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

Poor image quality

so after sending in my mini4p i got a new mini4p back (different SN, nice). i am getting now tired of the image comparison and i know you too :)
summary:
the new mini4p has less edge distortions - great (only bottom right seems a little worse)
however, the image quality is still disappointing, considering all the hype i am reading about, because i still l rate it worse than the mini2

my conclusion from best to worse:
mini2 jpeg ( i have not tried mini 2 raw)
mini4p dng
mini4p jpg (only slightly worse than DNG)

i use rawtherapee to look at the dng file

for the mini4p i used the AF mode and made sure my subject was selected green.

i still challenge anyone to show me a clear mini2 vs mini4 comparison of jpegs where the mini4 is clearly better. this should not include low light images because we should believe the f1.7 and larger sensor at same 12MP produces better low light images.
yes, the mini4p has nice features overall but if anyone thinks he/she needs to buy the way more expensive mini4 or mini3 over the mini2 because of better still images i would say forget it.
On the left, the image is oversharpened. This is not due to the optics, but to the drone's digital processing, and it's not pretty.
 
This just looks like a post processing shaprness difference. M4P has a adjustable shaprness settings. Have you tried adjusting it?
 
Jpg or dng shouldn't make any difference to what you've complained about.

Selected green ??
What are you talking about?

I still challenge you to make the effort to shoot proper images that make it easy to see if the cameras have the problems you complain about.

Yes, you would, but should anyone accept your opinion?
no, obviously not, it is a free country (at least for now). however, maybe other people take a critical look too and if they agree maybe DJI will react to it. my statistic is 3 drones (2x mini2, 1x mini3, 2x mini4p). for me it is simple, i am just disappointed, here is a drone with amazing capabilities but fails in the key feature
 
This just looks like a post processing shaprness difference. M4P has a adjustable shaprness settings. Have you tried adjusting it?
no, i have not adjusted anything - i shared the exif data before. i played with the sharpness in the DNG file but makes no difference. look, in the end it should not depend on some tweaking here and there, many reviewers say the image quality is greatly superior over mini2 if that is the case i should not to see worse images with DJI's default settings. would be nice if some of he professional photo sites like DPREVIEW would take a look.
 
This just looks like a post processing shaprness difference. M4P has a adjustable shaprness settings. Have you tried adjusting it?
valid point. you could be right but i was not really successful in DNG file but it improved a little. i think but it does not seem to be available for stills in the drone per manual

1721347175647.png
 
in-spite my issues i want to compliment DJI for reaching out to me after i received the replacement drone to see how my experience was. obviously i shared my experience mention above but they do seem to care and i had almost 1hour discussion with them. they promised my to have an image expert get in touch with me - will see.
 
i did send mini2 and mini4p jpeg and DNG to DJI for evaluation. here the reply from DJI:

" According to the assessment from the DJI senior technical team, the pictures shooter by your DJI Mini 4 Pro drone are normal, which means they are good.

The deviation may be more obvious in the DJI Mini 3 Pro/DJI Mini 4 Pro lens due to the larger aperture and shallower depth of field. We recommend choosing the central position for focusing to balance the resolution of the peripheral positions of the view."

i did use the central focus position.
so my personal conclusion: if you do mainly daylight (at least not very low light) still shooting and you do not need the fancy features of the m4p stick with mini2.
i find it still interesting that the image quality issue/comparison was not highlighted more in other reviews or postings
 
i find it still interesting that the image quality issue/comparison was not highlighted more in other reviews or postings

Likely because, as in this thread, others don't share your criticism. The rest of us think the M4P images are superior.

Since there's a healthy share of subjectivity in such analysis, arguing further about it seems a waste of time.
 
Likely because, as in this thread, others don't share your criticism. The rest of us think the M4P images are superior.

Since there's a healthy share of subjectivity in such analysis, arguing further about it seems a waste of time.
thanks for your comment, you are probably right but to me and 2 experienced photographer friends it does not look like a subjective issue. m4p i worse. i wish there would be a way to reduce the aperture to f2.8 for good light use
 
thanks for your comment, you are probably right but to me and 2 experienced photographer friends it does not look like a subjective issue. m4p i worse. i wish there would be a way to reduce the aperture to f2.8 for good light use

So the subjective opinion of 3 people is somehow objective? There are at least three people in this thread that shoot professionally that have a different opinion.

I do have sympathy for you that it just doesn't meet your needs. Regardless, judging the quality of imagery is always a subjective endeavor.
 
thanks for your comment, you are probably right but to me and 2 experienced photographer friends it does not look like a subjective issue. m4p i worse. i wish there would be a way to reduce the aperture to f2.8 for good light use
Well, you could've tried what most of the rest of us do with fixed aperture drone cameras.... drop the EV to negative values and ETTL (Expose To The Left)... put an ND filter in front of the lens... go manual and speed up the shutter.

