I have my main album installed on two systems and, after doing some
changes to metadata on one system, I copied the image files to the other system. Well 'copied' is a misnomer, it's still doing it after several hours. I took a look to see why and discovered that Digikam has written its current version into the metadata of every single image, since I have around 30000 images that makes for a *lot* of file copying! It also means that file modification dates are all changed for no particular reason. Why does Digikam do this? I don't really care what software has been used to do things to my images, at least I don't want it written in to the metadata. Also it writes the Digikam version so even changing Digikam version will cause an update of every file in the album. Is there any way of turning this off? -- Chris Green · |
I don't know about the "signature", other software does the same thing.
However, disabling the option to update the timestamp (Settings, Configure digikam, Metadata, "update filestamp when files are modified) on every picture it's the first thing I do when I install digikam. I lost lots of dates that were not saved to the metadata (the original date when the picture was taken) because of that. I guess the signature is only written to files that have been modified. Which is all of them in your case. I don't know why it's not disabled by default. -- Sent from: http://digikam.1695700.n4.nabble.com/digikam-users-f1735189.html |
woenx <[hidden email]> wrote:
> I don't know about the "signature", other software does the same thing. Other software doing the same thing doesn't necessarily make it a good idea! :-) > However, disabling the option to update the timestamp (Settings, Configure > digikam, Metadata, "update filestamp when files are modified) on every > picture it's the first thing I do when I install digikam. I lost lots of > dates that were not saved to the metadata (the original date when the > picture was taken) because of that. > My pictures are stored in a year/month/day hierarchy so their directory tells me their date. Relying on the file creation or modification date as the 'picture' date isn't a good idea IMHO. I need to rely on that date to ensure the right files get copied to the other system (using rsync). > I guess the signature is only written to files that have been modified. > Which is all of them in your case. > I didn't modify all of them, it's quite difficult to do something that modifies 30 thousand files! I probably made changes that I wanted to do to a few hundred files. I just wanted those files copied to the other system. > I don't know why it's not disabled by default. > Yes, it may be something that some people want to do but I would have thought that most people *don't* want it. -- Chris Green · |
I think that the option to update the timestamp (Settings, Configure > digikam, Metadata, "update filestamp when files are modified") SHOULD BE DISABLED BY DEFAULT. I just saw this today and am very sorry that I didn't notice it before. On Thu, Sep 20, 2018 at 11:16 PM Chris Green <[hidden email]> wrote: woenx <[hidden email]> wrote: |
On vendredi 21 septembre 2018 10:22:54 CEST Jack Marxer wrote:
> I think that the option to update the timestamp (Settings, Configure > > > digikam, Metadata, "update filestamp when files are modified") > > SHOULD BE DISABLED BY DEFAULT. > I just saw this today and am very sorry that I didn't notice it before. > No need to yell, it won't replace factual arguments... And i disagree, as a default, the "file modified" timestamp should be updated when files are, well, modified. At least for me, the expected behaviour for the "modification" timestamp is to reflect the last time the file was modified, Anything else would be surprising. And one of the principles of deciding what should be the default is that it should be the least surprising behaviour. That modification timestamp helps when you want to collect a number of edits, or need to remove sidecars: just order by timestamp in a file manager, and you'll have all the original files together. That means there's less risk of deleting an original image file. Don't forget that the image metadata contain the original date the image was created (i.e. the shutter was pressed). That one should never be touched automatically (and isn't, afaik), and manually only for a good reason (e.g. camera clock was wrong). Remco |
In reply to this post by Chris Green
On jeudi 20 septembre 2018 23:07:42 CEST Chris Green wrote:
> woenx <[hidden email]> wrote: > > I don't know about the "signature", other software does the same thing. > > Other software doing the same thing doesn't necessarily make it a good > idea! :-) > But it does allow the software to see what kind of fields can be expected in the metadata, and how to interpret them. And if there's information about the processing, the program and version that created that information is really needed to make sure any replay can be done correctly. Different programs, and sometimes even different versions, can use different algorithms or interpret parameters differently. No information about creating program/version would make storing processing info useless: no reliable way to reuse it, and it is not really human-readable (you see the numbers, but you'll have no idea what they mean). Remco |
But, for instance, let's say I tag a picture with the tag "dog". Is it really
necessary to include which software has tagged it, the version, and update the modification (or creation) date? Specially the last thing. Some older digital cameras did not store the date in EXIF metadata, and the only date was the file date. After digikam scanned my whole collection, all dates were changed to that day's date, so I lost quite a big of information. -- Sent from: http://digikam.1695700.n4.nabble.com/digikam-users-f1735189.html |
On vendredi 21 septembre 2018 12:00:00 CEST woenx wrote:
