RAW is processed, JPG is not

What you're complaining about isn't caused by DJI, it's caused by Adobe Camera Raw's lack of specific profile for the Phantom 4 Pro. DJI isn't applying tonality to the DNG. Adobe Camera Raw is, when it debayers it. Adobe Camera Raw is the core RAW/DNG processing engine that is implemented and used by all Adobe products, including Lightroom and Photoshop.

Perhaps DJI and Adobe will work together to implement a profile, just like Adobe already has profiles for the Phantom 3.

Or you could try using a different Raw Converter, like DXO, Capture 1, Rawtherapee, etc... Perhaps they treat Phantom 4 Pro DNGs in a way that you prefer.


I will tryout those software to see how it goes.
Thanks
 
  • Like
Reactions: Tomas Wangen
What you're complaining about isn't caused by DJI, it's caused by Adobe Camera Raw's lack of specific profile for the Phantom 4 Pro. DJI isn't applying tonality to the DNG. Adobe Camera Raw is, when it debayers it. Adobe Camera Raw is the core RAW/DNG processing engine that is implemented and used by all Adobe products, including Lightroom and Photoshop.

Perhaps DJI and Adobe will work together to implement a profile, just like Adobe already has profiles for the Phantom 3.

Or you could try using a different Raw Converter, like DXO, Capture 1, Rawtherapee, etc... Perhaps they treat Phantom 4 Pro DNGs in a way that you prefer.

Don't waste your time.
It's just the most banal discussion.

The basic principal of raw again JPEG is clearly misunderstood here.
 
  • Like
Reactions: SmurfN and Andrei V
Andre and Daniel

Again. The question is not if the RAW file can be modify in post, or not. In any camera, at least on my experience, It doesnt matter if it´s a medium format file, a full size sensor or it is Phantom 3 Pro or Phantom 4, never the DNG has come out contrasted, saturated and underexposed(regardless is it´s possible to tweak it in post) never ever before I have seen ACR or any other post processing image software apply a "adjustment" to the Raw Images. Never!

We are not misleading anyone here. Maybe some of you are super happy with the RAW on your P4P´s we are not. Regardless of the megapixels or dynamic range, Phantom 4 RAW images came out flat and without any adjustment applied to it.

I know this is wrong, at least to me. Maybe not to you and I respect that, but to me it is wrong.

Here´s the link to the straight from the camera a DNG and a JPG if you guys want to see it. If after seeing them you still tell me that ACR applies and adjustment to the DNG and that´s why it doesn´t come out flat etc... then why does the JPG comes out flat and without saturation etc.... it has NOTHING to do with ACR I am 100% sure, otherwise that would be applied to any other RAW files from any other camera, and that never happened before.

Again, we are not misleading anyone. This are facts. If you are content having your DNG being applied some process, great! I am not!

The firs screen shot is the JPG and the second one is the DNG

And here is a link for the files themselves: Dropbox - RAW - JPG
View attachment 73880

Dropbox - RAW - JPGView attachment 73879

Here are screenshots taken with Rawtherapee(linux) from the two.

JPG:

JPG001.png


RAW:

RAW001.png

Both are imported with the neutral profile so IMHO we can compare this.
I observe that JPG is a little bit more stretched is to the bright side.

Not a photographer but a linux freek.
 
Here are screenshots taken with Rawtherapee(linux) from the two.

JPG:

View attachment 73883

RAW:

View attachment 73884
Both are imported with the neutral profile so IMHO we can compare this.
I observe that JPG is a little bit more stretched is to the bright side.

Not a photographer but a linux freek.

Compare what?
What are you all comparing? How a raw processor shows the files?
 
Here are screenshots taken with Rawtherapee(linux) from the two.

JPG:

View attachment 73883

RAW:

View attachment 73884
Both are imported with the neutral profile so IMHO we can compare this.
I observe that JPG is a little bit more stretched is to the bright side.

Not a photographer but a linux freek.

It's possible the DJI in-came jpeg processing is automatically applying a shadow boost -- like Nikon/Sony's dynamic range optimizer setting. That could account for why the JPEG's histogram appears to be more stretched towards the right than the DNG.

