Welcome, Autel Pilots!
Join our free Autel drone community today!
Join Us

Color science on Evo 2 Pro

GeoDrone SRL

Well-Known Member
Joined
Aug 18, 2021
Messages
475
Reaction score
273
Age
46
Hello to all and especially to BLADE STRIKE.
Is there any chance to review the color science from EVO2Pro and fix that damn magenta ??????
I mean, a color profile is the easiest thing to fix in this drone....all DJI forums are full with this BS color science from Autel Evo 2Pro....and as an owner of both DJI and Autel, I can tell you this: Autel is kicking Mavic 3 *** on details and quality, but is LAME on color science...
Come on guys...you had 3 years for this fix...
USE YOUR PILOTS, YOU HAVE PLENTY TO GET YOUR INFORMATION TO MAKE IT BETTER !!!!
 
Color science is a thing? lol.... Reminds me of what retailers at TV areas do. Take the on sale items, mess with the controls to make it look worse than the higher end models.
You can shoot SOOC, or you can us your own profiles. Also if it is darker (for any camera use) do not use AWB as the sun sets or lighting changes from afternoon to early evening sun.
The color science for us here with a fleet of E2 drones in our company matches the Sony color better than anything else. The newer A7RV cameras we bought are bit more leaning warm, but again you can adjust all that stuff yourself.
Too much weight is used on the color science marketing terms... It really for most of them is what they find in DJI vs Autel or Canon vs Sony claims. Many using Canon use DJI as a preference especially those that self edit and developed their own workflows and custom presets in LR, PS, PP, etc etc. Some hate the "color science" of Adobe, and prefer what they see in Davinci... again a silly claim based on what ???
It's as bad as the BS claims about Base ISO, and so many other details people latch onto instead of just working the controls... Everyone thinks they have to shoot at XYZ settings... you don't !! If that is a workflow thing for these content creators especially on YT.. well then how is that being fair to the company you dislike because something doesn't fit your editing (yes all those content creators are total pro editors too) needs.
The internet is full of terms, made up claims, and the worse yet VS videos... I can't stand most of them. Many YT CC's are full of their own bias when it comes to products they use. From cameras, drones, gimbals, mics, software, and their own computers.... BIAS is everywhere.. not to mention affiliate links, free gear they often get.

Sony followed some of the CS change recommendations trying to win over Canon users, and in the end more complained about the CS being too warm now in the A7IV, A7RV and saying and mostly repeating garbage they heard from YT videos suggesting "If I wanted Canon colors, I'd buy a Canon" ..... lmao at this crap
For us here at our company, most of what we find is the Autel products in the E2P and even the Lite and Nano+drones we use routinely simply outclass the camera performance of the DJI drones we had purchased just to see what the HYPE was about. So far, Autel still wins over our media shooters just like many of our Canon shooters had to relearn their methods when they shot with our Sony Alphas. They too learned there are dozens of ways to work CS in any camera... We also do not edit in house. We use editing companies that work our media just perfectly no matter what we use for capturing. What they deliver in color, contrast is very similar no matter the device. They have told us though that Autel and Sony are a great match and the costs we have had in editing is far less given those similarities of the CS...
 
Last edited:
Actually color science is a thing… and if it’s set up correctly in DaVinci Resolve it will give you a pretty accurate base grade. It’s really surprising that every pro “cinema” camera has a native color science for their log footage / sensor but Autel doesn’t seem to have (or know) about it. All the footage I grade from my other cameras has a known color science (for color grading). Seems like no one really knows what the hell the color science is of the sensor in the Autel drones, at least in a year and a half of looking I’ve never found anything that definitively talks about the color science be it Sony S log, Canon C log, Braw or Arri log. Just tweaking knobs and dials is great to get a finished product “look” but it really doesn’t let you begin from a consistent color profile like every other pro camera that shoots in log. Basically when you set your color science, and log profiles correctly in Resolve it does the base grade for you and then all you have to do is get your look. Maybe color grading science is irrelevant for drone users but to a cinematographer / colorist it is absolutely critical.
 
