Am 23.06.2016 um 16:11 schrieb Tobias Ellinghaus:
On Thursday 23 June 2016 14:38:37 Christian Mandel wrote:
Am 23.06.2016 um 13:54 schrieb Tobias Ellinghaus:
On Thursday 23 June 2016 13:07:11 Christian Mandel wrote:
Hi!
I wonder if the best/only/recommended method to correct this is still
using the geotagging module?
Yes, as it's the only way.
Does the time shift propagate to the xmp
No, and it probably never will.
That's a pity. What's the reason? Is there another option on the horizon
or at least thinkable?
Unfortunately your answer makes me more and more willing to alter the
date in the raw files. I know about the risks, but a big advantage of
the sidecar files is that I do not have to care about the database, I
could always reconstruct it from the xmp files*. But this answer breaks
this behaviour, doesn't it?
The request is quite old and I guess everyone agrees that having something
only in database and not the XMP is bad. However, we haven't found a way to fix
the issue. Details are here:
>
> https://redmine.darktable.org/issues/10569
K, I got the problem. Unfortunately the proposed patch is gone, I have
to dig deeper in the git to find it.
Anyway, does something argue against the following procedure while
loading an xmp file:
- Check if the xmp originates from the actual file (e.g. metadata or
filename comparison if file without metadata, I would even do a pixel
data checksum¹)
- If file belongs to xmp, take the correction value in xmp
- If not, silently drop it (it is unlikely that the other file's date
is wrong) or ask the user how to proceed.
This way, processing another file would not break this file's timestamp,
and a user that knows how to load an xmp file with another image is
likely able to respond the proposed user interaction (keep the time
offset?). And IMHO loading a foreign xmp is a rare event compared to
wrong camera time stamps anyway.
¹This could anyway a reasonable feature request :-)
____________________________________________________________________________
darktable user mailing list
to unsubscribe send a mail to [email protected]
