That's not an enough report, show examples, process the same picture in both MLVApp and CR2HDR, adjust the exposure so the two images match, then post the results.
Also, share the original not processed files.
MLVApp Dual ISO works fine here.
We will never know without examples and direct comparison which will show why MLVApp is *faulty* and why CR2HDR is *correct*.
Dual ISO algorithm isn't designed for video processing in first place, I can have flickering issues also with CR2HDR.
During my experiments, there are many types of flickering issues, one of them is about white level output, I think a1ex did add a function in CR2HDR (you need to add a command for it to make it work) for making white level constant, and this fixes one of flickering issue. I think this one MLVApp doesn't have it.
I will make real tests and double check in future, and start a thread for Dual ISO flickering problems in general.
If someone can show (real examples) how cr2hdr performs better than MLVApp, go ahead
.
Sources at the bottom are included.
The same photo: DNG - CR2HDR, and dng - MLVApp, in which the photo was transcoded
in mlv, ticked in front of Dual ISO and saved due to the preset "CinemaDNG Uncompressed".
Photoshop only saw the difference that after CR2HDR the result is brighter, but the photos are more or less similar for comparison.
In the case of MLVApp - Camera RAW did not recognize on which camera and lens the photo was taken. The weight of the photos are almost the same.


For speed I better send pictures of my timelapse to CR2HDR.
But I cared about the video. After the inclusion of the Dual ISO video in MLVApp - everything becomes dark and I increase the exposure, after which a lot of colored noise comes out.



(I had a question: is it possible to shoot in Crop_Mode V2 - Preset "Centered x5 zoom" with fps override module? Received videos were not with x5 zoom.)
I decided to shoot more in different variations and getting a new result - made conclusions:
I think it is still right to include Dual ISO in MLVApp, and then export the video to DNG and continue processing in Lightroom or Photoshop (I thought it would be better to set up Clarity and Sharpen in MLVApp before exporting the video to the photos, but somehow the result with these manipulations is not saved and in Camera Raw I re-twist the corresponding settings).
As a result, my Camera Raw does not appear color noise on the same elevated levels of exposure as MLVApp, although the information CR2 photos have more and they weigh 30 with something megabyte, and the video saved in dng - 5 with something megabyte. Therefore, I believe that with CR2 photos more possibilities than with video for post-processing.
By the way for the source video I will attach my
Cat, where there were bright rays of the sun and medium light corridor. I think the detail could have gone better, but I don’t know how to achieve this result through MLVApp, and these shots themselves came out spontaneously

(sharpness is more on noise than on the details of the picture itself).
(password: dualiso)
If you have ideas - you can sit with the sources video:
https://mega.nz/folder/j0QVlTgb#DI90sloYobh1R3oiQSF8ig