Standardize dash cam test video?

[boilers]

I was just thinking, why not have one or more movies/videos created that can run very precise, exact images/videos on a large 4K screen, and people can set up a dash cam on a tripod and dim the lights, and point the dash cam at it and we (all) can have consistent recording SOURCE from which to compare the resultant video file outputs?

Think like a test pattern or such. Especially good for low-light resolution, catching compression artifacts, color reproduction, etc.

 

[jokiin]

not easy to emulate in other environments, some factories do have test rooms setup like this but it's very hard to try and match what you find out in the real world with lab tests

 

[boilers]

I don't even mean real world looking images/video. Test patterns-ish.

 

[jokiin]

I know what you meant but something that would be very hard to replicate from one user to the next

 

[Dashmellow]

There is simply no way a movie screen or video monitor of any type or resolution can reproduce the luminance values, dynamic range, specular highlights or fine details of a real world environment. Test patterns are generally used for calibration with the exception of comparing lens performance and would be of little value in comparing camera performance because the consistent "source" you are suggesting would be inadequate.

 

[boilers]

all I'm saying is have one REFERENCE to compare from test video to test video. Not intending to replicate real world. And my 65" 4K is pretty dynamic. Not retina-searing, stare at the sun bright, but no dash cam should be judged on such a situation.

 

[Dashmellow]

all I'm saying is have one REFERENCE to compare from test video to test video. Not intending to replicate real world. And my 65" 4K is pretty dynamic. Not retina-searing, stare at the sun bright, but no dash cam should be judged on such a situation.

What may seem dynamic to the human eye is perceived very differently by a digital camera.

 

[Feitelijk]

If you get close enough to the screen to fill the entire view of the dashcam, things would be out of focus, And worse, the corners would be mucht further from the screen than the center. There's no way this will give reliable results comparable to real world.

 

[boilers]

I don't equate reliable with consistent, though. Again if that would hold true for every DC, then that can still be a common benchmark to observer and analyze. my 2¢

 

[2000rpm]

-

Here's my theoretical test result, cropped and magnified as required from video footage in order to read a number plate.

Which cam is the best?
(Yes, there's a catch!)

-

 

[boilers]

must be same cam

 

[russ331]

Here's my theoretical test result, cropped and magnified as required from video footage in order to read a number plate.

Which cam is the best?
(Yes, there's a catch!)

The cam on the left has a narrow FOV, & hasn't been magnified & cropped as much as the image from the cam on the right, which has a wider FOV. In practise, however, I'm guessing that the cam on the right is the better performer overall.

 

[2000rpm]

The cam on the left has a narrow FOV, & hasn't been magnified & cropped as much as the image from the cam on the right, which has a wider FOV. In practise, however, I'm guessing that the cam on the right is the better performer overall.

 

[2000rpm]

must be same cam

Two very different cams - @russ331 worked it out.

One cam has a narrow field of view while the other has a wide field of view.
The narrow one sees more detail and can look impressive, but it misses a lot of what's happening at the sides and sometimes even cars passing on a bend are completely outside the field of view of a narrow-angle cam.
The narrow one also has a much inferior lens/sensor arrangement which impairs night performance.

If I could have only one of those two cams, I'd choose the one that produced the 'inferior' image on the right.

-

 

[Sutton]

-

Here's my theoretical test result, cropped and magnified as required from video footage in order to read a number plate.

Which cam is the best?
(Yes, there's a catch!)
one on the right is terrible but both are so inferior to quality exposure
-

 

[2000rpm]

one on the right is terrible but both are so inferior to quality exposure

The 'terrible' one on the right is a SGZC12RC.

I made it appear to be a 'terrible' cam by carefully choosing the comparison. I did mention, in pale text in my earlier post that 'there's a catch'.

I hope @jokiin doesn't mind. The SGZC12RC is nowhere near as 'terrible' in the real world as I made it appear in my 'test' (which is why I didn't mention the make/model in my initial, biased, carefully chosen screenshot).
The whole point of my earlier post was that it's possible for good cams to appear poor and mediocre cams to appear good, depending on the test.
For the record, the SGZC12RC is one of my favourite cams.

 

[2000rpm]

-

And here, for fairness, are some similar comparisons, although it's a Panorama II S as a proxy for the SGZC12RC (I find recordings from the two cams are virtually indistinguishable).

Night-time is a different matter - the 'terrible' cam is two or three steps ahead of the other one at dusk.

Second and third pics show the huge difference in field of view - the 'terrible' cam reveals that there's actually a whole lane beside me which the other cam's blinkered view can't see.



-

 

[Sutton]

The 'terrible' one on the right is a SGZC12RC.

I made it appear to be a 'terrible' cam by carefully choosing the comparison. I did mention, in pale text in my earlier post that 'there's a catch'.

I hope @jokiin doesn't mind. The SGZC12RC is nowhere near as 'terrible' in the real world as I made it appear in my 'test' (which is why I didn't mention the make/model in my initial, biased, carefully chosen screenshot).
The whole point of my earlier post was that it's possible for good cams to appear poor and mediocre cams to appear good, depending on the test.
For the record, the SGZC12RC is one of my favourite cams.

Ohhhh ok then it was not the original raw footage .It was a doctored up sample because I own a SG and I find that it is the best DVR that I ever owned. I own an A118-c., An A119 and the SG is far superior

Street Guardian SG9665GC v2 2016 edition

 

[2000rpm]

-
Original screenshot captures from which I cropped and zoomed the earlier pictures.

-