Manage video support progress in 6.0.0
caulier.gilles at gmail.com
Mon Mar 12 08:45:25 GMT 2018
The video metadata reader based on ffmpeg is mostly completed in 6.0.0
branch. I will start soon to write the second part : the video metadata
There are some point reported in bugzilla since a very long time about
advanced rename operations (BQM, Import, dialog) which fails with video
file, especially when timestamp is used.
The date is very well extracted with ffmpeg. This problem will probably
fixed with 6.0.0.
>From bugzilla :
These one are certainly fixed with Exiv2 and video support but as this
Exiv2 version is not suitable, it must be closed again with 6.0.0 :
2018-03-11 12:09 GMT+01:00 Gilles Caulier <caulier.gilles at gmail.com>:
> Hi all,
> Some feedback here about the video support in digiKam.
> I can said that 6.0.0 will have really this feature, compared to older
> releases where the missing Exiv2 video support was a big black hole.
> Before digiKam 6.0.0, video support was not usable in the goal to manage
> video as photo.
> Now, since we switch all video metadata suport from Exiv2 to FFmpeg, we
> boost the application to another space.
> I tested with my big collection of video (MP4, MOV, MKV, AVI - more than
> 3000 items), and i can said that scan on multi-core is fast and stable. The
> video metadata extraction is powerfull and populate properly the database
> with the suitable information to perform a video search.
> All video format supported by FFMpeg will be registered to the database.
> The metadata extraction is optimum for MP4 and MOV, and currently i
> finalize the wrap of RIFF files metadata to database (AVI and more).
> Note : WebP image format from Google dedicated to use on internet is also
> a RIFF based container, so it will be automatically supported in digiKam.
> The MKV metadata support is partial and i plan to optimize this part in
> the future.
> Note : The HTML5 video format WebM from google based on Matroska will be
> also automatically supported by digiKam.
> This is a search view screenshot over my database :
> The video thumbnailer is now based on a pure FFMpeg implementation and
> work well, compared to QtAV frame extractor.
> The plan for the future is to support the metadata writting in video file.
> With FFMpeg it's not too much complicated in fact, especially with whole
> XMP data to store as text in a metadata text field. This want mean to not
> touch internal video metadata with digiKam but to create an XMP bytearray
> in a dedicated chunck (something like done with PNG or JPEG in fact). but
> it's for the future...
> Voilà. Have a good week end...
> Gilles Caulier
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Digikam-devel