Could we imagine a workflow where we open a raw file, start the preprocessing and wait until it is ready to continue the editing, without saving a DNG in the middle ?
I see timing of 7 seconds to export one picture using DeepPrime… for a small batch of picture or some trials it is acceptable.
This could be optional for people who have time and powerful computers.
Could be, but there’s also not so much work to be able to display DeepPRIME output in the main viewer, so for the final solution I don’t think it’s worth going through DNG at all. We’ll see…
Exactly these one (distortion, vigneting, lens sharpness and chromatic aberration) + denoising and demosaicking. No color rendering applied (picture remains in sensor color space).
It’s good to have enhancements but, as mentioned in another thread on DNGs, the format never really took off in the way that it was envisaged and I can’t see any good reason for using it…
This is for most multi app users a major plus.
i think me personal wil only use this when i go for “multiple image files” stacking in an other application. (i tries earlier and DNG did change my colorWB things to much so i stopped and go to Tiff 16b adobeRGB as source file.)
One thing i am interested in is Sharpening vs denoising. Does react DeepPrime in his calculations when i sharpen before export? (when it’s DNG it’s a demosiaced file and any sharpening is direct visible in dxOPL and other apps.)
second thing:
Say i set Auto exposure compensation active and the other “magic Wants” of DxOPL do they all get “baked” in the DNG export? Because some are colorspace(workingspace) srgb/adobergb related due there boundery’s.
like this one:
This could be an influence from my large format film work, where I would never dream of disposing of a negative.
I already have (e.g.) a 79MB Nikon RAW file. Doing this would then give me a 128MB DNG file, either to replace the RAW or to store in addition.
In some ways, I can see the reason for ditching the original RAW, once the de-noised DNG has been created. After all, why would I want to keep a rough looking, noisy, file when I can now see, on screen, exactly how the image would look when I have applied any other corrections.
Now, that’s what I call a clean image. I find myself saying, if I got this quality out of the camera, I would be a very happy bunny. My only misgiving is having to find 50% more storage for every file. But, then again, if I had a 51 Mpx Fuji GFX 50, I’d be looking at about the same file size and, then, I’d still have to clean up any noise.
Just think, stunning quality clean RAW images - Do I keep or do I throw (the originals) ?
(I would say, even if that is not convenient to store, keep originals in any case: we never know, DeepPRIME and/or lens corrections might be improved even further in a future version of PhotoLab, and they would require these original files to work)
i can see this as a big advantage for professional photo editors:
use a dedicated pc for digesting RAWfiles push it throug DXOPLv4
create DNG’s on a shared folder/cloud/ server and use that as startingpoint for editing and developing.
storagespace is then much less an issue then at home.
Archive the Rawfile’s on a different offline storage in case of disaster.
For me as a few image’s a month and one by one developing route is DNG mostly storage filling extra file with no meaning for me unless i use it as transport container to an extra program.
But still i like that “floating WB” it has now. i can export and use the second application which treat it as a “cleaned rawfile”.
2 Likes
m-photo
( Marc (macOS Sonoma on MBP16" Intel))
24
Never !
…if you would like your picture to benefit from DeepDeepPrime one day
Storage costs less and less everyday. Except if you want full SSD…
But I agree with your post. Great news for big projects.
I’m not likely to need this feature except perhaps occasionally but it’s a great option to have.
One question I have. If I export such a DNG file, what attributes does it have? Will Lightroom still see it as coming from my camera? Or is it in some bespoke DNG sub-format? And what about the meta-data?
I’m trying to understand if I push a camera-created DNG through PL4 and then on somewhere else, what is missing/different (apart from the great processing) compared to the original?