A129 Pro Quality Issue

[Ariontor]

Hi everyone, I recently started using the A129 Pro and I'm getting some quality issues. The color seems distorted and perhaps overly compressed. I've especially noticed this on trees and other greenery. I added some snapshots from the video footage with the original size and then a cropped down version to highlight the issue. Is this normal quality for the camera? I tried with and without WDR but had the same issue. It came with firmware v. 2.8 and I tried downgrading to 2.7 but I don't think that made a difference.

 

[Nigel]

Considering the time of day, I think they are pretty good!

Moonlight always gives a slightly odd colour balance

 

[Nigel]

overly compressed

Which bitrate setting are you using?

When driving past trees, there is a huge amount of detail to be stored, every leaf needs recording, so it does require a high bitrate and can at times start to show a bit of pixelation. Generally a lot less than other dashcams though, since the A129 Pro can use very high bitrates relative to most.

I recommend keeping WDR off for the A129 Pro.

Colours look OK to me, the evening images are a little blue, but without a sunset that is reality. What is it you don't like about the colours?

 

[Milos7]

Hi, and thanks for posting your findings. Ill try to post some frames later.
I noticed the same compression artifacts in many situations in my A129Pro. I also have to tilt the camera down so that horizon is up at about 60-70% so I don't get weird contrast and dark video in the lower more important half. I always use maximum bitrate and have tested both WDR ON and OFF.
For the day use WDR does help in dark areas resolving the details while not overexposing bright lit areas. In the night I am still testing WDR ON vs OFF in v2.7

p.s.
VIOFO support just confirmed to me that WDR in Version 2.8 is recommended NOT to be used for night. It's optimized for day use.

 

[kamkar]

A expensive 4K dashcam are nothing near a expensive 4K action camera, even if you are using the highest bitrate / image quality,,,,, which you should always use.

 

[Milos7]

Good thing is, they are getting better just like the ones in cellphones did 5 years ago. Thats why I did not want to spend $500 fo a dash cam today. VIOFO 4k is a great price for the quality of image comparing to other dash cam competitors.

 

[Nigel]

A expensive 4K dashcam are nothing near a expensive 4K action camera, even if you are using the highest bitrate / image quality,,,,, which you should always use.

I've seen plenty of comments about pixelation on DJI drones, the action cameras tend to do better because there is normally a lot less movement when they are being tested, and it is the movement that requires the high bitrates, unfortunately when driving we normally have a lot of movement!

 

[Nigel]

I noticed the same compression artifacts in many situations in my A129Pro. I also have to tilt the camera down so that horizon is up at about 60-70% so I don't get weird contrast and dark video in the lower more important half.

It is recommended that you have the horizon at 50% height for Viofo cameras, if you put it a little higher then you start to loose the clouds to overexposure at times, while putting it lower will underexpose the road every time the sky gets to the centre of the frame.

I think you will find the colours better in daytime rather than evening, and the overall contrast gets better when you don't have midsummer sun.

As for the "compression", I'm not sure what I am seeing since it is difficult to judge how much you have zoomed in. Can you put a bit of original video on Google drive or equivalent so that I know what I am looking at? Preferably some taken in good daylight, plus some evening if that is when the issue occurs.

Remember that it is a dashcam, and exposure is set to record details like plates in the shadows in preference to just photographing scenery.

 

[Milos7]

It is recommended that you have the horizon at 50% height for Viofo cameras, if you put it a little higher then you start to loose the clouds to overexposure at times, while putting it lower will underexpose the road every time the sky gets to the centre of the frame.

I tried 50% and the lower half was dark and green full of compression artifacts that Ariontor is showing I think. The plates of the cars and anything below parked near by were in darkness only 5-10m away during day if the sky was showing in upper 40%. After I moved camera one click down, the exposure and overall color balance is much better in the lower half.

Do you keep WDR OFF during the day as well? And which firmware do you have? Thanks.

 

[Ariontor]

Thanks for the information everyone!

Considering the time of day, I think they are pretty good!

