A119 V2 with 12mm lens experiment

DAP

Well-Known Member
Joined
Sep 5, 2014
Messages
547
Reaction score
283
Location
California
Country
United States
Dash Cam
Two Viofo 129 PRO cameras and a Tesla Model 3 4 Channel
I had previously done a test with a 12mm lens on the A129 pro duo rear camera.
I had an A119 v2 lying around, not in use because it was prone to silently stopping recording. (I have since learned that the silent part of this failure seems to be that the speaker is not working).
Since it is not reliable, I decided to test it with an extra 12mm lens I had lying around.
When I disassembled the camera, I unfortunately broke two of the latches that hold the moving part of the camera enclosure together, but I found that that will be held together by the not moving part of the camera enclosure.
I also discovered that this camera uses the same lens mount as the A129 pro rear camera (19mm screw spacing). Having experience with the A129 pro rear camera, I knew that I would have to build a spacer between the sensor circuit board and the lens mount to allow the 12mm lens to focus.
The moving part of the camera housing is much smaller than the one on the A129 pro rear camera, so I had to make my spacer much thinner, and still had to trim away part of the camera housing to allow the extended housing to fit.
The camera is assembled and focused. I have not done any driving tests with it because it is Halloween, and I don't want to be driving around when there are a bunch of kind running around half blinded by their masks.
I do have a focus test photo to include.
I have had to re-format this from png to jpg so that this forum will allow me to attach it.
The focus targets are 90' from the camera, the hay stack looking thing is 140'.
It looks like the sharpness of the lens is matching the resolution of the sensor pretty closely as the target images degrade to grey just before they would begin aliasing. I would have preferred aliasing as it would mean that this lens could be used on a higher pixel density camera.
vlcsnap-2022-10-31-16h14m03s800.jpg
 
It looks like your lens doesn't have an IR filter?
 
It looks like your lens doesn't have an IR filter?
Yes, this camera lens does not have an IR filter. This is intended as a secondary camera with a better chance of capturing license plates. No IR filter means it is more sensitive to light, and will therefore use a shorter exposure. The main camera will capture the scene in normal color.
 
No IR filter means it is more sensitive to light, and will therefore use a shorter exposure

This statement is not true; it doesn't work that way.

Without the IR-cult filter a digital camera's sensor is sensitive to the near infra-red wavelengths of approximately 650-1200 μm (nanometers), especially red and magenta between about 625 μm and 750 μm and this is why you end up with the magenta false color image. Depending upon the colors (frequencies) the camera is seeing in the visible range of the electomagnetic spectrum some colors will be far less visible and discernible than others in the magenta image where no IR-cut filter is in the light path. The result will be that many license plates and other image details that might otherwise be clearly visible in a full color visible spectrum image may render as washed out or even almost non existent in the magenta image when there is no IR-cut filter.

Also, in an uncorrected infra-red image that is otherwise in focus, sharp details of objects that would be clear in a full color image may render as fuzzy or hazy, (especially foliage), as in your posted example.

Having the IR-cut filter in the light path of the lens has a negligible, essentially undetectable effect on shutter speed.

The notion that the camera sensor is somehow "more sensitive to light" and therefore will "use a shorter exposure" without the IR-cut filter in place is simply wishful/magical thinking.
 
Last edited:
I would get a IR filter on that too.

BTW, guy next door ( 10310 Anderson Rd ) , nice house.

Not much to brag about where i am 56.46750224669076, 10.025679919038486 and aint no drive from here to the hills, as hills / mountains pretty much do not exist in Denmark.
 
More work to do...
vlcsnap-2022-11-01-16h48m57s723.jpg

For some reason the right side of the image is much sharper than the left. Could be I got a fingerprint on the lens, or it could just be the cheap lens I'm using. In any case, I will have to disassemble the camera again to try and see what is going on.
And no, I did not add an IR filter. You only really see the effects of IR on vegetation, and incandescent lights.
 
  • Like
Reactions: Mtz
For some reason the right side of the image is much sharper than the left.
I've sometimes experienced one side being sharper than the other, more often with longer focal lengths = shallower depth of field. I think a small amount of mis-alignment between the sensor and the lens holder, either angular or displacement, can cause one side to drop out of focus.

I've spent the past 3 days checking and adjusting focus on my 6mm camera. It's rare that I get it right first time.

