https://bugs.kde.org/show_bug.cgi?id=388936

--- Comment #2 from aer...@aerysl.net ---
(In reply to caulier.gilles from comment #1)
> Can you give me a very good reason to provide a way to change the database
> image format in thumbnail database ?
> 
> - PGF is used because it use wavelets compression to optimize space,
> quality, speed.
> - the other format support provided in digiKam core for thumbs database is
> present for testing only. It's a legacy of test done many years ago when we
> choose the right one.
> - there is currently no mechanism to detect the thumb format stored in DB.
> including this king of control will slow down the thumb extraction, and we
> won't to complex this task.
> 
> For me PGF is prefect and there is no reason to changes it. The only
> possible way to change the format id to use a new compression which provide
> a better optimisation than PGF. Perhaps HEIF, WepP, or JPEGXR, but none
> provide an easy to use through a small library under Linux as PGF. There are
> also patents problem in background...
> 
> Gilles Caulier

One use case that I can think of (and am currently working on, as a personal
project), is that the I'd like to be able to use my Digikam database on other
platforms (say my Android phone, or via HTTP), so I have been working on a PHP
frontend of sorts. As such I have done local builds with JPEG thumbnails
(simply changing the previously mentioned line, and regenerating all my
thumbnails), so I can directly serve the thumbnails in the database extremely
quickly. I imagine others may find this useful, but it is a pretty niche case.

I definitely agree with PGF being a great format for use within Digikam, but
unfortunately I could not find many other libraries to interface with them.

I'm using the MySQL (MariaDB) backend, but it appears to me the "type" field in
the "Thumbnails" table corresponds with the type - and have had success using
JPEG based thumbnails. However, I am completely ignorant of how performant this
is, and any other considerations involved, and my observations may be
completely incorrect.

Thanks

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to