  • Like
Reactions: gschulzuio
Actually color science is a thing… and if it’s set up correctly in DaVinci Resolve it will give you a pretty accurate base grade. It’s really surprising that every pro “cinema” camera has a native color science for their log footage / sensor but Autel doesn’t seem to have (or know) about it. All the footage I grade from my other cameras has a known color science (for color grading). Seems like no one really knows what the hell the color science is of the sensor in the Autel drones, at least in a year and a half of looking I’ve never found anything that definitively talks about the color science be it Sony S log, Canon C log, Braw or Arri log. Just tweaking knobs and dials is great to get a finished product “look” but it really doesn’t let you begin from a consistent color profile like every other pro camera that shoots in log. Basically when you set your color science, and log profiles correctly in Resolve it does the base grade for you and then all you have to do is get your look. Maybe color grading science is irrelevant for drone users but to a cinematographer / colorist it is absolutely critical.
As a Resolve Studio user along with E2P operator, since you mention both and color science, what are you using for settings, other experience or tips would be interesting to hear, thks in advance.
 
As a Resolve Studio user along with E2P operator, since you mention both and color science, what are you using for settings, other experience or tips would be interesting to hear, thks in advance.
Unfortunately I haven't found anything that really works yet. I have tried setting up Davinci Color Managed and ACES cct (which seems to be marginally better that DCM) and almost every log profile there is in Resolve but I have yet to discover the actual log profile. DJI provides them (natively in Resolve) so I'm perplexed as to why Autel doesn't..? I'm sure it's one of the major manufacturers that builds the sensor (Sony?) so why not just tell us the log profile? After searching far and wide I have yet to find a source for the EP2 sensor. I'm in a BMD Resolve color grading class next week and will ask around there, it's usually attended from users world wide, maybe someone has finally figured it out.

BTW for those of you who are not colorists a LUT is a mathematical formula to translate a sensor's log profile to something that doesn't look washed out. Color management lets the colorist work in the full native color space of the sensor as it saw it on set (or in the air) so it doesn't look like anything like log, it looks as it did when the sensor captured it. Colorists will set up color management first, then grade (as it looked in reality) then, if needed apply a LUT as the very last step (to get a particular look) before going out to rec709. I understand most drone pilots just grade the footage and there is nothing wrong with that but using color management in your software lets you get a more consistent look across projects especially if you are also using terrestrial cameras and /or cameras from different manufacturers. Just my 2 cents.
 
  • Like
Reactions: gschulzuio
Unfortunately I haven't found anything that really works yet. I have tried setting up Davinci Color Managed and ACES cct (which seems to be marginally better that DCM) and almost every log profile there is in Resolve but I have yet to discover the actual log profile. DJI provides them (natively in Resolve) so I'm perplexed as to why Autel doesn't..? I'm sure it's one of the major manufacturers that builds the sensor (Sony?) so why not just tell us the log profile? After searching far and wide I have yet to find a source for the EP2 sensor. I'm in a BMD Resolve color grading class next week and will ask around there, it's usually attended from users world wide, maybe someone has finally figured it out.

BTW for those of you who are not colorists a LUT is a mathematical formula to translate a sensor's log profile to something that doesn't look washed out. Color management lets the colorist work in the full native color space of the sensor as it saw it on set (or in the air) so it doesn't look like anything like log, it looks as it did when the sensor captured it. Colorists will set up color management first, then grade (as it looked in reality) then, if needed apply a LUT as the very last step (to get a particular look) before going out to rec709. I understand most drone pilots just grade the footage and there is nothing wrong with that but using color management in your software lets you get a more consistent look across projects especially if you are also using terrestrial cameras and /or cameras from different manufacturers. Just my 2 cents.
No worries thks, fwiw the camera sensor on E2P is Sony. Concur would be nice to have a profile for Autel to use with DCM or ACEs, however can work around it.
 
I have tried the Sony Log(s) and none give me the results I would expect.. With ARRI Raw & Resolve DCM the raw images look awesome. Maybe it's something in the way the Autel encodes the log images E2P
 
  • Like
Reactions: gschulzuio
I have tried the Sony Log(s) and none give me the results I would expect.. With ARRI Raw & Resolve DCM the raw images look awesome. Maybe it's something in the way the Autel encodes the log images E2P
Concur.
 
Here are a few examples of some (downsampled) 12k BRAW images, one with Davinci YRGB, one with DCM and one with ACEScct color management.

DYRGB is the RAW FIle, the DCM looks the most natural the ACEScct looks the most vibrant. The only difference was setting up the color management in Resolve, no LUT or grade. This is all done prior to touching any of the footage... Now if color grading the E2P was this easy it would be an absolute joy in post and save so much time.