An experiment you might want to carry out -- lift the shadows or boost the exposure of the DNG to match the histogram of the JPEG. Add some noise reduction. Is the result better, or worse then the JPEG?
 
  • Like
Reactions: Tomas Wangen
Lets approach this differently then, since some of you dont understand what some of us are trying to say.

Its all digital, all the light coming trough the lens, hits the sensor, and from there its all a bunch of 0101010101's put into a system controlled by software. That software can do ANYTHING you program it to do, no matter if its DNG, mov, mp4 or jpeg, if you want to add a sidecar file with a preset, you can. It ALL happens in the camera, before the data is saved to the memory card, both presets and processing.

At this point, either of us, have any information about the facts, that if both DNG and/or jpeg files have been embedded with some presets, other than the lens profile. If you know, how do you know, and what is your source? and where is your documentation? and we are talking about files from THIS camera, and not DNG in general.
What we DO know by now, is that there is more than just one experienced photographer being used to work with these kind of files day in and day out, and not just theoretical knowledge, that have drawn attention to this.
Or are you implying that all of a sudden a bunch of experienced people suffer from problems with their eyes?

So, back to the starting point in ACR when you open the DNG.
If the "neutral" point where the CENTER is, STARTS with a profile that is considerably way off the "neutral" point, then you will not have the same space to work with, cause the distance to the ends is less, cause the sliders have a maximum range, you cannot just pull and pull until the sliders disappears out of the screen, IF there was a preset added in ACR, it should have shown the sliders moved to the the respective points, so we could pull them back to neutral point and start from scratch. We cannot even delete the presets that might be there, so we still end up with the same issue, something "added" that we can remove, in a non destructive way. Do we really need to argue more about this?

If the HISTOGRAM on your camera displays a "perfect" exposure being in centre, what good is that, if the saved file is greased with black? (exaggerate just to illustrate)
Have you ever had a raw file in ACR you tried to lighten up the shadows, reaching the end point, even if you would and could have pulled it further?
What you do think happens if you just have 2 steps on the slider, compared to 5 steps of adjustments? That said, shadow areas in these files should barely be touched at all, but thats another story.

Alternative is to overexpose to compensate for the incorrect histogram, but it all still boils down to the same thing, it has TOO much CONTRAST, which means the DIFFERENCE between shadow and highlights makes it really hard(er) to achieve a balanced picture, where the dynamic between shadow and highlights looks good.

Or if we forget everything weve discussed so far, for a second, and ask this;
WHY does the DNG retouched to your taste, have LOWER quality than the processed jpg that have been saved twice already?
That alone is also a mystery to me.

I still believe all this is software related, and the data saved FROM the camera is causing this, nothing else makes sense considering that a raw file contains 50% more data, and on top of that are "unchanged" raw data.
Dji have done weird things before in their software that didnt make sense, and its not unlikely they have done it again, for all i know.
Im confident the quality of these files can be improved further, by tweaking the software.

If everyone that participates in this thread, could upload their own set of untouched DNG and jpegs, then we can at least compare it to eachothers files, especially the ones that categorically keeps disagreeing that theres is something odd with these files.
 
It's possible the DJI in-came jpeg processing is automatically applying a shadow boost -- like Nikon/Sony's dynamic range optimizer setting. That could account for why the JPEG's histogram appears to be more stretched towards the right than the DNG.

An experiment you might want to carry out -- lift the shadows or boost the exposure of the DNG to match the histogram of the JPEG. Add some noise reduction. Is the result better, or worse then the JPEG?

That is the point. The Raw can be "fixed" or improved in post. It shouldn´t be the case. That´s the point of the thread. I think should be stretched and have more latitude out of the camera than de jpg
 
That is the point. The Raw can be "fixed" or improved in post. It shouldn´t be the case. That´s the point of the thread. I think should be stretched and have more latitude out of the camera than de jpg
And it has up to two more stops of dynamic range than jpeg.
 
Lets approach this differently then, since some of you dont understand what some of us are trying to say.

