so you have to read the specs to see what the exact model does.

[TylerD321]

2 minutes ago, Electronics Wizardy said:

This really depends on the camera, and Im not sure for the pi camera.

 

But most cameras will use pixel binning/line skipping to produce a low res video output with the full field of view of the camera. But really depends, so you have to read the specs to see what the exact model does.

The camera is Arducam IMX219, I can attach an Amazon link. I am very unfamiliar with exactly how cameras work!

Amazon.com: Arducam 8MP IMX219 175 Degree Ultra Wide Angle Raspberry Pi Camera Module, Compatible with Raspberry Pi 4 Model B, Pi 3/3B+, and Pi Zero 2W : Electronics

[An0maly_76]

In my experience (it's been awhile), lens size makes far more of a difference. I was using static lens cameras (no pan/zoom), 3.6mm was generally the widest viewing angle, but pretty much had to be in the air for best results .7-9mm is about what you want for a typical room, maybe 12mm. Anything larger and you're getting into tight shots from yards away with a very narrow viewing angle. 25mm is for license plates.

[PDifolco]

1 minute ago, An0maly_76 said:

In my experience (it's been awhile), lens size makes far more of a difference. I was using static lens cameras (no pan/zoom), 3.6mm was generally the widest viewing angle, but pretty much had to be in the air for best results .7-9mm is about what you want for a typical room, maybe 12mm. Anything larger and you're getting into tight shots from yards away with a very narrow viewing angle. 25mm is for license plates.

Yes, the image of OP with various res at same pixel size is confusing, it doesn't work like this, what happens is that the lens define the FoV, then the camera snapshot it at its resolution

For a given lens it's rather the same rectangle with more or less large pixels !

 

[tikker]

The sensor samples the focal plane of the camera. If you want to sample a larger field of view you'd want to sample more of the focal plane and thus want a physically larger sensor. If you increase the resolution of the sensor by adding more pixels, but keep the pixel size the same, then your sensor will naturally get bigger and thus offer a larger field of view. Conversely if you keep your sensor size the same and shrink the pixels when increasing resolution you'll sample the same field of view (the same part of the focal plane is sampled), but you'll be able to capture more detail. If you keep your sensor size the same and enlarge the pixels while decreasing resolution you'll sample the same field of view, but you'll capture less detail (but gain signal-to-noise ratio which can be useful in low-light situations, for example).

 

The datasheet for your camera here mentions pixel binning as the / a possible method for the lower resolution modes: https://www.arducam.com/product/arducam-imx219-auto-focus-camera-module-drop-in-replacement-for-raspberry-pi-v2-and-nvidia-jetson-nano-camera/

Quote

5-2 Pixel Binning Mode
Binning read-out can be used to obtain an image of lower resolution for full field of view. It has advantage on frame
rate than using digital scaling, and on signal-to-noise ratio than using sub-sampling. See Binning Capability Registers,
for detail of available configuration

 

[Fascinated Viewer]

On 7/27/2022 at 7:05 PM, TylerD321 said:

I am having a hard time wrapping my head around how exactly field of view works in terms of a camera.

 

From what makes sense to me, the field of view would be for the camera itself during a still image and then video would be taken from pieces of that.

 

For an 8 megapixel camera, the picture size will 3280x2464 pixels. Of course, the most common video resolutions are 1080p, 720p, or 480p. (1920x1080, 1280x720, 640x480)

 

I have attached a picture that describes how I imagine field of view works, is anyone more familiar on this topic that could shed some light for me? I have an arducam raspberry pi camera advertised at 175 degree field of view, the horizontal and vertical resolution is 96 dpi, and the bit depth is 24. I want to know what the field of view is of each video recording resolution.

 

Thank you!

This will be ample information dumb I am a uni 3rd-year student studying photography:
Your senor is a low-resolution senor; i.e. my cameras are 24 megapixels and 8, but I have used and worked with 110 megapixels (Phase one). Your details will be low, and they may appear low res as there are not enough pixels in the sensor. For example, if you zoom in on 110 megapixels, you'll see all the fine details in the image and crop in further, but if you use 8, you won't have enough pixels to make it appear as high res as you could. The sizes you're showing will be HD 1080p but not as good as they could be. It also depends on the bit rate. The lower the bit rate, the worse the image; basically, i.e. my camera is recording ins 1080p and will show me 1080p on Youtube, but the quility of it will be poor and will make you think that it is not 1080p, but it is just lousy 1080p.

 

It's not just the senor you need to think about. It's also the lens as, without it, your senor is just as good as someone with terrible eyesight. For example, if your vision is awful, you can't see without your glasses which are lenses that help you see (generalizing). FOV is a term that a of photographers use. Tbh most use mm or millimetres; for example, can I get an 18mm or 50mm, which describes how zoomed or wide angle the lens is. However, going into more detail, a seneor will determine how zoomed the lens is, for example, a crop seneor which is typically your cheaper beginner camera. It will be more zoomed in due to the sensor being smaller (but most manufacturers that make lenses for that camera will facotr this in and display the true mm on the lens unless that camera supports more than one type of lens mount), so you would need to consider a crop factor which can vary between manufacturers, i.e. sony is 1.50x. And a full-frame camera is 1x this critical in figuring out how zoomed your lens is, for example. I have a canon 77D and an 70 - 300MM lens. My camera has a crop factor of 1.6, so when I times 1.6 x 70mm, that tells me the "true" zoom of the lens (112mm), which is good to know for some photography apps)

some fancy lenses have a display that tells you the true crop factor.

If you can find out the MM of the lens, then you can find out how zoomed or wide angle the lens is as no one says 175 FOV bc that doesn't mean anything to most photographers, to be brutally honest with you.