DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Does Mavic Shoot Video or Watercolor Paintings? Don’t settle. Get Involved!

Perhaps I was overzealous by using the ND16. I do think that footage is a little dark. I did use the Polar Pro app though. I had a shutter speed of 1250 with no filter, and with a target shutter of 1/60 the Polar Pro IOS App advised that ND16 was the correct filter to use. p.s. I havent tried an ND Filter with Auto on! Will try that and see what happens.
See what happens if you stop trying to hit 1/60th. Anything less than 1/1000th is OK. Goes against all the experts advice for getting smooth footage I know but for me, at the moment at least, it's fixed the issue that this thread was set up to address.
 
I just completed another controlled test on my 3rd Mavic unit (refer to post#683 and #421 for previous tests). This test was aimed at assessing the effect of an ND filter and lower shutter speeds.
The results were -- shall I say -- reasonable??!!!

Summary:
Test Scenario 1
* Taco ND8, manual Iso100, 1/60s, -1EV, color=none
* Watercolor or mushy image was evident with sharpness 0,-1,-2.
* At sharpness +1, it looked good! (NOTE: +1 looks too edgy with higher frame rates)

Test Scenario 2
* No Filter, Auto (Iso=200, 1/100s, 0EV, color=none)
* Watercolor or mushy image was evident with sharpness -1,-2. Sharpness 0 was pretty good. Sharpness +1 was excellent, though a little edgy.

Mavic Pro Imaging Conclusions
* If you have issues getting razor sharp focus edge-2-edge across screen with high shutter speeds, iso100, Sharp=0, color=none -> RMA the unit
* If you are shooting with low shutter speed; less than 100/120 (ie. ND filter or low light), use sharpness 0/+1. For fast movements near objects, I suggest +1, else it will blur into a mushy oblivion!
* My experience with DJI units (P3P, I1,etc.) is shutter speed 160 or less is fine for enough "motion blur" effect. If you want to add more blur, most NLE's allow you to add. Focus on capturing a beautiful image!!
* My personal opinion is the 2k30 looks better than the 4k30. DJI even says (behind the scenes) the sensor is optimized for 2k30. But looks about the same on YouTube, so.....

Sorry, I did not get a chance to do a controlled test with the color filters, but do have some ART footage from 150ft that looked good... :p

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
4k30, Sharp-1, iso200, 1/100s
4k30, Sharp-2, iso200, 1/100s
4k30, Sharp+1, iso200, 1/100s

EDIT: corrected fps reference to shutter speed in 1/s
 
Last edited:
  • Like
Reactions: Willik
I just completed another controlled test on my 3rd Mavic unit (refer to post#683 and #421 for previous tests). This test was aimed at assessing the effect of an ND filter and lower shutter speeds.
The results were -- shall I say -- reasonable??!!!

Summary:
Test Scenario 1
* Taco ND8, manual Iso100, 60fps, -1EV, color=none
* Watercolor or mushy image was evident with sharpness 0,-1,-2.
* At sharpness +1, it looked good! (NOTE: +1 looks too edgy with higher frame rates)

Test Scenario 2
* No Filter, Auto (Iso=200, 100fps, 0EV, color=none)
* Watercolor or mushy image was evident with sharpness -1,-2. Sharpness 0 was pretty good. Sharpness +1 was excellent, though a little edgy.

Mavic Pro Imaging Conclusions
* If you have issues getting razor sharp focus edge-2-edge across screen with high shutter speeds, iso100, Sharp=0, color=none -> RMA the unit
* If you are shooting with low shutter speed; less than 100/120 (ie. ND filter or low light), use sharpness 0/+1. For fast movements near objects, I suggest +1, else it will blur into a mushy oblivion!
* My experience with DJI units (P3P, I1,etc.) is shutter speed 160 or less is fine for enough "motion blur" effect. If you want to add more blur, most NLE's allow you to add. Focus on capturing a beautiful image!!
* My personal opinion is the 2k30 looks better than the 4k30. DJI even says (behind the scenes) the sensor is optimized for 2k30. But looks about the same on YouTube, so.....