Its all digital, all the light coming trough the lens, hits the sensor, and from there its all a bunch of 0101010101's put into a system controlled by software. That software can do ANYTHING you program it to do, no matter if its DNG, mov, mp4 or jpeg, if you want to add a sidecar file with a preset, you can. It ALL happens in the camera, before the data is saved to the memory card, both presets and processing.

At this point, either of us, have any information about the facts, that if both DNG and/or jpeg files have been embedded with some presets, other than the lens profile. If you know, how do you know, and what is your source? and where is your documentation? and we are talking about files from THIS camera, and not DNG in general.
What we DO know by now, is that there is more than just one experienced photographer being used to work with these kind of files day in and day out, and not just theoretical knowledge, that have drawn attention to this.
Or are you implying that all of a sudden a bunch of experienced people suffer from problems with their eyes?

So, back to the starting point in ACR when you open the DNG.
If the "neutral" point where the CENTER is, STARTS with a profile that is considerably way off the "neutral" point, then you will not have the same space to work with, cause the distance to the ends is less, cause the sliders have a maximum range, you cannot just pull and pull until the sliders disappears out of the screen, IF there was a preset added in ACR, it should have shown the sliders moved to the the respective points, so we could pull them back to neutral point and start from scratch. We cannot even delete the presets that might be there, so we still end up with the same issue, something "added" that we can remove, in a non destructive way. Do we really need to argue more about this?

If the HISTOGRAM on your camera displays a "perfect" exposure being in centre, what good is that, if the saved file is greased with black? (exaggerate just to illustrate)
Have you ever had a raw file in ACR you tried to lighten up the shadows, reaching the end point, even if you would and could have pulled it further?
What you do think happens if you just have 2 steps on the slider, compared to 5 steps of adjustments? That said, shadow areas in these files should barely be touched at all, but thats another story.

Alternative is to overexpose to compensate for the incorrect histogram, but it all still boils down to the same thing, it has TOO much CONTRAST, which means the DIFFERENCE between shadow and highlights makes it really hard(er) to achieve a balanced picture, where the dynamic between shadow and highlights looks good.

Or if we forget everything weve discussed so far, for a second, and ask this;
WHY does the DNG retouched to your taste, have LOWER quality than the processed jpg that have been saved twice already?
That alone is also a mystery to me.

I still believe all this is software related, and the data saved FROM the camera is causing this, nothing else makes sense considering that a raw file contains 50% more data, and on top of that are "unchanged" raw data.
Dji have done weird things before in their software that didnt make sense, and its not unlikely they have done it again, for all i know.
Im confident the quality of these files can be improved further, by tweaking the software.

If everyone that participates in this thread, could upload their own set of untouched DNG and jpegs, then we can at least compare it to eachothers files, especially the ones that categorically keeps disagreeing that theres is something odd with these files.



English is not my first language so I cannot express my self as I would like to. But you did it perfectly. I agree 100% with you, and as a photographer who works with photos (RAW files) every singles day, well, there´s something off about this RAW files. Someone (forgive if I forget his name) suggested was something on the ACR that wasn´t updated. It made sense. But I opened the file in other programs he suggested and the same thing happens. That means, of course, that the proble (yes! there is a problem) lies from de camera. Even though there´s a kind of rude guys insisting it isn´t that we are dumb and don´t know what we are talking about.

I am sure the only fix to this will be an firmware update once they realize the parameter are mixed up or wrong.
 
Also, respectfully I would advise to everyone who doesn´t know how to fix this or they think we are absolutely wrong and that the DNG coming out from this camera are perfect, it would be best for them to leave the tread, is not my thread but it is only a suggestion, because you wont convince us and we wont convince you.
 
  • Like
Reactions: gr8pics
Also, respectfully I would advise to everyone who doesn´t know how to fix this or they think we are absolutely wrong and that the DNG coming out from this camera are perfect, it would be best for them to leave the tread, is not my thread but it is only a suggestion, because you wont convince us and we wont convince you.