Moonlight always gives a slightly odd colour balance

Nigel, the date and time stamps in these photos are incorrect; I just hadn't set it on the camera. These were all during the afternoon or evening.

Which bitrate setting are you using?

When driving past trees, there is a huge amount of detail to be stored, every leaf needs recording, so it does require a high bitrate and can at times start to show a bit of pixelation. Generally a lot less than other dashcams though, since the A129 Pro can use very high bitrates relative to most.

I recommend keeping WDR off for the A129 Pro.

Colours look OK to me, the evening images are a little blue, but without a sunset that is reality. What is it you don't like about the colours?

I was using the high bitrate option. Later I tested briefly with the maximum bitrate but didn't notice a difference in the video. As for the colors, it's not so much that they're wrong, it's mostly just that it looks compressed in places with some chunks being all of a more or less single shade without much gradient or definition.

A expensive 4K dashcam are nothing near a expensive 4K action camera, even if you are using the highest bitrate / image quality,,,,, which you should always use.

That might be my main issue lol. I was probably hoping for too high of quality from something that costs a good bit less than a GoPro.

Milos7, Nigel, I might try changing the height angle for the camera. Thanks for the suggestion!

 

[Milos7]

Ariontor,
How are you watching your videos? Which software? I use Movist Pro and I see very well these compressions artifacts. When I upload them to Youtube they get transcoded differently into 1080 and most of those artifacts are gone, so overall video appears much cleaner. Thats probably also why VIOFO videos on YouTube from many reviews and other user uploads appear better than yours.

 

[kamkar]

If that is the case you are not the first one.

My Osmo action camera record with 100 mbit, that allow for it to deal with a lot of stuff, where as the 30 - 40 mbit of a 4K dashcam have less of a "spine" to do the heavy lifting.

Playback software indeed also have a huge say in the footage you see on your screen, the standard windows media player er pretty poor, i use pot player or VLC myself.

 

[Nigel]

Nigel, the date and time stamps in these photos are incorrect; I just hadn't set it on the camera. These were all during the afternoon or evening.

Yes, I had worked that out, and had noticed that the ones from the afternoon have significantly nicer/less blue colour than the ones from the evening.

Later I tested briefly with the maximum bitrate but didn't notice a difference in the video.

It will only make a noticable difference when there is a lot of information to store, such as when driving fast under trees. There is definitely a difference in some situations.

As for the colors, it's not so much that they're wrong, it's mostly just that it looks compressed in places with some chunks being all of a more or less single shade without much gradient or definition.

Ahh, the codec tends to compress colour information first, before removing resolution, since as it is a dashcam we want to be able to read plates, where resolution is critical, but colour resolution is unimportant. If it was a GoPro then it might be done the other way around!

That might be my main issue lol. I was probably hoping for too high of quality from something that costs a good bit less than a GoPro.

It is more an issue of different compromises. It is trying to record evidence, not beautiful movies, it will record detail in shadows when a GoPro would simply show the shadows as black, it always uses fast exposure times to minimise motion blur, while a movie camera will be more worried about having enough motion blur to hide the steps between frames, you never use a GoPro at 24/30fps, unless you also set slow exposure times, but you normally use a dashcam at 30fps because recording 60fps is a waste of bitrate as far as evidence is concerned.

How are you watching your videos? Which software? I use Movist Pro and I see very well these compressions artifacts. When I upload them to Youtube they get transcoded differently into 1080 and most of those artifacts are gone, so overall video appears much cleaner.

Just converting the videos from H264/H265 into Youtube's VP9 or AV1 will do the same thing, all the pixelation in the sky disappears! Works if you keep the original resolution too.

 

[Milos7]

....but you normally use a dashcam at 30fps because recording 60fps is a waste of bitrate as far as evidence is concerned.

Can you elaborate a bit on 60fps please.
I was testing 60fps 1440p vs 30fps 2160p and chose 4k/30 over 2k/60.
I was reading that the 60fps is fake and that the extra frame is just copy of previous, but thats BS! A129Pro is definitely a true 60fps. Deffinitely less motion blur, but less detail in many situations because of lower rezolution. I checked frame by frame and counted them. Actually it does more like 64fps
But can you share your thoughts on "evidence" part.