If I can't get the whole image perfectly sharp, my preference is to get the driver's side correct since that is where oncoming and passing cars will be for most of the time.

You may also find that some lenses are only sharp in the middle and not around the edges, not so much through distortion which can be an issue, but as if the focal plane is curved so that rotating the lens will render either the middle or the edges in good focus.

Good luck finding the sweet-spot for your camera & lens combination.
 
Last edited:
It looks like the asymmetrical focus problems are due to poor collimating of the lens.
DSC01851_Cropped.JPG
The reflections of the ring light should be concentric, instead they are all over the place. I will have to buy a better lens.
The lens appears to be OK for the 1080P sensor on the A129 pro rear camera sensor, but definitely not OK for the 1440P lens on the A119 V2.
Usual rule of thumb, you get what you pay for appears to apply here. Two lenses for $12 is too cheap.

Anyone know a good source for lenses? Amazon seldom lists the brand of lenses they sell, and there is no way to judge the quality from the descriptions of what they are selling. It is also clear that the "5MP" claim means nothing on Amazon.
 
Last edited:
No IR filter means it is more sensitive to light, and will therefore use a shorter exposure.

How much effect depends on how sharp the cut off is on the filter being used, expensive filters can have very little effect on things that don't reflect NIR light, the auto exposure is normally looking mainly at the road, which should not reflect much NIR. Normal IR cut filters can give quite a large reduction in visible light towards the red end of the spectrum. Of course if things are giving off NIR light, like in your image above, then removing the IR filter will affect the auto exposure and reduce the exposure times significantly, the sensor sees the NIR light as magenta light since its RGB filters are not very effective at blocking NIR, so the auto exposure sees the NIR on red and blue pixels, and a fair amount on the green, it doesn't know that it is NIR so can't ignore it.

post-329120-0-37878500-1637594492.png

For some reason the right side of the image is much sharper than the left.
Always worth rotating the lens holder 180 degrees to check if the focus issue reverses, it doesn't take much for a lens holder to cause a focus issue, even the tightness of the screws can have an effect.
 
Baader Filters are specifically designed for astrophotography where longer exposures are the norm and the desired light transmission wavelengths are extremely different than dashcams or other typical digital cameras used in mundane daylight settings. Astrophotography filters are completely unrelated to the kind of IR-cut filters used in dash cameras. The reason your chart compares the Baader BCF filter to the Canon is that the company specializes in astrophotography hot mirror conversions for Canon DSLRs which are popular for this use.

"The Baader Planetarium BCF filter, allows light with wavelengths of 420-680 nm to reach the sensor. The camera becomes sensitive to light in the ultraviolet, visible and infrared ranges (starting at 300 nm) DSLRs are very often used for astrophotography because they offer large sensor and low electronic noise. This means a 4 times higher sensitivity for H-alpha than an unmodified camera. In order to keep a proper color balance in daylight use, DSLRs use a filter placed in front of the sensor that is not completely transparent, especially at longer wavelengths corresponding to red. In astronomical application that is a problem because the most important emission line, the H-alpha (emitted by many nebulae for example) is right in the red part of the spectrum. The 800Da camera replaces the original filter with Baader BCF specifically designed for astrophotography, increasing the camera sensitivity and allowing you to take brighter pictures."


The IR-cut filters used in dash cams tend to use very thin glass and have a sharp cut-off at 650 nm and as such tend to have excellent transmission characteristics and low filter factors. Not that much of the NIR the spectrum above 650 nm is usually blocked by these filters.

Why use an completely unrelated filter chart to try make points that are not accurate? !
 
Last edited:
Anyone know a good source for lenses? Amazon seldom lists the brand of lenses they sell, and there is no way to judge the quality from the descriptions of what they are selling. It is also clear that the "5MP" claim means nothing on Amazon.

 
Aak! that is the mobile site! works terrible on a computer with a 4K screen! I finally figured out that I needed to remove the "m." at the beginning of that link.
In any case, I ordered this: https://www.aliexpress.us/item/2255...c10Evnl&gatewayAdapt=glo2usa&_randl_shipto=US
Won't arrive until December though.
Looks like it has a bigger aperture than the cheap junk I bought before. Better light capture, worse depth of field. Everything is a tradeoff.
It is a 4K lens, I hope this time that means something.
 
Back
Top