Compared to the P4P I'm finding:
1)  Lens flare is a lot better. Mostly gone are the terrible red/green blotches. Excellent.

2) Shadow noise is bad to start with compared to P4P. I've processed thousands of images from the P4P for commercial use, but now I'm having to nurse the M2P images much more.

Here's an example at 100ISO/RAW DNG at 100%, with very little LR post processing. I had assumed the M2P has the same sensor as the P4P. If it does then there seems to be some poor in camera processing?  Really not impressed by the image noise - seems a significant step backwards.  Hopefully a firmware update can improve something?


What are other photogs seeing?

Yes same result for me. Lens flare is very well controlled and now I can shoot in front of  the sun. I find a bit more definition than the P4pro because I need to add less sharpness in post BUT noise in very very important ! In the shadows of course but in blue sky too !  If sensor is the same pleaseDJI  can  you fix the problem ?

Hi Rob8888, thanks for reaching DJI Forum. We do apologize for the issues you had recently. Can we get a copy of the photo that is unprocessed for comparison? We're here to help you. Thank you.

El Diablo Posted at 2018-9-24 23:13
You are bumping the Clarity slider to +76, that will also introduce artifacting... is that your normal workflow? Can you share the RAW files with the rest of us, i would like to edit them myself if you don't mind.

How do I share these? Whats the best way please?

stefgo Posted at 2018-9-24 22:24
Yes same result for me. Lens flare is very well controlled and now I can shoot in front of  the sun. I find a bit more definition than the P4pro because I need to add less sharpness in post BUT noise in very very important ! In the shadows of course but in blue sky too !  If sensor is the same pleaseDJI  can  you fix the problem ?

Yes blue skies are a real mess of noise and artefacts at 100ISO.  What the heck is going on with DJI's processing of this sensor.  I'm tempted to go back to the P4P for most shoots because of these problems, which I'm not impressed about.  Come on DJI, you must be able to improve this?

"I had assumed the M2P has the same sensor as the P4P. If it does then there seems to be some poor in camera processing?

RAWs don't undergo in-camera processing. That's the whole point about RAW.
As El Diablo implicated - you are creating the problem yourself by taking Clarity way up too much.
People love it for dramatic effect but there are limits. Try denoising afterwards.

Eric13 Posted at 2018-9-25 15:08
"I had assumed the M2P has the same sensor as the P4P. If it does then there seems to be some poor in camera processing?

RAWs don't undergo in-camera processing. That's the whole point about RAW.

You're missing my point, like El Diablo.  I'm comparing the sensor to the P4P which I work with daily, regardless of where the clarity sensor is.  
By comparison, the M2P sensor has some weird stuff going on in the shadows (and also clean blue skies). I have to treat the RAW files with kid gloves for them to be usable - often having to put a median filter layer through areas such as skies and calm water areas to clean them up, even from perfectly exposed well lit files at 100ISO.  This shouldn't be so.  I don't know if mine is worse than others as I haven't seen any detailed look at RAW photos (plenty going on about video quality though). Would appreciate your observations of your files

Rob8888 Posted at 2018-9-25 15:27
You're missing my point, like El Diablo.  I'm comparing the sensor to the P4P which I work with daily, regardless of where the clarity sensor is.  
By comparison, the M2P sensor has some weird stuff going on in the shadows (and also clean blue skies). I have to treat the RAW files with kid gloves for them to be usable - often having to put a median filter layer through areas such as skies and calm water areas to clean them up, even from perfectly exposed well lit files at 100ISO.  This shouldn't be so.  I don't know if mine is worse than others as I haven't seen any detailed look at RAW photos (plenty going on about video quality though). Would appreciate your observations of your files

I don't see a comparison. A comparison is when you shoot the same image, at the same time, with the same exposure/settings...and then compare them. It's hard (without that) for us help evaluate your statement (or for you, honestly, to make it). Raw's would help, but you never know how the raw profiles are applied (for instance DJI could have adjusted the metadata so that there is less noise reduction going on on the M2 vs the P4P, which would reduce the appearance of noise, but also smear detail and cause artifacts). Seeing as the assumption is that these are the same sensor, it's hard to imagine that one has more shot noise than the other. The M2 has a unique color palette compared to the other DJI cameras, and that profiling could be causing increased noise, but in theory, if the shot noise is the same, then we should be able to adjust the raw files to look similar in terms of sharpness, and noise, if not color.