Im happy that you said it, if i did, everyone would gang up on me :tearsofjoy:
 
Im happy that you said it, if i did, everyone would gang up on me :tearsofjoy:
Sure. I know the natural tendency is to defend DJI because if DJI product has a problem then my drone might have one. But it has one. At lest in your drone and my drone. Not saying all are like that though. Also it´s so very comon people thinking they are professional photographers because they take pictures, or aerial photographers because they have a drone with a camera.
Whoever wants to think the DNG is amazing or that the problem are the computer or the retouching software, it´s wrong.
I know I wont gain anything but I will talk to DJI today and see where it get´s me.
I´m still confident, because they are not stupid, that they will correct this in a firmware update. Meanwhile, I will do overexpose my photo by 1 or 2 stops so I don´t clip the highlights but recover more info in the shadows. Specially short duration photos at night. I know I can shoot seconds in it, but I want to try not to do so so much.

Anyways.... let´s hope something is resolve soon.
 
  • Like
Reactions: gr8pics
gr8pics,

If you have the time/means, there is great info on assessing RAW files here: RAW Files | imatest

And you can get a demo version of Imatest to run on a DNG file and see what it produces.

That or one of the other programs mentioned in the link should help you analyze the DNG file and confirm if it is RAW or cooked.

Update: For what it's worth, I bought the P4P to use mostly for photography and, besides a bit more noise than expected, I have been pleasantly surprised by the DNG files including their latitude in post. I would put them ahead of my old Canon 50D files (though they can't match the 5DmkIII).
 
  • Like
Reactions: Tomas Wangen
gr8pics,

If you have the time/means, there is great info on assessing RAW files here: RAW Files | imatest

And you can get a demo version of Imatest to run on a DNG file and see what it produces.

That or one of the other programs mentioned in the link should help you analyze the DNG file and confirm if it is RAW or cooked.

Update: For what it's worth, I bought the P4P to use mostly for photography and, besides a bit more noise than expected, I have been pleasantly surprised by the DNG files including their latitude in post. I would put them ahead of my old Canon 50D files (though they can't match the 5DmkIII).

Im on Mac, maybe someone with Pc can do...
 
Make sens what you are saying, but also when I try opening them in Lighroom is the same thing. I will do further test tomorrow and see how it goes.

What bothers me a lot is, OK, ACR is not updated and therefore doesn´t look the way it should. But then why JPG that should come out with some process apply to it, it comes out flat and without contrast, as the DNG should. Is like if the process or the lack of it would be inverse (bye mistake of DJI of course, which actually could be the case).

Such is the complexity of RAW and it could very well be DJI is applying incorrect metadata resulting in a default underexposed looking shot. Or it could be the lack of the correct profile for the tool you are using as Andrei suggests. But the important part of it is that if the shot was properly exposed in camera, the data is still there. Move the sliders and you'll see that you can achieve the full quality of the shot.

P.S. I have the same challenge with Davinci Resolve showing all my CinemaDNG shots from the X5R as darker than they were shot. It's a bit of a pain but the important point to remember always is the detail is there. It just needs the settings to be adjusted.
 
You wont convince us and we wont convince you.
The trouble is that you have been saying "P4P have problems in RAW" and don't have problems in RAW!! You are looking a problem when it's not there.

I understand all of you point of view. Maybe you can't listen and learn because it's hard to recognize for you.

1) RAW is not processed.
2) RAW is what the sensor sees. RAW it's like a negative in analog photography. If you see different the same DNG it's for baseline profiles that the program loads with the DNG. It's not a problem in P4P, it's a missed interpretation (or not) of the RAW viewer.
3) JPG is processed into P4P based on your profiles (color profile, picture profile, contrast, sharpening, saturation, WB, etc).
4) And most important: Histogram live feed is based on what you see in the screen, WITH the profiles applied live.
5) If you don't like what you see on the DNG straight to P4P, buy a Inspire 2 with X5. Don't waste your time waiting for a firmware that solves a problem, because WE DON'T HAVE A PROBLEM HERE!!!!!!!!!

If you shoot stills, attach to a NONE and 0-0-0 for a better histogram live feed, and when you have the RAW file, reveal as your taste. If you shoot video, it's another story because P4P don't have RAW data in video.

Sorry, I don't know you, but is really hard to think that you live as professional photographer. You can't not know this...
 
Last edited:

Recent Posts

Members online

No members online now.

Forum statistics

Threads
143,086
Messages
1,467,528
Members
104,965
Latest member
Fimaj