 

[Nigel]

I was reading that the 60fps is fake and that the extra frame is just copy of previous, but thats BS! A129Pro is definitely a true 60fps.

You are correct, except that at night it may choose exposure times longer than a 1/60th of a second, in which case it can't record separate exposures for each frame since there is not enough time, so in the dark it is possible that it will duplicate frames to keep the 60fps while obtaining reasonable exposures. I'm not actually sure that the A129 Pro does, but I expect it to.

Deffinitely less motion blur,

That is generally not true, normally the exposure time is the same as it would be for 30fps. If it is using a 1/2000th second exposure, then that is the correct exposure whatever the frame rate. There may be some situations where there is less motion blur, but if you put two cameras side by side, one set to 60fps and one to 30fps, I wouldn't like to take a bet on which would show the least motion blur! I think any difference is most likely due to the two modes having their exposures calibrated a little differently, not because of the use of a different frame rate.

but less detail in many situations because of lower rezolution.

I think it is always the case that there is less detail because of the lower resolution. It is possible that it is sometimes the same amount of detail, but only if there is enough motion blur throughout the image because you are turning a corner in darkness.

Can you elaborate a bit on 60fps please.

In general the exposures are the same, the motion blur is the same, and the bitrates are approximately the same, so the amount of information stored is the same. In fact the bitrates are sometimes less for 60fps because the processor is having to do more work processing twice the amount of frames, in which case less information is being stored, so less evidence.

So it is best to think of 60fps as a tradeoff, where you sacrifice some spatial resolution in order to gain some temporal resolution. When it comes to evidence for insurance and court purposes, 30fps almost always gives us plenty of temporal resolution, it is the spatial resolution that is lacking.

From a base of 2K 30fps, I would always choose to move to 4K rather than move to 60fps, unless I was movie making. The only extra issue with 4K is that the sensors are not as sensitive, so tend to have more motion blur, which is why we tend to use 2K rather than 4K dashcams, but if you are making the choice within the A129 Pro then you are using the same sensor whatever the choice.

 

[Milos7]

You are correct, except that at night it may choose exposure times longer than a 1/60th of a second, in which case it can't record separate exposures for each frame since there is not enough time, so in the dark it is possible that it will duplicate frames to keep the 60fps while obtaining reasonable exposures. I'm not actually sure that the A129 Pro does, but I expect it to.


That is generally not true, normally the exposure time is the same as it would be for 30fps. If it is using a 1/2000th second exposure, then that is the correct exposure whatever the frame rate. There may be some situations where there is less motion blur, but if you put two cameras side by side, one set to 60fps and one to 30fps, I wouldn't like to take a bet on which would show the least motion blur! I think any difference is most likely due to the two modes having their exposures calibrated a little differently, not because of the use of a different frame rate.


In general the exposures are the same, the motion blur is the same, and the bitrates are approximately the same, so the amount of information stored is the same. In fact the bitrates are sometimes less for 60fps because the processor is having to do more work processing twice the amount of frames, in which case less information is being stored, so less evidence.

Im not sure what it does at night in driving mode, but I can check next chance I get to see if 60p are duplicates or 60 distinct moments in time.
One thing A129 Pro does do during night (parking mode) is lower resolution and frame rate to 1440/30p no matter what the settings are during the day driving. I don't think there is a way around that.

As for 60fps having same amount of motion blur as 30. Why would exposure time stay the same? It should be half, and it should be gain/tone curves adjustments that compensate the brightness of the image. Otherwise whats the point of 60fps unless you want some slow motion as result
I don't know what they do , but if you are correct and exposure is fixed, then I can't explain how I saw difference in motion blur between 30 and 60 fps. I recorded myself walking fast from left to right in front of the car on sunny day, and played it back frame by frame to see details. Edges appeared cleaner in 60p. I might have gotten tricked, or the exposure is indeed halved in 60p.
Best regards
Miloš