Guys, I'm a working pro, I deal with files every day from medium format dslr to DJI drones - I get a strong feeling for resilience and quality of files from all my equipment.  I get all the stuff you're saying, BUT....  
I'm simply asking for others impressions of the M2P RAW image files.  I'm seeing there are some great improvements and some really poor things - like bad noise, even at 100ISO.  
I'm saying that I have to treat M2P files gently, and that the starting point for noise from this sensor is really poor.  What do you see? Tell me. I'm not after primers on sharpness/clarity blah blah blah.

Rob8888 Posted at 2018-9-25 16:07
Guys, I'm a working pro, I deal with files every day from medium format dslr to DJI drones - I get a strong feeling for resilience and quality of files from all my equipment.  I get all the stuff you're saying, BUT....  
I'm simply asking for others impressions of the M2P RAW image files.  I'm seeing there are some great improvements and some really poor things - like bad noise, even at 100ISO.  
I'm saying that I have to treat M2P files gently, and that the starting point for noise from this sensor is really poor.  What do you see? Tell me. I'm not after primers on sharpness/clarity blah blah blah.

Hi Rob,

As one working pro to another (one who began color correction images nearly 35 years ago), I can't tell you how many times my "strong feelings" about something have been bashed when I do a side by side comparison between the things I have strong feelings about. You say you are comparing something, but you show us no comparison. You say you want our opinions of the Raw Files,  in relation to the P4p, but you don't show us any raw files from either craft. There are plenty of opinions about the m2P image quality floating out there, but you've asked something very specific. Don't you think it would be for the good of the forum, and other people who might read this thread, if you demonstrated what you're talking about. I know I'd benefit. Not everybody owns both drones, but you do...so why don't you show us what you mean?

I haven't had any noise issues, unless, its user error of wrong settings.  when its too under exposed and you lift exposure over 1 then jes ofcs noise is coming...  

Heres a test night shots.

JarmoK Posted at 2018-9-26 00:16
I haven't had any noise issues, unless, its user error of wrong settings.  when its too under exposed and you lift exposure over 1 then jes ofcs noise is coming...  

Heres a test night shots. [view_image]

Although you don't show a comparison between the MP2/PP4, I trust your experienced eye.

Perhaps show an exact comparison for exact proof to put to rest your doubters.

It has been extensively documented elsewhere that the MP2 sensor has potential issues.

There seems to be some disagreement but little doubt there could be a problem.

Rob8888 Posted at 2018-9-25 15:27
You're missing my point, like El Diablo.  I'm comparing the sensor to the P4P which I work with daily, regardless of where the clarity sensor is.  
By comparison, the M2P sensor has some weird stuff going on in the shadows (and also clean blue skies). I have to treat the RAW files with kid gloves for them to be usable - often having to put a median filter layer through areas such as skies and calm water areas to clean them up, even from perfectly exposed well lit files at 100ISO.  This shouldn't be so.  I don't know if mine is worse than others as I haven't seen any detailed look at RAW photos (plenty going on about video quality though). Would appreciate your observations of your files