> But, for instance, let's say I tag a picture with the tag "dog". Is it > really necessary to include which software has tagged it, the version, and > update the modification (or creation) date? > > Specially the last thing. Some older digital cameras did not store the date > in EXIF metadata, and the only date was the file date. After digikam scanned > my whole collection, all dates were changed to that day's date, so I lost > quite a big of information. I agree that there are situations where it's not the right thing to do. And for those cases it is nice if the program can be set to not updating the timestamp. But this was about what to use as a *default*. And I maintain that the default should be "when modifying a file, set the modification timestamp to the time of the modificiation". Oh, and when the dates were changed after *scanning* the collection, you might want to file a bug report. Scanning should not modify the files or their timestamps (as it's only reading the files). But do make sure that it was indeed just a pure scan. Remco |
In reply to this post by Remco Viëtor
Remco Viëtor <[hidden email]> wrote:
> On jeudi 20 septembre 2018 23:07:42 CEST Chris Green wrote: > > woenx <[hidden email]> wrote: > > > I don't know about the "signature", other software does the same thing. > > > > Other software doing the same thing doesn't necessarily make it a good > > idea! :-) > > > But it does allow the software to see what kind of fields can be expected in > the metadata, and how to interpret them. > how Digikam uses the metadata? > And if there's information about the processing, the program and version that > created that information is really needed to make sure any replay can be done > correctly. Different programs, and sometimes even different versions, can use > different algorithms or interpret parameters differently. No information about > creating program/version would make storing processing info useless: no > reliable way to reuse it, and it is not really human-readable (you see the > numbers, but you'll have no idea what they mean). > Yes, but in reality my images have been processed by several different programs. Say I do major image enhancement/changes using Gimp or Photoshop, then some changes to metadata using Photini and then I use Digikam to manage the inmages. Having 'Digikam 5.9' in the metadata would be totally misleading! Also Digikam writes 'Digikam 5.9' into the metadata even when it hasn't changed anything else at all. -- Chris Green · |
totally agree, there should be some kind of differentiationAlso Digikam writes 'Digikam 5.9' into the metadata even when it hasn't changed anything else at all. On 21.09.2018 13:12, Chris Green wrote:
Remco Viëtor [hidden email] wrote:On jeudi 20 septembre 2018 23:07:42 CEST Chris Green wrote:woenx [hidden email] wrote:I don't know about the "signature", other software does the same thing.Other software doing the same thing doesn't necessarily make it a good idea! :-)But it does allow the software to see what kind of fields can be expected in the metadata, and how to interpret them.Possibly, but how many programs are there out there that 'understand' how Digikam uses the metadata?And if there's information about the processing, the program and version that created that information is really needed to make sure any replay can be done correctly. Different programs, and sometimes even different versions, can use different algorithms or interpret parameters differently. No information about creating program/version would make storing processing info useless: no reliable way to reuse it, and it is not really human-readable (you see the numbers, but you'll have no idea what they mean).Yes, but in reality my images have been processed by several different programs. Say I do major image enhancement/changes using Gimp or Photoshop, then some changes to metadata using Photini and then I use Digikam to manage the inmages. Having 'Digikam 5.9' in the metadata would be totally misleading! Also Digikam writes 'Digikam 5.9' into the metadata even when it hasn't changed anything else at all. |
Ok, interesting discussion. If I look at the definition of Exiv2:
Exif.Image.ProcessingSoftware => The name and version of the software used to post-process the picture. Would I understand it so that only with changes to the image data is a software entry. So I would suggest that only the image editor sets the software entries, but also the lossless JPEG rotation. Metadata changes write no digiKam software entrys. Maik Am Freitag, 21. September 2018, 13:28:04 CEST schrieb Stefan Müller: > Also Digikam writes 'Digikam 5.9' into the metadata even when it > hasn't changed anything else at all. > totally agree, there should be some kind of differentiation > > > On 21.09.2018 13:12, Chris Green wrote: > > Remco Viëtor <[hidden email]> wrote: > > On jeudi 20 septembre 2018 23:07:42 CEST Chris Green wrote: > > woenx <[hidden email]> wrote: > > I don't know about the "signature", other software does the same thing. > > Other software doing the same thing doesn't necessarily make it a good > idea! :-) > > > But it does allow the software to see what kind of fields can be expected in > the metadata, and how to interpret them. > > > Possibly, but how many programs are there out there that 'understand' > how Digikam uses the metadata? > > > > And if there's information about the processing, the program and version > that created that information is really needed to make sure any replay can > be done correctly. Different programs, and sometimes even different > versions, can use different algorithms or interpret parameters differently. > No information about creating program/version would make storing processing > info useless: no reliable way to reuse it, and it is not really > human-readable (you see the numbers, but you'll have no idea what they > mean). > > > Yes, but in reality my images have been processed by several different > programs. Say I do major image enhancement/changes using Gimp or > Photoshop, then some changes to metadata using Photini and then I use > Digikam to manage the inmages. Having 'Digikam 5.9' in the metadata > would be totally misleading! > > Also Digikam writes 'Digikam 5.9' into the metadata even when it > hasn't changed anything else at all. |
https://cgit.kde.org/digikam.git/commit/?