Sorry, I did not get a chance to do a controlled test with the color filters, but do have some ART footage from 150ft that looked good... :p

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
4k30, Sharp-1, iso200, 100fps
4k30, Sharp-2, iso200, 100fps
4k30, Sharp+1, iso200, 100fps

Thanks for doing those tests. It's a very good observation to make with regards to bumping sharpness up to +1 when using an ND filter. This comes back to theory that the issue is software/algo related, at least hopefully for the majority of us.

I'm looking forward to getting an opportunity to replicate these tests with my Mavic and report back.

Thanks to everyone who has contributed!! I think this is good news right?
 
  • Like
Reactions: 2nd2non
Thanks for doing those tests. It's a very good observation to make with regards to bumping sharpness up to +1 when using an ND filter. This comes back to theory that the issue is software/algo related, at least hopefully for the majority of us.

I'm looking forward to getting an opportunity to replicate these tests with my Mavic and report back.

Thanks to everyone who has contributed!! I think this is good news right?
Should we more accurately be saying to bump sharpness when we have long shutter speeds? I'm anticipating replies from people thinking it's the actual ND is the cause.
 
Bumping sharpening with high shutter speed looks to edgy and artificial for my taste. Perhaps there are slight tolerances in sharpening between mavic units, but before this assumption is made, a person should check edge to edge sharpness with high shutter first. There's probably a hardware issue if you're having to set sharpness to +1 just to make the overall image look sharp at high shutter rates
 
Should we more accurately be saying to bump sharpness when we have long shutter speeds? I'm anticipating replies from people thinking it's the actual ND is the cause.

Yes, you're right. But don't the long shutter speeds and ND filters go hand in hand? I suppose night flying is the exception.
 
I completed test on my Mavic. This test was aimed at assessing the effect of an ND filter, resolution 2.7 vs 4 and D-log vs None

Summary:
Test Scenario 1
* Taco ND8, 4K, Auto Iso100, 24fps, 0EV, color=d-log SS=1/200, -1,0,0
* Image OK
* At sharpness 0 and color=none it looked good!
* 2.7 K also OK

Test Scenario 2
* No Filter, 4 K, Auto Iso=100, 24fps, 0EV, color=none/d-log, SS=1/2000, -1,0,0
* Image OK
* At sharpness 0 and color=none it looked good!
* 2.7 K also OK

Test Scenario 3
* Taco ND32, 4K, Auto Iso100, 24fps, 0EV, color=d-log SS=1/50, -1,0,0
* Watercolor or mushy image was evident and at d-log image is worse muddled
* At sharpness 0 and color=none also Watercolor or mushy image
* 2.7 K color none and d-log Watercolor or mushy image was evident but slightly less


Conclusions

The reason smudged image is too little light and too slow shutter speed when you put too strong ND filter. Shutter speed should not be reduced below 1/100
 
Here's footage, Litchi waypoint repeated 12 times, lumetri color on d-log sections shots, resolution 720

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
I completed test on my Mavic. This test was aimed at assessing the effect of an ND filter, resolution 2.7 vs 4 and D-log vs None

Summary:
Test Scenario 1
* Taco ND8, 4K, Auto Iso100, 24fps, 0EV, color=d-log SS=1/200, -1,0,0
* Image OK
* At sharpness 0 and color=none it looked good!
* 2.7 K also OK

Test Scenario 2
* No Filter, 4 K, Auto Iso=100, 24fps, 0EV, color=none/d-log, SS=1/2000, -1,0,0
* Image OK
* At sharpness 0 and color=none it looked good!
* 2.7 K also OK

Test Scenario 3
* Taco ND32, 4K, Auto Iso100, 24fps, 0EV, color=d-log SS=1/50, -1,0,0
* Watercolor or mushy image was evident and at d-log image is worse muddled
* At sharpness 0 and color=none also Watercolor or mushy image
* 2.7 K color none and d-log Watercolor or mushy image was evident but slightly less


Conclusions

The reason smudged image is too little light and too slow shutter speed when you put too strong ND filter. Shutter speed should not be reduced below 1/100

Great tests mate, thanks for sharing the results with us.

I agree that too little light entering the sensor is the cause of my (own) issue.

So, is there anyone out there that can rock slow shutter speeds and not experience mush?
 
Conclusions

The reason smudged image is too little light and too slow shutter speed when you put too strong ND filter. Shutter speed should not be reduced below 1/100
Thank you for taking the time to do this and for sharing the results!

