Hi !
I made some effort to gt a feeling if 4k delivers more Details then 2,7k.
If that would NOT be the case then i better go for 2,7k as the "per pixel" compression ratio is better.
I even had "compression sharpness pumping" in my maiden flight bt i am very picky aout it (60% grass, 30% trees, 10% cloudy sky, bad lighting).
I did NOT have compression sharpness pumping at 2,7k yet shooting similar clips.
(There is ALWAYS some due to I/P frames but if i dont notice them they are non of my business )
So i produced a simple test setup:
Siemensstern & USAF 1952 Printouts at 5m distance, Mavic centered towards the targets, GOOD Lighting, Focus done in 2x Mag while doing 1080p streaming on a Retina 10" Android Tablet, triple checked that focus wont get better, THEN left it unchanged.
Took 4second clips in 4k Cinematic@24FPS, 4k@30FPS, 2,7k@30FPS, 1080p@30FPS, 720p@30FPS, JPG, DNG.
Mavic set to -2,0,0 as DJI does a horrible sharpening job on the data otherwise IMO.
I interprete the results in the way that i will continue to shoot in 2,7k and that i will ALWAYS shoot JPG + DNG as DNG not only has its "usual" advantages but also a very noticable Detail advantage.
Numbers:
Siemensstern ( Moiré radius / Total radius, smaller is better ):
USAF 1951 (First number Reslution Group, -1 BETTER than -2, Second number subfield, bigger is better)
This was far away from scientific and it is valid for MY Mavic but having seen quite some clips now i doubt that the variance is VERY big.
I wanted to imagine that 4k contains more Detail but it wasnt in the numbers.
I went into a little more (exhausting) Detail in my German post but thats way to much for this place, send me a PM if interested.
And remember this was done for MYSELF to judge what to do, your decision probably variees
Greetings,
Ender
P.S. Todo list: DNG profile (actually part of my current job is writing a DNG profiler) and ICC profiles for some standard lightings, dynamic range measurement JPG vs. DNG, Multiexposure effeciency, ISO noise profile, but at first: GETTING GOOD WEATHER TO FLY
I made some effort to gt a feeling if 4k delivers more Details then 2,7k.
If that would NOT be the case then i better go for 2,7k as the "per pixel" compression ratio is better.
I even had "compression sharpness pumping" in my maiden flight bt i am very picky aout it (60% grass, 30% trees, 10% cloudy sky, bad lighting).
I did NOT have compression sharpness pumping at 2,7k yet shooting similar clips.
(There is ALWAYS some due to I/P frames but if i dont notice them they are non of my business )
So i produced a simple test setup:
Siemensstern & USAF 1952 Printouts at 5m distance, Mavic centered towards the targets, GOOD Lighting, Focus done in 2x Mag while doing 1080p streaming on a Retina 10" Android Tablet, triple checked that focus wont get better, THEN left it unchanged.
Took 4second clips in 4k Cinematic@24FPS, 4k@30FPS, 2,7k@30FPS, 1080p@30FPS, 720p@30FPS, JPG, DNG.
Mavic set to -2,0,0 as DJI does a horrible sharpening job on the data otherwise IMO.
I interprete the results in the way that i will continue to shoot in 2,7k and that i will ALWAYS shoot JPG + DNG as DNG not only has its "usual" advantages but also a very noticable Detail advantage.
Numbers:
Siemensstern ( Moiré radius / Total radius, smaller is better ):
Code:
4kCinematik@24 FPS 0,42
4k@30 0,43
2,7k@30 0,43
1080p@30 0,61
720p@30 0,68
JPG 0,36
DNG 0,28
USAF 1951 (First number Reslution Group, -1 BETTER than -2, Second number subfield, bigger is better)
Code:
4kCinematik@24 FPS -1;1
4k@30 -1;1
2,7k@30 -1;1
1080p@30 -2;2
720p@30 -2;1
JPG -1;2
DNG -1;3
This was far away from scientific and it is valid for MY Mavic but having seen quite some clips now i doubt that the variance is VERY big.
I wanted to imagine that 4k contains more Detail but it wasnt in the numbers.
I went into a little more (exhausting) Detail in my German post but thats way to much for this place, send me a PM if interested.
And remember this was done for MYSELF to judge what to do, your decision probably variees
Greetings,
Ender
P.S. Todo list: DNG profile (actually part of my current job is writing a DNG profiler) and ICC profiles for some standard lightings, dynamic range measurement JPG vs. DNG, Multiexposure effeciency, ISO noise profile, but at first: GETTING GOOD WEATHER TO FLY