VIOFO A119 Mini 2 - Test and Review RCG

[Nigel]

Seen some posts on this form, saying that Ultra wide is "better" for overall picture quality.

Ultra wide is great for 4K, but doesn't work so well for the 2K cameras that have narrower fields of view.

 

[rcg530]

Would it be possible, if u still have the 2x A119 Mini2's setup, to do a direct compare of 16:9 vs 21:9 (ultra Wide).
That is one cam set to 2560x1440 60fps and the other 2560x1080 60fps.

I only have one A119 Mini 2 installed in my car that the moment. I've been prepping for an A229 Pro / A229 Plus / A139 Pro video comparison, but my A229 Plus firmware testing has been encountering problems that are preventing me from completing my power consumption testing. I'll see if I can temporarily install a second A119 Mini 2 again.

 

[speedingcheetah]

Ultra wide is great for 4K, but doesn't work so well for the 2K cameras that have narrower fields of view.

The A119 mini 2 has a large vertical field of view though…. It shows a crap ton of dash and sky in 16:9.

 

[Nigel]

The A119 mini 2 has a large vertical field of view though…. It shows a crap ton of dash and sky in 16:9.

Well you only lose the top and bottom, so if you still have enough vertical FoV then use the wide aspect ratio.

On image quality, I normally only notice a difference when going fast under trees, hard to see any difference in the city where there is plenty of bitrate anyway.
Unless you are not using Maximum bitrate, in which case it will help in easier situations.

I've never noticed a difference in the exposure since I've never had two identical cameras to test with, interested to see the result of the side by side test...

 

[rcg530]

@speedingcheetah here's the video you requested comparing the 2560x1440 60fps video resolution with 2560x1080 60fps video resolution. The 2560x1080 video resolution is not a simple cropping of the vertical field of view. There's a small amount of compression of the vertical field of view taking place. I have a couple quick comparisons segments so you can see the vertical field of view in both resolutions.

 

[speedingcheetah]

@speedingcheetah here's the video you requested comparing the 2560x1440 60fps video resolution with 2560x1080 60fps video resolution. The 2560x1080 video resolution is not a simple cropping of the vertical field of view. There's a small amount of compression of the vertical field of view taking place. I have a couple quick comparisons segments so you can see the vertical field of view in both resolutions.

Awesome! Thank you! (for some reason, this site did not send me a e-mail notification of your post

Anyways, the "compression" thing, yea, i noticed that and mentioned it in my post linked to previous. It is less "fish eye" looking.
In 16:9, vehicles look a tad squished vertically to me. in Ultrawide, looks more natural, but maybe a tad stretched horizontally.
People here had said the cam was just cropping, but, in fact, it is doing a tad more as we can see.
But, i think it is a metdata issue, as i can take the video, run it through MKVToolNix, and change the aspect ratio to 21:9, and it no longer has that compression issue. (or preset value 2.21)
For the raw footage from my cams, MPV live stats reports the 2560x1080 to be 2.37:1, while media info reports 2.40:1
(I do not own a "Ultrawide" monitor, to view footage on, to see if it is correct ratio to fill whole screen, and, I assume there is varying ratios in those type monitor models too...)

The main things i was looking for:
1. there was any noticeable exposure changes, or how the cam processes the image, as some had said they get better results with less sky for the cam to process.
-Do not see anything noticeable, both footage look the same to me in this reguard.
- When u go under the bridges at 3:04, the cams adjust the same, at the same time... so, if there is anything to others claims, it is very minimal, or not seen in this test of a bright sunny day.

2. Higher bit rate make much differences. 16:9 ~26.6Mb/s 21:9 ~30.3Mb/s
- Can't really tell. Would need to see very zoomed in images, say of plates or signs to see if it made any differences, but, i think it would be marginal if any.

3. Cut off things.
- I do see times, as u move under street signal lights, the ultrawide will cut off the signal light a tad sooner. But it is quite minor a differences, imo.

Overall, of cource, it is down to personal preference as to what you want to use.
(Ultrawide is just a bit easier on the eye to me)

One more test i want to do myself, is the person test.
Stand directly infront of my vehicle, legs touching front bumper, to see if ultrawide cuts off my head....
This test, of cource, would vary greatly on how tall your vehicle is, how tall you are, and exact mount position and angel of the cam lens.
But idea is to get a "pedestrian" shot, in case that situation comes up, where a crazy person tries to jump infront of me to get hit... or similar.

 

[ogzogz]

@speedingcheetah here's the video you requested comparing the 2560x1440 60fps video resolution with 2560x1080 60fps video resolution. The 2560x1080 video resolution is not a simple cropping of the vertical field of view. There's a small amount of compression of the vertical field of view taking place. I have a couple quick comparisons segments so you can see the vertical field of view in both resolutions.

