Reviewing footage on pc

[HonestReview]

I remember watching a documentary program about Police tactics. One of the programs showed a speedcam shot where the owner had installed these anti-flash plates that are normally impossible for speedcams to read, especially at night, so they fit them and speed through cameras.

The operator then just flipped the image into negative mode and there was the number, plain as day - gotcha

Mythbusters tested those anti-flash plates and proved they're nothing but garbage. None of them worked.

 

[Dashmellow]

I would expect that to work fine, my comment was based on the plates in your images all being too dim to be considered highlights. Possibly your reflective plates reflect a lot less light than european ones, but that seems unlikely, presumably they are all retro-reflective plastic and not pressed metal plates with reflective paint?

There does seem to be an exposure issue in your images too, the dark areas in your last set of images are a long way from black and seem very noisy compared to those from other cameras, but not many people are posting images that don't have streetlights, and your very last image which does have some lighting looks fine. I doubt that is affecting the performance of the HDR though, just looks like Viofo haven't calibrated this firmware for use in full darkness with no streetlights, not easy for them as they live in a brightly lit city, you could point that out to them and maybe they will give you a better calibrated firmware to test out in your lighting conditions.

Every US state has a different approach to the style and manufacture of license plates. Some use reflective plastics but many use embossed or debossed steel or aluminum often with unique hybrid stenciling for the lettering using highly reflective glass beaded coatings or film laminations that are retro-reflective. Vermont plates are aluminum with debossed highly reflective lettering on a green reflective background.

The reflectivity of the plates is not of issue here as evidenced by anyone who more typically experiences completely blown out license plates at night with virtually any dash camera or firmware. I believe the smaller size of the plates makes them more difficult to capture at night.

As for the V2.6T firmware, which I've only tested briefly so far for one short night time excursion, I need to do further evaluation before reaching any definitive conclusions. I did find an unexpected fingerprint on the CPL which may well be contributing to the poor results and the haze I'm seeing. Not sure how I got a fingerprint on the CPL but I will start testing again from scratch.

As for Viofo calibrating their A119 V3 firmware for driving in full darkness you will no doubt recall the video I posted of a barn owl striking my windshield last December. The camera's performance in full darkness was superb! (I don't recall at the moment which firmware version was installed but I believe it may have been V.2.3)

 

[HonestReview]

This is a bit off topic to the A119 but does apply. It amazes me how every camera tries to handle scenery and plates differently. @TonyM's video shows the wide differences on how well a camera picks up a plate between the varying Dash Cam Models.

Does HDR apply only to the 119 or does it exist in the A129? A139? T130? firmwares.

 

[Nigel]

As for Viofo calibrating their A119 V3 firmware for driving in full darkness you will no doubt recall the video I posted of a barn owl striking my windshield last December. The camera's performance in full darkness was superb! (I don't recall at the moment which firmware version was installed but I believe it may have been V.2.3)

Yes, A119 v3 is superb, however I think the calibration needs to be different when using HDR, the sensor is operating in a different mode, so when HDR is on, it is possible the current calibration may be poor in some situations.

 

[Kremmen]

I reviewed my drive home footage earlier which was in the dark. There is no way on poor or unlit roads anyone is going to capture a front reflective plate at a closing speed of 60mph or higher.

With exactly the same settings on my daylight outbound journey the clarity is superb.

What is noticeable, same settings, daylight - no motion blur, in the dark - motion blur.

 

[gse]

There is no way on poor or unlit roads anyone is going to capture a front reflective plate at a closing speed of 60mph or higher.

 

[dAk1nG]

which mod firmware was used here ?

 

[Kremmen]

Here's the best I can offer :



With the same settings on the outbound journey :

 

[Dashmellow]

I started again from scratch with a fresh install of V.2.06T and a clean CPL. This was my best result at dusk. So far I've been unable to achieve anything like some of the other examples we've seen posted to the forum.

The camera is making it appear brighter than it really was at this hour. When I get an opportunity I will try again in completely dark conditions.

 

[Kremmen]

That's very similar to what I'm getting. It does look like darker conditions introduce motion blur.

I'm wondering whether gse shot was with a highly tweaked firmware and if so is it temperature safe as I seem to recall a mention of the bitrate being increased from 30k to 50k ?

 

[Dashmellow]

That's very similar to what I'm getting. It does look like darker conditions introduce motion blur.

I'm wondering whether gse shot was with a highly tweaked firmware and if so is it temperature safe as I seem to recall a mention of the bitrate being increased from 30k to 50k ?

Obviously, it is much harder to capture the smaller American license plates but if you are getting similar results in the UK there is something else going on here.