id=f969875cb3a9ff70e927364f5d406229505b70fd Maik Am Freitag, 21. September 2018, 21:43:36 CEST schrieben Sie: > Ok, interesting discussion. If I look at the definition of Exiv2: > > Exif.Image.ProcessingSoftware => The name and version of the software used > to post-process the picture. > > Would I understand it so that only with changes to the image data is a > software entry. So I would suggest that only the image editor sets the > software entries, but also the lossless JPEG rotation. Metadata changes > write no digiKam software entrys. > > Maik > > Am Freitag, 21. September 2018, 13:28:04 CEST schrieb Stefan Müller: > > Also Digikam writes 'Digikam 5.9' into the metadata even when it > > hasn't changed anything else at all. > > totally agree, there should be some kind of differentiation > > > > > > On 21.09.2018 13:12, Chris Green wrote: > > > > Remco Viëtor <[hidden email]> wrote: > > > > On jeudi 20 septembre 2018 23:07:42 CEST Chris Green wrote: > > > > woenx <[hidden email]> wrote: > > > > I don't know about the "signature", other software does the same thing. > > > > Other software doing the same thing doesn't necessarily make it a good > > idea! :-) > > > > > > But it does allow the software to see what kind of fields can be expected > > in the metadata, and how to interpret them. > > > > > > Possibly, but how many programs are there out there that 'understand' > > how Digikam uses the metadata? > > > > > > > > And if there's information about the processing, the program and version > > that created that information is really needed to make sure any replay can > > be done correctly. Different programs, and sometimes even different > > versions, can use different algorithms or interpret parameters > > differently. > > No information about creating program/version would make storing > > processing > > info useless: no reliable way to reuse it, and it is not really > > human-readable (you see the numbers, but you'll have no idea what they > > mean). > > > > > > Yes, but in reality my images have been processed by several different > > programs. Say I do major image enhancement/changes using Gimp or > > Photoshop, then some changes to metadata using Photini and then I use > > Digikam to manage the inmages. Having 'Digikam 5.9' in the metadata > > would be totally misleading! > > > > Also Digikam writes 'Digikam 5.9' into the metadata even when it > > hasn't changed anything else at all. |
In reply to this post by Maik Qualmann
Maik Qualmann <[hidden email]> wrote:
> Ok, interesting discussion. If I look at the definition of Exiv2: > > Exif.Image.ProcessingSoftware => The name and version of the software used to > post-process the picture. > > Would I understand it so that only with changes to the image data is a > software entry. So I would suggest that only the image editor sets the > software entries, but also the lossless JPEG rotation. Metadata changes write > no digiKam software entrys. > the metadata really shouldn't be leaving its 'fingerprint' behind. -- Chris Green · |
I understand that Digikam writes to sidecar what he needs to write, and no
more. GIMP does it In fact, as someone, wrote here, Digikam does NOT write to sidecar while scanning. If so, I should noticed it each time I launch Digikam, with a +300K photos database like mine, cause it will be writing to sidecars in the disk. However, my case (http://digikam.1695700.n4.nabble.com/digiKam-users-Anyone-with-a-300K-photos-internal-mysql-database-td4706440.html) is not related with sidecars (at first sight). -- Sent from: http://digikam.1695700.n4.nabble.com/digikam-users-f1735189.html |
Free forum by Nabble | Edit this page |