I reported this situation to Viofo about 2 months ago.
In my opinion, there should be no compression. It is good to cut directly evenly.

The purpose of using this resolution is to eliminate the unnecessary vehicle hood image and extra sky, as well as to obtain a clearer image as a result of the bit value and operation in the field of view.
Compression means that the clarity will not be better than the clarity you get in a 16:9 image.

 

[speedingcheetah]

I reported this situation to Viofo about 2 months ago.
In my opinion, there should be no compression. It is good to cut directly evenly.

The purpose of using this resolution is to eliminate the unnecessary vehicle hood image and extra sky, as well as to obtain a clearer image as a result of the bit value and operation in the field of view.
Compression means that the clarity will not be better than the clarity you get in a 16:9 image.

"Compression" not meaning video encode codec compression, but physical squishing of the image. The bitrate is higher at ultrawide resolution, so, technically, should have slight more image clarity.

As i mention though, seems to be a metadata thing, as i can u can change that and it fixes the image, with no re-encode or distortion.

How VIOFO has this video resolution set to encode, 21:9. aspect ratio, but looking that up, the actual aspect ratio varies based on resolution.

 

[speedingcheetah]

as well as to obtain a clearer image as a result of the bit value and operation in the field of view.

I do not see any real differences in the image in his tests. Look at the same center line, look identical to me.
There is no significant or noticeable image quality differences with the cam set to this resolution and cropping off vertical filed of view.

 

[ogzogz]

"Compression" not meaning video encode codec compression, but physical squishing of the image. The bitrate is higher at ultrawide resolution, so, technically, should have slight more image clarity.

As i mention though, seems to be a metadata thing, as i can u can change that and it fixes the image, with no re-encode or distortion.

How VIOFO has this video resolution set to encode, 21:9. aspect ratio, but looking that up, the actual aspect ratio varies based on resolution.

View attachment 68492

There is also this detail.
In the recordings taken with a wide angle lens at 16:9 resolution, the vehicles in the middle of the frame are almost the same as those seen with the naked eye. Those included in the side view experience a shape change.
When recording 21:9 it looks the opposite due to compression. In other words, the ones joining from the side are the same as seen with the naked eye, while the ones in the middle are flattened.

Therefore, using the fisheye effect in the 21:9 selection can make all visible cars look the same as seen by the eye.

Yes, I mentioned that compression suppresses the image.
The other compression is to increase the clarity as a result of applying the required bit value to the cropped area. This cannot meet its full purpose due to pressure.

I make a calculation below:
2560x1440p standard 16:9 . The video size is the same at 2560x1080p resolution.
At 2560x1080p, the video appears to be somewhat suppressed from the bottom up. In other words, 1440-1080=360 360/2=180 equal pixels should be applied decreasing from top to bottom. At the current 2560x1080p it has been scaled down to around 150 pixels and compressed from top to bottom.

 

[speedingcheetah]

2560x1440p standard 16:9 . The video size is the same at 2560x1080p resolution.
At 2560x1080p, the video appears to be somewhat suppressed from the bottom up. In other words, 1440-1080=360 360/2=180 equal pixels should be applied decreasing from top to bottom. At the current 2560x1080p it has been scaled down to around 150 pixels and compressed from top to bottom.

I am not sure i understand what u are trying to say.
The ultra wide setting is not just "compressing" the full 16:9 image from the 2560x1440 resolution nor is it the same vertical size.
It is cropping the image, removing from view much of the dash and sky.
After that, it then appears to be applying a very slight "compression" to the aspect ratio. (or using ever so slightly wrong aspect ratio)
The pixels are still there, reguardless of how you squish or stretch an image, they just change shape/ratio.

And as i mention before, BOTH settings, to me, distort the image a bit.
One is fish eye, and make cars look more tall and boxy then they are.
The other, flattens them out, making them a tad more wide and skinny then they are.

16:19 had the advantages of being "taller" and still will capture street light signals as u pass under them, or more image of things above u, like in a tunnel or parking garage.
Ultrawide, viewer focus is more centered on the road, less distraction of sky, but can cut off things. How much, depends on your vehicle and mount position.
Still, both are not using the full camera sensor ability or full resolution. Too bad Sony made it a 4:3 sensor...but who wants 4:3 dash cam footage??

 

[speedingcheetah]

@speedingcheetah here's the video you requested comparing the 2560x1440 60fps video resolution with 2560x1080 60fps video resolution. The 2560x1080 video resolution is not a simple cropping of the vertical field of view. There's a small amount of compression of the vertical field of view taking place. I have a couple quick comparisons segments so you can see the vertical field of view in both resolutions.

