search

darktable-org / darktable

darktable is an open source photography workflow application and raw developer
https://www.darktable.org
GNU General Public License v3.0
9.45k stars 1.12k forks source link

Blackmagic Design Pocket Cinema Camera 6k Pro #11495

Open lexxy23 opened 2 years ago

lexxy23 commented 2 years ago

Hi, all,

please provide support for the DNG files of the Blackmagic Pocket Cinema Camera 6k Pro.

I posted already a sample raw picture on the raw.pixls site (look for 3 spheres)

Quadisti commented 1 year ago

I also would like to edit bmpcc 6k pro images on darktable.

kmilos commented 1 year ago

Does it only output DNGs in uncompressed mode?

A bit weird this one indeed - the RPU sample DNG is uncompressed and should in theory just work (unlike known issues with some other Blackmagic models that output some special and not yet supported compression scheme), but rawspeed complains with:

[rawspeed] (Blackmagic Design - Pocket Cinema Camera 6k Pro - 12bit (16_9).dng) virtual std::unique_ptr<rawspeed::RawDecoder> rawspeed::RawParser::getDecoder(const rawspeed::CameraMetaData*), line 91: No decoder found. Sorry.

It could be (TBC) because there are unusually no mandatory Make nor Model tags in this file (just the optional UniqueCameraModel) - if not fixed by some later firmware, this should be probably reported to Blackmagic...

lexxy23 commented 1 year ago

@kmilos Sure. But it makes me wonder why tools like Affinity Photo and Luminar Neo are able to decode them ootb.

Or is Darktable doing something weird here?

If you need further pictures, let me know.

kmilos commented 1 year ago

They don't do any strict checking? 😉

Checking for compliance is not "weird" IMHO, more of the opposite.

In any case, the RPU sample w/ software 7.5.1 is breaking the baseline TIFF/EP spec - there are at least those two mandatory tags missing. I haven't checked if there is anything amiss re the DNG spec.

kmilos commented 1 year ago

Actually it fails even earlier than that! The TIFF IFD structure itself is borked ("Two IFD's overlap. Raw corrupt!" because the IFD chain is not terminated properly) and cannot be cleanly parsed.

Even with that fixed in the hex editor, then we hit the missing make/model problem...

So these are definitely all issues to be raised with Blackmagic to fix in their camera firmware (if not already).

Edit: The TIFFEPStandardID tag also seems to be stored in the wrong byte order.