Dahua WizColor 5x59-PRO and SmartLight 5x59-IL new series

[Techie007L]

But but but… 8 megagixels has to be better than only 4, I mean it’s twice the megagixels!

Because there are two axis involved, the total number of pixels has to square (multiply by four) to double the actual resolution. 4k (8MP) is only double the resolution of 1080p (2MP), a 100% increase. But because of the Bayer pixel color pattern, you need an 8MP sensor to have a true 1080p resolution video without algorithmic interpolation. 4MP (2k) sensors find themselves in a sweet spot that provides enough extra resolution to result in near perfect 1080p video, which is why video from a 2k camera looks so much sharper than from a native 1080p camera—even when viewing at 1080p (because some interpolation is better than no interpolation, but despite quadrupling the number of pixels, interpolation cannot double the actual resolution). Going from a 4MP (2k) sensor to 8MP (4k) is a mere 50% increase in resolution, identical to jumping from 720p to 1080p true resolution. If you want true 2k (2560x1440) resolution, you would need a 14.8MP sensor; and for true 4k (3840x2160) resolution, you would need a 33.2MP sensor. This is why going from 4MP to 8MP results in a negligible improvement in detail resolving power. And this is why many newer pro phones have 48MP sensors in them even though they only take 12MP photos.

 

[wittaj]

So your saying a 1/1.2 8mp vs a 1/1.8 4mp there is no difference identify distance? Both at 2.8mm.

Would not a larger sensor and higher mp have longer identify distance at the same focal length unless I'm missing something?

That has been the experiences of most of us here. I do not see a measurable improvement.

We still don't have a camera that mimics digital zoom we see in movies and TV.

As pointed out, the larger sensor also provides a wider field of view and thus means MP are spread over a bigger filed of view, which does nothing for identify at distance.

 

[raidflex]

That has been the experiences of most of us here. I do not see a measurable improvement.

We still don't have a camera that mimics digital zoom we see in movies and TV.

As pointed out, the larger sensor also provides a wider field of view and thus means MP are spread over a bigger filed of view, which does nothing for identify at distance.

My logic is if I take for example the color4k-t and 5442 S3 and zoom in on a snapshot of a person I should be able to make out more details in the image of the 4k-T. This is keeping everything else equal meaning same distance of subject and same lighting conditions.

Since the camera will be in the front, I need a wide angle of the area hence using a 2.8mm. But if I can get a little more details of a person walking down the path that's coming up to the doorway at a further distance, that would be a benefit to me. Like I said I'm not talking about 50 ft here. I'm talking around 20 ft.

 

[bigredfish]

Because there are two axis involved, the total number of pixels has to square (multiply by four) to double the actual resolution. 4k (8MP) is only double the resolution of 1080p (2MP), a 100% increase. But because of the Bayer pixel color pattern, you need an 8MP sensor to have a true 1080p resolution video without algorithmic interpolation. 4MP (2k) sensors find themselves in a sweet spot that provides enough extra resolution to result in near perfect 1080p video, which is why video from a 2k camera looks so much sharper than from a native 1080p camera—even when viewing at 1080p (because some interpolation is better than no interpolation, but despite quadrupling the number of pixels, interpolation cannot double the actual resolution). Going from a 4MP (2k) sensor to 8MP (4k) is a mere 50% increase in resolution, identical to jumping from 720p to 1080p true resolution. If you want true 2k (2560x1440) resolution, you would need a 14.8MP sensor; and for true 4k (3840x2160) resolution, you would need a 33.2MP sensor. This is why going from 4MP to 8MP results in a negligible improvement in detail resolving power. And this is why many newer pro phones have 48MP sensors in them even though they only take 12MP photos.

Yes

 

[Techie007L]

I’m not sure identify ranges you’re talking about are correct. 4MP 1/1.8 2.8 and 3.6mm have same linear pixel density(identify) as those same focal lengths focused on an 8MP 1/1.2” sensor. The 8MP just gets bigger FOV.

As pointed out, the larger sensor also provides a wider field of view and thus means MP are spread over a bigger filed of view, which does nothing for identify at distance.

This has not been my experience. The FOV figures for all these cameras are readily available, and the Color4K series cameras have a narrower/smaller FOV (so slightly more zoomed in) than the 2K cameras, which should help further improve resolution, not negate the addition of pixels. The actual reason for the meager improvement in detail resolving power was explained in my previous post above. I suspect the focal length figures have been normalized to something, similar to how many lenses in the photography industry are normalized to 35mm.

Model	2.8mm H-FOV	3.6mm H-FOV	6mm H-FOV
Color4K-T	102°	88°
Color4K-X	102°	88°	55°
Color4M-TL	112° (wider)	94° (wider)
Color4M-XL	112° (wider)	94° (wider)	58° (wider)
T24PRO-X	106° (wider)	87° (slightly narrower)
5442-S3	112° (wider)	88° (same)	61° (wider)
5459-IL	112° (wider)	89° (slightly wider)	60° (wider)
T54PRO-AS	111° (wider)	95° (wider)

 

[OD1]

So your saying a 1/1.2 8mp vs a 1/1.8 4mp there is no difference identify distance? Both at 2.8mm.

Would not a larger sensor and higher mp have longer identify distance at the same focal length unless I'm missing something?

Yes. At 2.8mm focal length, if that lens if focused on a smaller sensor, the FOV is smaller.

If you double sensor size but also megapixels, you get the same pixel pitch and pixel density is f/pitch. So basically exactly the same. 24-25 PPF at 50’

I guess you could say the problem is that camera makers always ship the same lenses with vastly different sized sensors so they have vastly different “zoom” despite both being 2.8mm.

edit: I had to read a little now after reading they advertise similar FOV. It's very possible that "2.8mm" is more of a suggestion instead of a specification. That's what I get for trying to apply optics logic to the surveillance industry. verafocal models seem to be a little more honest.

edit2: after more math and reading, Dahua just uses focal length as a suggestive or normalized out of thin air number. Axis specifies actual focal lengths. There is no standard, but since this is supposed to be an optical specification, I would much prefer Dahua use their actual lens specifications or just not publish these numbers at all. Always use FOV because f is completely useless and madly confusing.

 

[raidflex]

Yes. At 2.8mm focal length, if that lens if focused on a smaller sensor, the FOV is smaller.

If you double sensor size but also megapixels, you get the same pixel pitch and pixel density is f/pitch. So basically exactly the same. 24-25 PPF at 50’

I guess you could say the problem is that camera makers always ship the same lenses with vastly different sized sensors so they have vastly different “zoom” despite both being 2.8mm.

edit: I had to read a little now after reading they advertise similar FOV. It's very possible that "2.8mm" is more of a suggestion instead of a specification. That's what I get for trying to apply optics logic to the surveillance industry. verafocal models seem to be a little more honest.

Yeah I think it's a little more complicated the way each camera maker listed the specs and not as logical as it should be based on the optics alone.

Either way these camera are light years ahead of the consumer junk most people use.

 

[raidflex]

So in the end it seems either get the IPC-T54IR-ZE
or the color4k-T at this point. Seems like more AI features then anything was added with the newer models which with Frigate I don't need any of that.

I did notice the older models are more expensive then the newer ones though.