I suppose the biggest point being missed here , is two years ago P4Pro was retailing at €1699, two years on M2P retailed at €1449, so a difference of almost €250, so where does that €250 go, well for those who have flown a P4Pro and M2P it’s pretty clear as an aircraft M2P is showing to be a lot better, more sensors, better stabilization, longer battery as well as more features like quickshots hyperlapse better AT, etc.
The expectations seem to be that we want a better craft better camera but want it at reduced price, people need to get real and see there are economy’s of scale to be applied, Comparing Cameras is all very well and good, but take into consideration what you paid for them, what you need them for, is Camera more important, then pay more, is portability more important, then you have a choice.
What you have with M2P is a great package that will suit those looking for a good Camera craft that works well, great portability, and reasonably priced.
When you compare , take note of why your M2P is cheaper than P4Pro, try make the very best out of what you have.
It is also great to have debate and to put forward ideas in the hope that dji will improve on elements of their camera, which they have with both P4Pro and orignal Mavic , so hopefully from this discussion we will see tweaks that will enhance cameras capabilities.
But we must realize that you can’t make a silk Purse out of a Sows ear.

Hi,
the sensor may, may not, have problem as you describe, i have made just 1 test flight in day conditions (M2P), so i cant say yes or no to that.
The main problem in this thead is:
showing a RAW which is a bit heavilly postprocessed in LR (i.e clarity)
To state the problem, You have to show unprocessed RAW.
And if comparing with P4, when attached one unprocessed RAW from P4 as well.
Then we can compare and judge.
Our eyes are realy good when comparing 2 same pictures beside each other, but not for one by one.

I see noise issues aswell at 100 iso.

Shadows seem to be way to light in unprocessed raw, seem to be pre boosted which would show hidden noise right?..

Skyris Posted at 2018-9-26 02:44
Shadows seem to be way to light in unprocessed raw, seem to be pre boosted which would show hidden noise right?..

Thank you Skyris. This is exactly what I'm seeing.  Something is happening to these RAW files before we get them that doesn't seen normal.  It's as if they've already been pre-pushed to their noise limits and there is little or no resilience left.

I'm not spending hours providing 'side by side' comparisons to convince the naysayers and knockers, I'd appreciate my considerable and long experience in field being considered and ideally added to by others.

hallmark007 Posted at 2018-9-26 02:24
I suppose the biggest point being missed here , is two years ago P4Pro was retailing at €1699, two years on M2P retailed at €1449, so a difference of almost €250, so where does that €250 go, well for those who have flown a P4Pro and M2P it’s pretty clear as an aircraft M2P is showing to be a lot better, more sensors, better stabilization, longer battery as well as more features like quickshots hyperlapse better AT, etc.
The expectations seem to be that we want a better craft better camera but want it at reduced price, people need to get real and see there are economy’s of scale to be applied, Comparing Cameras is all very well and good, but take into consideration what you paid for them, what you need them for, is Camera more important, then pay more, is portability more important, then you have a choice.
What you have with M2P is a great package that will suit those looking for a good Camera craft that works well, great portability, and reasonably priced.

Thanks Hallmark007.  Points taken. However, DJI makes a big play with the Hasselblad name on this camera.  Given this is a couple of years on from the P4P, coupled with the Hasselblad name I really had expected more.  If it is the same sensor as the P4P then something seems to be going wrong with onboard processing, to my eye

I wouldn't call it "very little LR post processing" with the clarity, vibrance and saturation as high as you have them. To me that amount of colour noise looks about right for the sensor and lighting conditions. f/4 would have been better to obtain the correct exposure and lessen the noise.

Run them through RawTheraphy.
It will strip the files from any profiles and camera adjustments.

Im getting a bunch of noise as well on my M2Z, but maybe thats because the aperature and lense are so much smaller on the zoom than they are on the pro. Hopefully its my camera settings, which im still learning and I can get that figured out....

-SkinJob2501

Rob8888 Posted at 2018-9-26 03:05
Thank you Skyris. This is exactly what I'm seeing.  Something is happening to these RAW files before we get them that doesn't seen normal.  It's as if they've already been pre-pushed to their noise limits and there is little or no resilience left.

I'm not spending hours providing 'side by side' comparisons to convince the naysayers and knockers, I'd appreciate my considerable and long experience in field being considered and ideally added to by others.

