4K: Gimmick or Truly Sensational Video?

I have the DR900S and DR750S installed on my vehicle now and have been very happy with the improvements to detail and resolution by stepping up to 4K. When I get home this week I’ll be uninstalling the 750 and giving it away.

I do notice a bigger effect when stationary versus moving, but I still like the improvements in both and feel it’s a worthy upgrade in terms of detail. I don’t have any 4K displays at home and my older laptop struggles to play 4K footage smoothly, but the improvements in detail are what I want most and I’m happy with the step forward.

Has anyone tried or done any testing with H264 vs H265 yet to compare the differences in detail capture when stationary or moving? I’d like to test it myself if it hasn’t been done yet.
 
I'm not sure about the 'ethical' bit either, even though I am a bit of a 'if you don't like' it, send it back person.
I always order stuff from Halfords on a 'collect in store' basis, as the distance selling 'perks' still apply as you haven't seen the item before purchase.
 
Has anyone tried or done any testing with H264 vs H265 yet to compare the differences in detail capture when stationary or moving? I’d like to test it myself if it hasn’t been done yet.
I've tried compressing using both on my computer, and came to the conclusion that with slow moving, or still, video the H265 could get somewhere close to half the bitrate for the same quality, or can do decent 4K at only twice the bitrate of decent H264 1080. However try driving fast under trees and it needs almost the same bitrate as H264, or nearly 4x for decent 4K instead of decent H264 1080. VP9 and it's successor AV1 seems to do a much better job with dashcam video, but that wont appear in dashcams until next year.

That is compressing the files on my computer though, and it has a lot more processing power and time available than a dashcam, so I'm very interested to hear what conclusions you come to...
 
Has anyone tried or done any testing with H264 vs H265 yet to compare the differences in detail capture when stationary or moving? I’d like to test it myself if it hasn’t been done yet.
I have the Mobius 2 which can record in either h264 or h265. The h265 bitrate needs to be more than half the h264 bitrate in high detail situations to get the same video quality.

A lot of people struggle to play h265 video on their computers, either through lack of computing power or suitable software. When it comes to providing your insurer, solicitor etc with video evidence, the last thing you want is for that evidence to be rejected because someone in the legal chain can't play your video! For that reason alone, I have stuck with h264 on my Mobius 2 and would recommend the same to anyone else who has the choice between h264 and h265.
 
I have the DR900S and DR750S installed on my vehicle now and have been very happy with the improvements to detail and resolution by stepping up to 4K. When I get home this week I’ll be uninstalling the 750 and giving it away.

I do notice a bigger effect when stationary versus moving, but I still like the improvements in both and feel it’s a worthy upgrade in terms of detail. I don’t have any 4K displays at home and my older laptop struggles to play 4K footage smoothly, but the improvements in detail are what I want most and I’m happy with the step forward.

Has anyone tried or done any testing with H264 vs H265 yet to compare the differences in detail capture when stationary or moving? I’d like to test it myself if it hasn’t been done yet.
DR900S is available to buy now?
 
I agree with you on the effect of low bitrate on video detail. However one benefit of lower bitrate is increased reliability which is very important in a dashcam. I'm testing a 2k camera at the moment which can run at a very high 60Mbps bitrate and has very good detail retention, but it is having some issues with certain microSD cards.

As you say, that's down to the cards. The big issue seems to be that manufacturers aren't manufacturing high endurance memory in U3 / V90 ratings. They're manufacturing in low speeds only on the assumption that everyone is running 1080P dashcams.
 
I have the DR900S and DR750S installed on my vehicle now and have been very happy with the improvements to detail and resolution by stepping up to 4K. When I get home this week I’ll be uninstalling the 750 and giving it away.

I do notice a bigger effect when stationary versus moving, but I still like the improvements in both and feel it’s a worthy upgrade in terms of detail. I don’t have any 4K displays at home and my older laptop struggles to play 4K footage smoothly, but the improvements in detail are what I want most and I’m happy with the step forward.

Has anyone tried or done any testing with H264 vs H265 yet to compare the differences in detail capture when stationary or moving? I’d like to test it myself if it hasn’t been done yet.

I tried quickly testing the difference between h265 and h264 on the DR900S-2CH. For some reason the file sizes were the same but I didn't look much into it after. Would love to see what your thoughts were on this.

-Jimmy @ BlackboxMyCar
 
Thanks -- and congrats on your purchase (I'm jealous)!

Just got a notification from the dealer and was advised that they’re going to send out the 900S in 10 days, and if I wanted to wait I could. Made another decision to change my mind again and wait the 10 days. My overall reason for my change back to the 900 S was that the Wi-Fi issues with the 750S 2.4 GHz and my iPhone and iPad being set to 5 GHz.

Unknown below attachment would happened to me, but I want to be on the safe side.


BE5BB420-5A78-4911-8A51-37CA38DE01DF.pngBE5BB420-5A78-4911-8A51-37CA38DE01DF.png
 
Last edited:
I tried quickly testing the difference between h265 and h264 on the DR900S-2CH. For some reason the file sizes were the same but I didn't look much into it after. Would love to see what your thoughts were on this.