You can't compare an f/2.8 with an f/1.7 aperture and expect parity, if you want that: you have to alter your settings before you fire the shutter - I would have expected your experienced photographer friends to have pointed this out.

This is something I've had to do myself when using a Mavic 2 zoom (fixed f/2.8) and a mini 3 pro (fixed f/1.7) during the same shoot. To balance the output from two distinct and different lenses requires alterations in camera settings before you take the shot and then careful adjustment of the exposure variables in post, which takes time, attention to detail and patience.
 
Last edited:
Well, you could've tried what most of the rest of us do with fixed aperture drone cameras.... drop the EV to negative values and ETTL (Expose To The Left)... put an ND filter in front of the lens... go manual and speed up the shutter.

You can't compare an f/2.8 with an f/1.7 aperture and expect parity, if you want that: you have to alter your settings before you fire the shutter - I would have expected your experienced photographer friends to have pointed this out.

This is something I've had to do myself when using a Mavic 2 zoom (fixed f/2.8) and a mini 3 pro (fixed f/1.7) during the same shoot. To balance the output from two distinct and different lenses requires alterations in camera settings before you take the shot and then careful adjustment of the exposure variables in post, which takes time, attention to detail and patience.
all the changes of these settings do not change the physics of the optical imaging as you well know. none of these settings remove corner distortions or changes DOF AND it should not be my job to have to tweak out of the camera images to make them better with all post processing. at least the sharpness should be good to begin with - all the other parameters depend a lot on personal taste. But i did go a step further and for fun i modified it into to a f2.8 system and now as expected DOF is as it should be for a f2.8 and overall image is not worse but still not as good as mini2.
sure i compromise low light shooting (which is not a problem for me) but i should still have less noise than the mini2 due to the larger pixel area.

1722131838339.png
 
"... do not change the physics of optical imaging..."

The mini 2 has a 12 MP Bayer matrix sensor (approximately 2.4microns per pixel) behind a fixed aperture f/2.8 lens.

The mini 3/4 has a 12 mp quad Bayer matrix sensor (the equivalent of 0.6microns per pixel) behind a F/A f/1.7 lens.

They're as similar as chalk and cheese.
 
"... do not change the physics of optical imaging..."

The mini 2 has a 12 MP Bayer matrix sensor (approximately 2.4microns per pixel) behind a fixed aperture f/2.8 lens.

The mini 3/4 has a 12 mp quad Bayer matrix sensor (the equivalent of 0.6microns per pixel) behind a F/A f/1.7 lens.

They're as similar as chalk and cheese.
ok thanks for that detail, but i assume with the larger sensor for the mini3/4 of 1/1.3 there is more light per pixel even without the f1.7 or then a simple question which sensor you expect to deliver the better image quality.
based on info i have i estimate this. so mini3/4 has about 2x pixel area. anything wrong with that?

1722145333036.png
 
ok thanks for that detail, but i assume with the larger sensor for the mini3/4 of 1/1.3 there is more light per pixel even without the f1.7 or then a simple question which sensor you expect to deliver the better image quality.
based on info i have i estimate this. so mini3/4 has about 2x pixel area. anything wrong with that?

View attachment 176531
Photosite (pixel) size is all about light gathering. Do you recall film photography? Film grain, or ASA: the smaller the film grain - the less light each 'grain' is capable of reacting to. Larger grain films collect more light due to the larger surface area of the silver halide crystals. Reduce the surface area (increase the grain) and the exposure has to change to compensate. Digital sensors work in exactly the same basic way.

You will get a richer, more detailed image from a sensor with larger photosites, which is why pro cameras have physically larger sensors with correspondingly larger pixels (usually 2.4micron or slightly larger).

The aperture may well be more open (f/1.7) but the 'quarter size' pixel on the mini 3/4 is both less sensitive and of a significantly smaller area than the full size pixel on the mini 2.

When it boils down to it: different sensor. Different lens. Different aperture. Different image output.

If you're interested, I'll post two shots later. Both the same location, the same light conditions, taken within a few minutes of each other. One using a Bayer 12mp sensor @ f/2.8: the other a quad Bayer '48mp' sensor @ f/1.7. Both are a straight batch convert from DNG original to 8bit JPG. Neither has been edited. The difference is stark and immediately noticeable.
 
Last edited:
Photosite (pixel) size is all about light gathering. Do you recall film photography? Film grain, or ASA: the smaller the film grain - the less light each 'grain' is capable of reacting to. Larger grain films collect more light due to the larger surface area of the silver halide crystals. Reduce the surface area (increase the grain) and the exposure has to change to compensate. Digital sensors work in exactly the same basic way.

You will get a richer, more detailed image from a sensor with larger photosites, which is why pro cameras have physically larger sensors with correspondingly larger pixels (usually 2.4micron or slightly larger).