 

[gse]

This was my best result at dusk.

At dusk the result will not be good, you have to watch it at night.

 

[Kremmen]

Thanks, but my 06:51 shot above was night, dawn not due on that day till 07:36. The snag here is that UK plates are highly reflective. I could be wrong but some countries plates give the impression of just having a white painted background.

Maybe if night plates are required you need 2 dashcams configured differently

 

[gse]

I could be wrong but some countries plates give the impression of just having a white painted background.

In Russia, the numbers are metallic with a reflective coating.

 

[Kremmen]

I wonder how much of a difference that makes.

 

[JockoT]

The camera has a fixed aperture. To capture the right amount of light you have to vary the "shutter speed". Now whether this is mechanical (as in old-time cameras), or electronic (as in modern digital cameras such as phones and dash cams) you need a longer "shutter speed" in lower light levels. This is why number plates appear blurred in lower light levels and pin-sharp in bright sunshine. At 60 mph and a "shutter speed" of 1/1000 of a second, the image will move 1.06 inches during the exposure. At 1/20 of a second "shutter speed", the vehicle will move 52.8 inches. This is the reason isolated frames at night are so blurred.
It is not as straightforward as this example as digital cameras also vary the "film speed" for low light images leading to less motion blur but much-degraded image quality (noise).

 

[Kremmen]

Makes perfect sense, thanks.

 

[Dashmellow]

The camera has a fixed aperture. To capture the right amount of light you have to vary the "shutter speed". Now whether this is mechanical (as in old-time cameras), or electronic (as in modern digital cameras such as phones and dash cams) you need a longer "shutter speed" in lower light levels. This is why number plates appear blurred in lower light levels and pin-sharp in bright sunshine. At 60 mph and a "shutter speed" of 1/1000 of a second, the image will move 1.06 inches during the exposure. At 1/20 of a second "shutter speed", the vehicle will move 52.8 inches. This is the reason isolated frames at night are so blurred.
It is not as straightforward as this example as digital cameras also vary the "film speed" for low light images leading to less motion blur but much-degraded image quality (noise).

What you say is quite true but with HDR it changes the equation (or tries to) because an algorithum produces two images simultaneously at different exposures and shutter speeds that are superimposed on one another. In theory this has the potential to capture an image with a fast shutter speed on top of a motion blurred slow shutter speed image. Apparently though this needs very specific conditions that are not always available and so it doesn't always work.


@gse's example shows the HDR effect working well with a sharp plate capture on top of a motion-blurred capture.

 

[Kremmen]

A couple of things I've noticed.

gse's image does have some secondary light as you can see a lot more than my image which was in total darkness other than headlights. I noticed similar with a vvs example where there was other lighting, almost like a lit shopping street.

The other thing is that in the UK you're not allowed to put anything on the windscreen where the drivers side wiper sweeps. Both gse and DM's image give the impression the cam is on the drivers side so maybe a better angle ?

I think in my case there's no improvements to make with my current firmware. I need to travel down a street, at night, with better side lighting as a test.

Problem is, I pick my travel times, because I can, and driving at night when the shops are open isn't going to happen

 

[Dashmellow]

A couple of things I've noticed.

gse's image does have some secondary light as you can see a lot more than my image which was in total darkness other than headlights. I noticed similar with a vvs example where there was other lighting, almost like a lit shopping street.

The other thing is that in the UK you're not allowed to put anything on the windscreen where the drivers side wiper sweeps. Both gse and DM's image give the impression the cam is on the drivers side so maybe a better angle ?

I think in my case there's no improvements to make with my current firmware. I need to travel down a street, at night, with better side lighting as a test.

Problem is, I pick my travel times, because I can, and driving at night when the shops are open isn't going to happen

I agree. There are too many variables to offer consistency with how well these mods seem to work. Also, it is likely that some of the example screen shots we see on the forum that look so impressive may simply be cherry picked. A raw video of these examples might offer a fuller opportunity to assess how well these mods are truly functioning.

Out where I live there are no street lights. Only when I get closer to the nearest town some ten or so miles away do I start to encounter street lights and greater illumination. Often I find myself driving in pitch black conditions on the back roads in my "hood". More often than not though, I don't do all that much driving at night anyway these days since the pandemic began. Probably the most driving I do in the dark is around the time of the winter solstice in December when night starts falling quickly here around 4pm in the late afternoon. So, it may be that night time plate capture doesn't matter so much for me. However, if these otherwise well crafted mods detract from the quality of plate capture at other times, like at dusk when I am more likely to be driving, then I may be better off not using them. More testing still required.