At 2:22, when u are stopped, and the car drives by, that may be a good point to take a freeze frame from the source video of each, and zoom in to the persons face and maybe the cars rims etc, to compare and see if there is any noticeable quality differences. More details at all with Ultrawide since it is using slightly higher bitrate.
With Youtube compression, they look the same to me

 

[speedingcheetah]

@speedingcheetah here's the video you requested comparing the 2560x1440 60fps video resolution with 2560x1080 60fps video resolution. The 2560x1080 video resolution is not a simple cropping of the vertical field of view. There's a small amount of compression of the vertical field of view taking place. I have a couple quick comparisons segments so you can see the vertical field of view in both resolutions.

Made a gif of one frame of your vid where u switch back and forth.
(not sure if the gif making website i used would have messed with the overall aspect ratio with its output. but, its it is at least a decent enough thing)
Pay attention to the dark Suburu in front of you, (not the white sedan on the left, as this is a couple frame later and that car as moved a tad)...

 

[rcg530]

I took two full size frame grabs from the point in my video you mentioned with the car driving by my car. The faces of the people in the car look the same (quality). There is a shift in the aspect ratio.

I've uploaded the full size frame grabs here: https://drive.google.com/drive/folders/1I0FBLCd1_Gkc1613VQRg4rIoUJ2zD2wF?usp=sharing

The auto resized by DCT versions:

 

[speedingcheetah]

I took two full size frame grabs from the point in my video you mentioned with the car driving by my car. The faces of the people in the car look the same (quality). There is a shift in the aspect ratio.

I've uploaded the full size frame grabs here: https://drive.google.com/drive/folders/1I0FBLCd1_Gkc1613VQRg4rIoUJ2zD2wF?usp=sharing

The auto resized by DCT versions:

View attachment 68507 View attachment 68506

what are you using to grab the frames? Should be .png, as .jpg adds additional compression to the image and is not lossless.

For example, my frame extraction from my 2560x1440 cam footage each file is about 10.5MB

But even so, yea, can see the shift in ratio

 

[ogzogz]

Summary of the situation I mentioned; A 21:9 frame is applied, but there is no actual 21:9 image. The image printed from the bottom and top was placed in a 21:9 frame.

That is, when you select a 21:9 frame, the vehicle image in the attached image should be the same as in 16:9.

As a result, in the same picture, the image of the car should remain the same regardless of the frame 4:3 16:9 21:9.

 

[speedingcheetah]

Summary of the situation I mentioned; A 21:9 frame is applied, but there is no actual 21:9 image. The image printed from the bottom and top was placed in a 21:9 frame.

That is, when you select a 21:9 frame, the vehicle image in the attached image should be the same as in 16:9.

As a result, in the same picture, the image of the car should remain the same regardless of the frame 4:3 16:9 21:9.

Sorry, Did u edit in some way those images you posted?
The form changes the "21:9" cropped image to 1920 by 810 (64 : 27)
ANd the car changes shape when u compare those images
Can u post your raw images in a google drive or something?
Nevermind... i see what i are saying... i can do this my self.

 

[rcg530]

I'm using MPC-BE video player to extract full size frame grabs. I uploaded the png versions of the frame grabs to the Google Drive folder I listed in my previous post.

I'm done for the night....

 

[speedingcheetah]

Here is crop only of the 16:9 image to 21:9.
Compare this side by side to the original "ultrawide" image the cam outputs... can see the compression/distortion.
Though make me wonder, i see vertical stretch in the 16:9 image output of the cam, does not look natural to me, the cams "ultrawide" looks more natural. "fish eye" effect perhaps of the 16:9 2560x1440p resolution maybe?? Can really see it in Hotel sign, its stretched vertically. Look more correct in the Ultrawide from the cam, not as tall, but, not knowing what it looks like in real life by human eye, just speculating here.
Not sure what mode is most "true to life". It could be that the 16:9 is the one that is more out of wack with reality....

DCT - Google Drive

drive.google.com

 

[ogzogz]

@rcg530 @Panzer Platform @viofo @VIOFO-Support

Have you noticed that the a119 mini 2 uses variable bitrate?

It is not stable, especially when looking at the file sizes of illuminated and unlit road records in the evening. This is a good thing in terms of saving memory space.

Recently, there was color variability in 2 images taken from the video. So the shade of red was changing. Variable bitrate may have affected this situation.

Finally, is there any data on which cameras are currently tested, especially in road recordings with variable light at night, which are fixed bitrate and which cameras are variable?

Apart from the high, medium and low bitrate options in the camera settings, what is an idea to make the bitrate value selectable as fixed or variable in the menu in order to optimize memory?
The memory card with high capacity can get maximum benefit by choosing fixed bitrate, and the one with low capacity can get maximum benefit by choosing variable bitrate.