-Jimmy @ BlackboxMyCar

Whoa, you're right. I saw your post this morning so I switched to H.264 before I driving. Looking at the footage now that I'm back home, I can confirm that file sizes with H.264 are the same as what I was seeing with H.265. They're 193 - 197 MB/min. What's up with that? I was expecting much larger file sizes. Time to shoot Blackvue an email...

DR900S is available to buy now?

For pre-order, yeah. A few of us have sample units from Blackvue for testing and review purposes. They are production hardware and firmware, but the packaging is a little different. Otherwise they're the same as the dashcams that will be shipping out shortly.
 
Whoa, you're right. I saw your post this morning so I switched to H.264 before I driving. Looking at the footage now that I'm back home, I can confirm that file sizes with H.264 are the same as what I was seeing with H.265. They're 193 - 197 MB/min. What's up with that? I was expecting much larger file sizes. Time to shoot Blackvue an email...
Are the video bitrates the same?
 
OK, so the bitrate for both is the same at 25Mbps, which I understand is the Extreme setting. The h265 video 'should' have the equivalent quality of h264 at 50Mbps.
dr900s-quality-settings-codec-h.265-h.264.png

Can you see any difference between them?
 
OK, so the bitrate for both is the same at 25Mbps, which I understand is the Extreme setting. The h265 video 'should' have the equivalent quality of h264 at 50Mbps.

Can you see any difference between them?

That makes sense.

In that case the H.265 file, in theory, should have better quality than the H.264 file and there shouldn't be any reason to run H.264 except for compatibility purposes. I'm also noticing that 4K H.264 clips play smoothly in VLC while H.265 clips stutter pretty heavily. On newer computers I'm sure H.265 would play smoothly too.

I can try some tests later and others could conduct some too. Stationary tests would be the easiest to test side-by-side, though moving tests may show more of a difference.
 
That makes sense.

In that case the H.265 file, in theory, should have better quality than the H.264 file and there shouldn't be any reason to run H.264 except for compatibility purposes. I'm also noticing that 4K H.264 clips play smoothly in VLC while H.265 clips stutter pretty heavily. On newer computers I'm sure H.265 would play smoothly too.

I can try some tests later and others could conduct some too. Stationary tests would be the easiest to test side-by-side, though moving tests may show more of a difference.
Accidents never happen when there is nothing moving! You really should do some tests with them driving under trees, that requires far more bitrate than a static image, much more compression effort, that is when you will see the differences.
 
Nigel is spot on.

You want to drive down roads lined with trees, bushes or grass. It's here that a higher bitrate comes into it's own as with a low bit rate, compression is upped and detail is lost - the trees / bushes can become smudged (used to be macro blocking but with the latest compression advances it now appears to be smudging), and if really pushed, the whole picture can become blurred.

In my opinion, this shot from an unknown camera, taken in a highly detailed roadside scene, shows lack of bitrate in this particular situation and thus high compression effects. (That doen't mean it's a bad camera, just that this particular scene is so challenging with detail, that it's had no option to resort to high compression given it's bit rate).

I've highlighted several parts, although there are many more places the effect of compression can be seen particularly in the trees.

In the bit by the lamp post I zoomed in on, you can see all the detail has been discarded and replaced by an area of similar colour producing just a khaki smudge. In the large tree furtherest down the road to the right (highlighted with the arrow), the codec appears to have discarded all the top branches and replaced them with a grey brown smudge, also seen elsewhere in the other closer trees. The driveway to the right which you can see is a highly detailed area of stoney mud, is reduced to just a smudge of brown colour as is the road to that side. Muddy areas of the banks where the roadside joins the bushy area, are just smudges. There are many more areas if you look. All of this in my opinion is typical of bit rate starvation as is a bit of an overall lack of sharpness:




You're unlikely to see bit rate starvation from a stationary image due to the way compression works - my understanding as an amateur is it saves the full frame in what's known as key frames and then tends to reference areas of the picture that are the same in subsequent frames to save saving all the detail again. No real change as when stationary = almost the whole picture referenced until next key frame = very little data saved = little to no compression effects . In my opinion, upping the bit rate for less challenging situations certainly can help the overall sharpness of a picture when moving. However, there is a law of diminishing returns here as in low detail situations the extra bit rate may have little to no visible effect on sharpness. That said, GoPros now use 60mbs (70mbs in protune - believe 4K and h265 compression). Yi+ (admittedly in 4K) up to 135mbs in h 264. However, storage considerations for continuous recording, mean you're probably unlikely to ever see that in dashcam.

I believe GoPros used to use 25mbs (Hero 2), so that isn't necessarily a bad rate. 50mbs is certainly better though and Y265 helps achieve that without additional storage requirements. Of course y265 @ 50mbs would be better still, but again it's storage and diminishing returns in ordinary low detail situations that dictates what many manufacturers are going to set.
 
Last edited:
These frame extracts show the potential of using a high 60Mbps bitrate with a very fast lens, compared to 24Mbps and an F1.8 lens in the A119.

The A119 has reduced the building stonework to an almost textureless grey surface, and there is loss of detail in both the road and the grass verges.
1525167548539.png
1525167594421.png


The fast lens and high bitrate also help to capture number plates on passing cars in high detail scenes like this:
1525168182346.png
1525168246636.png
 
Back
Top