But I do not agree on the conclusion. A framerate of 1x/2x the shutter speed is common and usually no issue. If the video image is exposed well at 30 fps with 1/60 s each frame, then there is no light missing. We use ND filters to deal with too much light ...

One might say now that the reason for mushy video is too slow shutter speed / too strong ND filter. But this is still because DJI internally messes it up with this setup.
 
Parrot Bebop 2's videos look like water color paintings. It was the biggest turn off for me when they launched their drone.

I bet the next Mavic will have a better camera.
 
Looks like several are getting similar results with testing.
For me its about optimizing the capabilities of the Mavic as it currently is.
We'll have to see what a future firmware may bring, but for now I'm able to make it work
 
Parrot Bebop 2's videos look like water color paintings. It was the biggest turn off for me when they launched their drone.

I bet the next Mavic will have a better camera.

I don't even think it's necessarily a camera issue. It's the marketing department's desire to throw a "4K video" label on computing hardware not really capable of 4K video. All these action camera and drone companies are pushing to up the resolution because that's easily marketable with a label everyone can understand, but they're neglecting what quality those images actually are, because there's no simple label that the marketing department can throw out there for superior image quality.
 
  • Like
Reactions: Robbyg
I completed test on my Mavic. This test was aimed at assessing the effect of an ND filter, resolution 2.7 vs 4 and D-log vs None

Summary:
Test Scenario 1
* Taco ND8, 4K, Auto Iso100, 24fps, 0EV, color=d-log SS=1/200, -1,0,0
* Image OK
* At sharpness 0 and color=none it looked good!
* 2.7 K also OK

Test Scenario 2
* No Filter, 4 K, Auto Iso=100, 24fps, 0EV, color=none/d-log, SS=1/2000, -1,0,0
* Image OK
* At sharpness 0 and color=none it looked good!
* 2.7 K also OK

Test Scenario 3
* Taco ND32, 4K, Auto Iso100, 24fps, 0EV, color=d-log SS=1/50, -1,0,0
* Watercolor or mushy image was evident and at d-log image is worse muddled
* At sharpness 0 and color=none also Watercolor or mushy image
* 2.7 K color none and d-log Watercolor or mushy image was evident but slightly less


Conclusions

The reason smudged image is too little light and too slow shutter speed when you put too strong ND filter. Shutter speed should not be reduced below 1/100

Can this be a function of the limit of a gimbal to stabilize things like aircraft vibration, turbulence, and the perturbation cause by wind forces?

In other words, there is a limit in hand-held photography that is difficult to go under, in terms of shutter speed. Each individual is different, and various forms of optical stabilization can only accomplish so much. This is in the hands of an accomplished photographer. 150 feet up, buffeted by wind, is a completely different set of challenges.

So, what is the lower limit of shutter speed, in terms of sharpness, from a common sense photography perspective? 1/60th seems super low in that environment, period.
 
Can this be a function of the limit of a gimbal to stabilize things like aircraft vibration, turbulence, and the perturbation cause by wind forces?

In other words, there is a limit in hand-held photography that is difficult to go under, in terms of shutter speed. Each individual is different, and various forms of optical stabilization can only accomplish so much. This is in the hands of an accomplished photographer. 150 feet up, buffeted by wind, is a completely different set of challenges.

So, what is the lower limit of shutter speed, in terms of sharpness, from a common sense photography perspective? 1/60th seems super low in that environment, period.


Plenty of us have done multi-second exposures without issue. This thread is about video issues though, so the entire purpose of going with a lower shutter speed is to create some amount of blur.
 
I just completed another controlled test on my 3rd Mavic unit (refer to post#683 and #421 for previous tests). This test was aimed at assessing the effect of an ND filter and lower shutter speeds.
The results were -- shall I say -- reasonable??!!!

Summary:
Test Scenario 1
* Taco ND8, manual Iso100, 60fps, -1EV, color=none
* Watercolor or mushy image was evident with sharpness 0,-1,-2.
* At sharpness +1, it looked good! (NOTE: +1 looks too edgy with higher frame rates)

Test Scenario 2
* No Filter, Auto (Iso=200, 100fps, 0EV, color=none)
* Watercolor or mushy image was evident with sharpness -1,-2. Sharpness 0 was pretty good. Sharpness +1 was excellent, though a little edgy.

Mavic Pro Imaging Conclusions
* If you have issues getting razor sharp focus edge-2-edge across screen with high shutter speeds, iso100, Sharp=0, color=none -> RMA the unit
* If you are shooting with low shutter speed; less than 100/120 (ie. ND filter or low light), use sharpness 0/+1. For fast movements near objects, I suggest +1, else it will blur into a mushy oblivion!
* My experience with DJI units (P3P, I1,etc.) is shutter speed 160 or less is fine for enough "motion blur" effect. If you want to add more blur, most NLE's allow you to add. Focus on capturing a beautiful image!!
* My personal opinion is the 2k30 looks better than the 4k30. DJI even says (behind the scenes) the sensor is optimized for 2k30. But looks about the same on YouTube, so.....

Sorry, I did not get a chance to do a controlled test with the color filters, but do have some ART footage from 150ft that looked good... :p

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
4k30, Sharp-1, iso200, 100fps
4k30, Sharp-2, iso200, 100fps
4k30, Sharp+1, iso200, 100fps
Thanks for posting those. That's interesting. On my Mavic (Art .-1,-1,0) The sharpness at minus 1 is almost too much. I'll try bumping it up to zero and plus 1. Just from I originally settled on RtPhilip Blooms
Thank you for taking the time to do this and for sharing the results!

But I do not agree on the conclusion. A framerate of 1x/2x the shutter speed is common and usually no issue. If the video image is exposed well at 30 fps with 1/60 s each frame, then there is no light missing. We use ND filters to deal with too much light ...

One might say now that the reason for mushy video is too slow shutter speed / too strong ND filter. But this is still because DJI internally messes it up with this setup.
Thank you for taking the time to do this and for sharing the results!

But I do not agree on the conclusion. A framerate of 1x/2x the shutter speed is common and usually no issue. If the video image is exposed well at 30 fps with 1/60 s each frame, then there is no light missing. We use ND filters to deal with too much light ...

One might say now that the reason for mushy video is too slow shutter speed / too strong ND filter. But this is still because DJI internally messes it up with this setup.
I agree. The issue is with the firmware or software. We should be able to reduce shutter speeds down to 2 x Frame rate with ND filters to get a cinematic look. We should also be able to get a setup nearly perfect then simply introduce a bit more sharpness by dialling it in. Same goes with contrast, saturation, ISO & exposure. & Having got that perfect setup we should be able to trust that they will work consistently and reliably. The problem DJI have at present is that none of this is possible with the existing FW and/or SW. Dial in one more stop of sharpness and it completely messes everything else up. It doesn't adjust to compensate, it just messes everything up. It's unpredictable so you can't reproduce issues so therefore as a user, you can't fix it. If there was a work around that worked for all, one of us would've nailed it by now but all our results are so random. Even when I think I've found a fix for my Mavic, next flight, apply the fix and it will not work. I'm beginning to think that this thing really is borderline not fit for purpose. You certainly can't trust that it will yield reliable results.
 
  • Like
Reactions: peros550
If this is a consumer drone then i would tell DJI not to call it "Mavic pro" and just "Mavic" lol . And if the sweet spot is 2.7@30fps then its just like a Phantom 3 standard. The price shouldn't be $999. It should be $600. Just my 2 cents
 
If this is a consumer drone then i would tell DJI not to call it "Mavic pro" and just "Mavic" lol . And if the sweet spot is 2.7@30fps then its just like a Phantom 3 standard. The price shouldn't be $999. It should be $600. Just my 2 cents
I shoot at 4k and I like what I see.
 
If this is a consumer drone then i would tell DJI not to call it "Mavic pro" and just "Mavic" lol . And if the sweet spot is 2.7@30fps then its just like a Phantom 3 standard. The price shouldn't be $999. It should be $600. Just my 2 cents
Quite right. I find it hard to agree with comments referring to how this thing is so cheap & what D'you expect for the money Etc. It's not cheap. It's an expensive piece of kit but they've obviously skimped on the software development. Just hope that, as some posts suggest, DJI are working on a fix
 

DJI Drone Deals

New Threads

Forum statistics

Threads
134,565
Messages
1,596,302
Members
163,065
Latest member
kyle465
Want to Remove this Ad? Simply login or create a free account