The aperture may well be more open (f/1.7) but the 'quarter size' pixel on the mini 3/4 is both less sensitive and of a significantly smaller area than the full size pixel on the mini 2.

When it boils down to it: different sensor. Different lens. Different aperture. Different image output.
ok my calculation gives the average pixel area. my understanding is that all the different colors from the bayer filter get combined to 1 pixel and that makes the pixel area for the mini3/4 sensor still bigger. i have no farther detail on the actual pixel area of the 2 drone types. from what you are saying is that you expect the mini3/4 sensor has a lower resolution do i understand this right? however, i do not understand your pixel area deduction.
 
Photosite (pixel) size is all about light gathering. Do you recall film photography? Film grain, or ASA: the smaller the film grain - the less light each 'grain' is capable of reacting to. Larger grain films collect more light due to the larger surface area of the silver halide crystals. Reduce the surface area (increase the grain) and the exposure has to change to compensate. Digital sensors work in exactly the same basic way.

You will get a richer, more detailed image from a sensor with larger photosites, which is why pro cameras have physically larger sensors with correspondingly larger pixels (usually 2.4micron or slightly larger).

The aperture may well be more open (f/1.7) but the 'quarter size' pixel on the mini 3/4 is both less sensitive and of a significantly smaller area than the full size pixel on the mini 2.

When it boils down to it: different sensor. Different lens. Different aperture. Different image output.

If you're interested, I'll post two shots later. Both the same location, the same light conditions, taken within a few minutes of each other. One using a Bayer 12mp sensor @ f/2.8: the other a quad Bayer '48mp' sensor @ f/1.7. Both are a straight batch convert from DNG original to 8bit JPG. Neither has been edited. The difference is stark and immediately noticeable.
ok from what i am reading the quad bayer sensor is not much better if at all in terms of image resolution and that is exactly what i am seeing - sadly, it actually is worse and even the 48MP is a wash. i think that mainly concludes the new sensor in the mini3/4 drones. due to the larger sensor there should still be lower noise and better low light behavior which i believe is a valid point. however, the f1.7 brings additional challenges to the lens system as i have seen with really bad corner distortions. ok, the replacement drone showed improvements in that regard which is good. so the possibility to perform in-situ HDR during video should also be beneficial but does not help the image resolution overall.
said all this i hope DJI is improving the camera quality because the drone with all its other features is really a gem.
 
Hi everyone,

I found myself here looking for a solution to my low image quality pictures of my new M4P. I have read the entire thread and will attempt to make some changes on my end to see if I can produce better quality pictures.

However, the default auto mode so far has produced some really bad picture quality even in RAW. Something is not "Autoing" correctly by the sounds of it. Last night I took a picture and it came out really bad. I also did a few hyperlapse's and those pictures are in jpg format....they looked MUCH better. They were not of the same subject or distance. So that tells me the M4P can take good pictures I just need to learn how to do it. I wish Auto would just work like on my phones but if I need to learn manual methods so be it.

If I find myself unable to improve my image quality I will come back for some education and help. :)

EDIT: my pictures look like the OP's currently.
 
However, the default auto mode so far has produced some really bad picture quality even in RAW. Something is not "Autoing" correctly by the sounds of it.
Whether you use auto or manual exposure, the camera is capable of producing good images.
It's likely that the camera isn't at fault, but the way it's being used might be.
So that tells me the M4P can take good pictures I just need to learn how to do it. I wish Auto would just work like on my phones but if I need to learn manual methods so be it.
Probably the best way to move forward would be to post a few of your problem images and members would be able to identify the issues and give you some pointers to avoid them and improve your shooting.

Upload a few full-sized, original jpg files to Google Drive or similar, start a new thread here and post a link to the image files.
 
  • Like
Reactions: bbertram
Whether you use auto or manual exposure, the camera is capable of producing good images.
It's likely that the camera isn't at fault, but the way it's being used might be.

Probably the best way to move forward would be to post a few of your problem images and members would be able to identify the issues and give you some pointers to avoid them and improve your shooting.

Upload a few full-sized, original jpg files to Google Drive or similar, start a new thread here and post a link to the image files.
i would be happy to look at some of your pics. i previously uploaded raw and jpegs and raw is already not great. also remember the jpegs and raw i did send to dji were evaluated as NORMAL. it is a pity. i think that still photographers are more bothered by this poor performance. i would be so happy with the m4p if it just had the same image quality as the mini2. sadly to say, i just bought a used mini2 again. would be great if DJI can take a serious look at that issue and come up with a SW upgrade
 

DJI Drone Deals

New Threads

Forum statistics

Threads
135,088
Messages
1,602,301
Members
163,575
Latest member
foooz
Want to Remove this Ad? Simply login or create a free account