God, man. Then perhaps you shouldn't spend time asking questions on forums. Several people have tried to help answer your question, only to have you tell us you're a pro, and that therefore you already know the answer...and you then just start calling us names.

Exactly how many hours does it take to take both craft out to the back yard, shoot two raw files, and post them on dropbox. I estimate .25 hours, but maybe more. It really sounds to me like you just want to have your opinion confirmed. So Here:

To me it looks like you have lots of noise there. Probably at least 3x more than the P4P. Maybe even more. Or maybe less than 3x. Definitely, probably, maybe more than the P4P. (I wouldn't really know because I've never seen a still photo taken of that lighthouse, at that time, with the P4P, with those lightroom settings. If the shadows have been lifted (or any other metadata adjustment made by the raw profile), something this naysayer --or am I a knocker -- already suggested as a possibility, I wouldn't know....  but I definitely agree with you. I'm sure of it. Or, as sure of it as I can be.

And what's wrong with knockers. I've always liked knockers. (maybe it's a language thing.)

Compare the RAW files in RawTherapy is a good suggestion.
There is 99% chance that the sensor is the same as the P4P.
Few chance as well to have some additional noise due to electronics, the CAN conversion is made as close as possible to the pixel and the digital CDS of the Exmor's is very efficient.
Maybe the difference could be the temperature (the dark noise is x2 every 6°C).
The blue channel is the less sensitive. Be sure to compare the noise in the same conditions.
On the other hand, the noise visible in the image looks 'sympathetic', should be easy to remove.
How was adjusted the sharpen in the image ?
Maybe DXO will test the MP2 in the future
https://www.dxomark.com/dxomark- ... one-camera-sensors/

If you’re going to push those levels that high, then try running it thru Dfine (Nik Software). It’ll clear that up. Can I get the link to download the RAW file?

Here's a zero'd out version - no shaperning, no noise reduction, no clarity, no saturation etc.

Any small positive raw adjustment emphasises the noise badly.  Much worse than the P4P, which I'm picking is the same sensor. Love to know why these files are so different and have such poor resilience. It's as if they're being pre-stressed somehow?  



This is a bit random, but what if the lens wasn't a f2.8 as claimed and was being 'pushed' in camera say a stop? That would totally fit with the feel I'm getting from working intimately with these files after a few days, for instance.

Rob8888 Posted at 2018-9-26 14:49
Here's a zero'd out version - no shaperning, no noise reduction, no clarity, no saturation etc.

Any small positive raw adjustment emphasises the noise badly.  Much worse than the P4P, which I'm picking is the same sensor. Love to know why these files are so different and have such poor resilience. It's as if they're being pre-stressed somehow?  

Could be a possibility, the Dji x7 lenses are around half stop darker/slower than the x5s (panasonic / olympus) lenses at f2.8

I took a look at some publicly available Raw stills from both the P4P and the M2 Pro. As the images were not side by side comparisons, it's always difficult to judge things definitively, but these are the things that jumped out to me.

Both cameras show considerable noise in "flat, cool, neutral areas". When normalized to each other in ACR, (all noise reduction turned off, sharpening set to 0 and white balance set the same, all other controls nulled), generally the M2Pro shows lower noise throughout the image, although the deepest shadows appear noisier, but more detailed. The P4P seems noisier overall, but has less noise and less detail in the shadow areas. This would imply that the two cameras are receiving different approaches to noise reduction in the Raw encoding.  Overall, I see what looks like considerable "smearing" in the P4P image...it looks less detailed and sharp. This could be a product of higher noise reduction, or it could be the lens. All of these differences are relatively subtle.

In particular white balance applied in Raw seems to affect noise levels in the flat, neutral, cooler areas of both cameras. This is one of the reasons why I've said from the beginning that "working pro feelings" are pretty suspect (including my own) without a proper side by side comparison. In initial view of the two cameras with settings on both synced to nulled settings, ACR doesn't adjust white balance. I noticed in that situation that variations in the tint slider in particular, caused significant changes in noise(the M2Pro looked slightly noisier in flat neutral areas). When I set both cameras to the same WB and null the tint slider, there was very little difference between the two than what is stated above.

