DJI Air 2S Twilight and Night test
1178 13 2021-5-7
Uploading and Loding Picture ...(0/1)
o(^-^)o
sduck
lvl.4
Flight distance : 4315098 ft
United States
Offline

Just a quick test to see how Twilight and Night videos look with it. Flying over Brentwood, Tennessee (soon after a storm!). Shot in 5.4K, auto settings, the default "Natural" color setting, no color grading. Shot in Cine mode, sped up 2X in Final Cut Pro.


2021-5-7
Use props
DAFlys
Captain
Flight distance : 312090263 ft
  • >>>
United Kingdom
Offline

Looks great.  
2021-5-7
Use props
DGCA3
lvl.4
Flight distance : 465062 ft
United States
Offline

Looks good.
2021-5-7
Use props
Bob-Mini-2
lvl.4
Flight distance : 73114 ft
United States
Offline

sduck, Nicely done - Twilight - one of my favorite times.  It gives me something to go for when I finally get to doing some real flying.
Thanks for sharing.
2021-5-7
Use props
K6CCC
lvl.4
Flight distance : 651683 ft
United States
Offline

Nice video.  I also love flying at twilight.  Making me jealous with being able to fly that far in the city.  Can't do that with my original Mini.
2021-5-7
Use props
DGCA3
lvl.4
Flight distance : 465062 ft
United States
Offline

K6CCC Posted at 5-7 19:29
Nice video.  I also love flying at twilight.  Making me jealous with being able to fly that far in the city.  Can't do that with my original Mini.

You'll love Ocusync, should you choose to upgrade down the road.
2021-5-7
Use props
DJI Gamora
Administrator

Offline

Great video! Keep it up and thanks for sharing.
2021-5-8
Use props
Frontliner 2020
lvl.4
Philippines
Offline

beautiful video...
2021-5-8
Use props
DowntownRDB
Core User of DJI
Flight distance : 1722 ft
  • >>>
United States
Online

Great footage!    Have a great weekend.
2021-5-8
Use props
K6CCC
lvl.4
Flight distance : 651683 ft
United States
Offline

DGCA3 Posted at 5-7 23:11
You'll love Ocusync, should you choose to upgrade down the road.

I know.  I expect to at some point.  I'm kinda waiting for the Mavic 3 Pro - when it comes out.  Although the Air 2S is sure tempting...
2021-5-8
Use props
Montfrooij
Captain
Flight distance : 2560453 ft
  • >>>
Netherlands
Offline

Looks very nice!
2021-5-10
Use props
Lops
lvl.3
Flight distance : 747772 ft
  • >>>
Germany
Offline

be honest and confess that you put heavy noise filters, because the air 2s is not better with respect to low light performance
2021-5-11
Use props
sduck
lvl.4
Flight distance : 4315098 ft
United States
Offline

Lops Posted at 5-11 07:31
be honest and confess that you put heavy noise filters, because the air 2s is not better with respect to low light performance

Sorry, being totally honest here - no filters or anything. I didn't touch the color palette or anything. It's just the straight 5.4K h265 files from the drone card dropped into Final Cut Pro, sped up 2x, and then downscaled to 4K using Apple Compressor (because youtube/vimeo can't really handle 5k yet).

What you can see is a pretty drastic loss of details/resolution in the night part of the video - which was shot only about 10-15 minutes after the first part - it wasn't even completely dark yet. Same auto settings, so the iso had to go up substantially, and hence the loss of resolution.
2021-5-11
Use props
Lops
lvl.3
Flight distance : 747772 ft
  • >>>
Germany
Offline

sduck Posted at 5-11 17:57
Sorry, being totally honest here - no filters or anything. I didn't touch the color palette or anything. It's just the straight 5.4K h265 files from the drone card dropped into Final Cut Pro, sped up 2x, and then downscaled to 4K using Apple Compressor (because youtube/vimeo can't really handle 5k yet).

What you can see is a pretty drastic loss of details/resolution in the night part of the video - which was shot only about 10-15 minutes after the first part - it wasn't even completely dark yet. Same auto settings, so the iso had to go up substantially, and hence the loss of resolution.

Ok, accepting your statement I suggest, that the downsampling process caused a denoise effect because good editing software uses interpolation algorithms for resampling which calculates the resampled pixel values from local pixel clusters (e.g. 3x3) which is leading to a certain noise reduction. Might be a good idea to take highest resolution under low light conditions and then just downsample...have to check that. And saying that, I realize that you also speeded up by factor two and the same argument as with downsampling applies here too - not in the image-spatial but in the image sequence temporal direction. That means by downsampling and speeding up you did a double-strike against the noise.

And btw. I like your video ;)


2021-5-13
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules