1080p 60fps at Night?

[TonyM]

All the dashcams I have tried that can record 1080p at 60fps seem to use an effective 30fps at night by frame-doubling, even those with a Starvis IMX291 sensor.

Are there any dashcams that can record true 60fps at night, and do they offer any benefit over 30fps video?

 

[jokiin]

they don't have to frame double, depends how the firmware is configured

 

[TonyM]

they don't have to frame double, depends how the firmware is configured

I know they don't have to. Are there any with the FW configured to record 60fps?

 

[jokiin]

Plenty don't do it, the reason it is normally done though is 60fps and low light don't go together that well

 

[estore009]

low light don't go together that well

+1

 

[Nigel]

I know they don't have to. Are there any with the FW configured to record 60fps?

Only one I know of is the Gitup Git2, and that shows the problem, when it is dark you get a very dark image when using 60fps!

Although with newer Starvis sensors, I guess it doesn't make much difference since under normal lighting they probably use shutter speeds above 1/120th to minimise motion blur whichever fps they are using. You would only see the difference when driving at night without headlights.

 

[TonyM]

Although with newer Starvis sensors.... You would only see the difference when driving at night without headlights.

Typically, we do not drive at night without headlights! So perhaps a Starvis sensor can record at night using a 'shutter speed' faster than 1/60s and still maintain a decent exposure?

 

[kamkar]

Yeah with 60 FPS you only have half the time to capture light on the sensor, cuz dashcams can and will drop to 1/30 second exposure, even the starvis ones,
we really need much better sensors.

 

[Dashmellow]

Yeah with 60 FPS you only have half the time to capture light on the sensor, cuz dashcams can and will drop to 1/30 second exposure, even the starvis ones,
we really need much better sensors.

Better sensors are good but dash cams really need bigger ones with bigger pixels.

Canon's 35MMFHDXS with its 19μm x 19μm pixels is showing the way forward.....ultra high sensitivity @ 100 fps! Hopefully this technology will eventually work its way into a smaller sensor, say 1 inch and maybe some enterprising manufacturer will finally produce a dash cam with such a sensor.

 

[TonyM]

Different cameras. Each with Starvis sensors. Does the faster frame rate make enough difference?

Dashcam 1080p60 (double-frame, effective 30fps)


Action Camera 1080p60 (true 60fps)

 

[Nigel]

It improves readability, but increases noise a little.

I suspect the engineers are sitting at a desk during the testing trying to get the best quality noise free images, and motion blur doesn't get considered because they are not moving!

 

[TonyM]

It improves readability, but increases noise a little.

I'm not sure if that makes it any better? Low noise video is more appealing to look at.

 

[TonyM]

Since the dash cam is doubling frames, I think a better comparison would be, dash cam at 30fps and the action camera at 60fps.

I think the dashcam IS recording at 30fps. It simply doubles each frame when it's set to create a 60fps video. I'll check later.

 

[Nigel]

I never use a dashcam at 60fps because I always want it to use maximum resolution, and at maximum resolution they never support 60fps!
Using maximum resolution will reduce motion blur on plates all day long, not just at midnight, due to being able to read the plates further away.

 

[TonyM]

I never use a dashcam at 60fps because I always want it to use maximum resolution, and at maximum resolution they never support 60fps!

Many cameras use the IMX291 which is not only one of the best sensors for dashcams in low light, but also has a maximum resolution of 1080p at both 30fps and 60fps.

Using maximum resolution will reduce motion blur on plates all day long, not just at midnight, due to being able to read the plates further away.

I like using 4K, but sometimes it's not sensitive enough to pick up much of anything at distance in low light to take advantage of that higher resolution.

 

[TonyM]

So, the general consensus seems to be that 60fps is no good at night. I've tried it a few times and had mixed results, but I'm not convinced it's going to work well enough in all conditions.

 

[jokiin]

So, the general consensus seems to be that 60fps is no good at night. I've tried it a few times and had mixed results, but I'm not convinced it's going to work well enough in all conditions.

60fps can be ok at night (with some work on the firmware) but all things being equal I think 30fps is still going to have an advantage

 

[Nigel]

60fps can be ok at night (with some work on the firmware) but all things being equal I think 30fps is still going to have an advantage

30fps can only give an advantage if you take advantage of the extra time available which allows a longer exposure. However longer exposure = more motion blur, and on a dashcam motion blur is normally a negative.

If I'm making a movie to be played at 30fps then having some motion blur is good, it smooths the movement between frames so that you don't notice the 30fps frame updates, so a movie camera should use low shutter speeds, but for a dashcam you want fast shutter speeds to remove the motion blur from plates.

So it is a bit odd that my dashcams use 30fps and have a lot of motion blur in the dark while my movie (action) cam uses 60fps with low motion blur, and has much more readable plates!

Filmed this on the A129 Pro prototype at 9:05 PM on a cloudy night, movement between frames is nicely blurred, ideal for movies, and much better than the Gitup Git2 action camera that I used for the sound and which had almost no motion blur at 60fps and had numbers that were clearly readable, while they are totally unreadable on the A129 video. But the A129 is a dashcam, it should capture those numbers, while the Gitup Git2 is my movie camera which i want some nice blur from, they are the wrong way around!

On a positive, the A129 Pro did manage to make things look almost like a cloudy midday rather than a cloudy 9:05 PM, would be hard to tell it wasn't midday if it wasn't for the blinding LED headlights and the cab full of flames - driver must have been a bit hot! The Gitup Git2 video on the other hand is dark and blurry, quite poor quality, but I can read the engine number from it!

So, it is a compromise, quality v readability, but I'm still not convince my dashcam and action camera are the right way around!

(No, it wasn't mounted on a drone, I was just playing with the zoom control!)

 

[TonyM]

@Nigel you're the only person I know who uses a dashcam on a tripod as an action camera. I might try that one day

 

[Nigel]

@Nigel you're the only person I know who uses a dashcam on a tripod as an action camera. I might try that one day

It is good for low light, you don't get Starvis sensors on action cameras! If it was daylight then I would use the Gitup F1, but it was too dark for a good result from that, and the A129 Pro still gives 4K allowing use of zoom. Also, the A129 fits nicely in my GoPro frame, very easy to mount on a tripod.