When I look at the images from both camera as a whole, the M2Pro appears slightly contrastier, and has more saturation. The P4p images look flat, and in need of color correction. This implies the M2Pro has a profile (Hasselblad Color?) that is different from the P4P Profile (DJI Flat and Dull?).  Assuming this is true, increased saturation and contrast in a profile would contribute to the appearance of more noise.

Finally, if shown two images from the two cameras, unlabeled, I would undoubtedly choose the M2Pro image. It appears to have more color information and detail, lower noise overall, and seems to need less post processing. Given the tools available in ACR and lightroom, if you'd appreciate a flatter, less saturated "starting point". I'm sure that can be achieved quite easily. As we know from various sources, the M2 Pro and P4P have identical sensors, but different processors. These results are not surprising.

I see nothing to back up the claim that the M2pro has higher noise generally than the P4P.

Edit: I performed an experiment on two similar images, trying to bring the two together more from a color profile perspective. After pushing and pulling contrast and saturation, (again leaving WB, NR, and sharpening alone), what's apparent is that the P4P image has much higher color noise in darker neutrals, and the M2Pro shows a little more luma noise and detail, and almost no color noise in those same areas. Again, I'd pick the M2Pro image easily over what I see from the P4P.

Barry Goyette Posted at 2018-9-27 07:24
I took a look at some publicly available Raw stills from both the P4P and the M2 Pro. As the images were not side by side comparisons, it's always difficult to judge things definitively, but these are the things that jumped out to me.

Both cameras show considerable noise in "flat, cool, neutral areas". When normalized to each other in ACR, (all noise reduction turned off, sharpening set to 0 and white balance set the same, all other controls nulled), generally the M2Pro shows lower noise throughout the image, although the deepest shadows appear noisier, but more detailed. The P4P seems noisier overall, but has less noise and less detail in the shadow areas. This would imply that the two cameras are receiving different approaches to noise reduction in the Raw encoding.  Overall, I see what looks like considerable "smearing" in the P4P image...it looks less detailed and sharp. This could be a product of higher noise reduction, or it could be the lens. All of these differences are relatively subtle.

Interesting comment thanks.  Did you get a sense of resilience/elasticity of the files when working with them? I'm experiencing a lot less from the M2P.  Perhaps the 'hasselblad' profile is already working them hard?

not really. The m2 pro has a little more of a look baked in, but I'm pretty sure thats really all profile, meaning it's metadata. If that's the case, then its just about backing out those corrections. I've pushed these images pretty hard, in every direction and I don't really see anything that doesn't look like every other DJI DNG I've ever processed. In general the M2P image is cleaner out the gate, and the lens on the M2P appears to be substantially better than the P4P (less smearing and CA), so to me the images seem "more" malleable. But I'm not working with either craft every day.

I have a simple conclusion regarding this camera and smartphone cameras, they are great, as long as the light is sufficient, in the dark no one is even close to yield any quality, noise level: extreme.

From shooting both P4 Pro and MP2, IMO both cameras even at ISO100 have the possibility of producing noise in the shadows.  I quickly started using 5 shot AEB on the Phantom and use it on the MP2 also.  It's the same sensor, I can't believe that DJI had Sony make a new 1" sensor, but if they did, it's not much better.   I always expose to the right with both drones, as I have found both will allow highlight recovery better than shadow push.  Shadows just have to be dead on.  

Just assumed it's an issue with the older Sony 1" in both products.   

All raw capture, LR or C1 used.  BTW C1 will give a slightly cleaner image, with less detail.  To use C1, you have to change the name of the camera from Hasselblad to DJI, and camera type to the type of the P4 PRo.  I realize they aren't the same lenses, but they are close enough that you can get a lot done.  Color IMO is also a bit better in C1 out of the gate.

Paul C

Did DJI respod to this? I have simmilar issue.