Thanks to Blackmagic for the images, these are particularly difficult skin tones which is why I picked them.
 

Attachments

  • Screenshot 2023-05-18 at 4.17.21 PM.png
    Screenshot 2023-05-18 at 4.17.21 PM.png
    3.6 MB · Views: 4
  • Screenshot 2023-05-18 at 4.16.25 PM.png
    Screenshot 2023-05-18 at 4.16.25 PM.png
    5 MB · Views: 6
  • Screenshot 2023-05-18 at 4.16.52 PM.png
    Screenshot 2023-05-18 at 4.16.52 PM.png
    4.8 MB · Views: 4
  • Like
Reactions: gschulzuio
Here are a few examples of some (downsampled) 12k BRAW images, one with Davinci YRGB, one with DCM and one with ACEScct color management.

DYRGB is the RAW FIle, the DCM looks the most natural the ACEScct looks the most vibrant. The only difference was setting up the color management in Resolve, no LUT or grade. This is all done prior to touching any of the footage... Now if color grading the E2P was this easy it would be an absolute joy in post and save so much time.

Thanks to Blackmagic for the images, these are particularly difficult skin tones which is why I picked them.
Yup
 
I have tried the Sony Log(s) and none give me the results I would expect.. With ARRI Raw & Resolve DCM the raw images look awesome. Maybe it's something in the way the Autel encodes the log images E2P
Have you tried using Autel's LUT? It's not in the Resolve package, but you can import it.

There is no Autel CST in Resolve, so Davinci Managed Color isn't really an option , but you could shoot an X-rite color checker and color match in Resolve. From there, you could easily make your own LUT for EVO footage.
 
Have you tried using Autel's LUT? It's not in the Resolve package, but you can import it.

There is no Autel CST in Resolve, so Davinci Managed Color isn't really an option , but you could shoot an X-rite color checker and color match in Resolve. From there, you could easily make your own LUT for EVO footage.
Thanks for the reply, I have not tried the Autel LUT, although I usually don't like to apply a LUT until the last stages of CC, and if I'm using different cameras (which I do a lot of) it creates workflow issues when I can't use DCM. DCM is my preferred workflow since it makes working with other raw/log footage more predictable. It's the one truly disappointing thing about Autel as most other cameras I use play well in a color managed environment. Do you know where I can get the Autel EVO 2 LUT?
Thanks again,
Mark
 
  • Like
Reactions: gschulzuio

Is this what you are looking for?
 
  • Like
Reactions: gschulzuio

Is this what you are looking for?

Yes, I downloaded the E2P LUT and I agree it's a fair starting place for clip based CC. Working in DCM /ACES gives you a much larger color space than Rec709 so applying this LUT (which essentially converts to Rec709) sort of bypasses all the advantages of using a large color space like ACES or DCM - which BTW shows other Log footage as the corrected images just as beautifully as the sensor saw it. This LUT only gives you a base grade in Rec709 and Rec709 is a compressed / limited color space, which seems to take away the main advantage of shooting in Log. I prefer to grade in Log/Raw throughout the entire pipeline, and as the very last step convert to Rec709 for final output. I suppose this LUT is great if you only need to grade footage from the E2P but if you shoot BRAW/DNG/ARRILog etc it's very disappointing to be limited to a compressed color space. This is why I hope some day Autel will release the actual Log profile. Just my 2c... M.
 
  • Like
Reactions: DadCorso80
I found this thread looking for the minimal workflow for HDR from the E2P 10-bit log video. Here is what I found that gives me a natural (pre-look) image in Resolve (and this also gives me a good HDR preview on my M1 MacBook Pro XDR display).

Project settings:

1687258867279.png

Clip settings on color panel:

1687256820843.png

To be honest, I'm not 100% sure why this works. My best guess is that the P3-D65 ST2084 (4000 nits) Inverse Display transform is close to the E2P LOG encoding's gamut, gamma and dynamic range. For example, selecting a lower nit variation results in clipping, while selecting Rec.2020 instead of P3-D65 results in an over-saturated image.

While this may not be exact, it seems reasonable given the complete lack of an ACES IDT for the E2P or any other official info.

Curious for others to try this on their own footage!
 

Latest threads

Members online

Forum statistics

Threads
11,210
Messages
102,481
Members
9,784
Latest member
0xD34D