[thomas.vriens]

Hi all, I'm new to the codebase of XBMC but not new to XBMC as a user. I am a developer by trade and it's one of my main hobbies as well.

Currently I am trying to understand the way XBMC translates video characteristics to media flags while with my own tests the encoded properties of a file don't always have a 100% match with 'proper' resolution for example but XBMC manages to match it anyway.

I know the images themselves are skin specific but from the XML of several skins I can tell that there is very little mapping involved for the flag values on the skin side, so it must be in XBMC itself. Any help to locating the part of XBMC that does this processing would be appreciated. I have spent a few hours on it already but since I'm new to the codebase and there is a lot of code to go through it's been like looking for a needle in a haystack.

Regards,
Thomas

[ronie]

xbmc presents the video characteristics to the skin using $INFO[] labels.
these labels gets set in GuiInfoManager.cpp

the detection of the video characteristics itself is done in StreamDetails.cpp i think...

[CapnBry]

the detection of the video characteristics itself is done in StreamDetails.cpp i think...

This is accurate that that's where they are published and the actual video width and height are quantized to standard values. "VideoResolution" is determined using only the closest width, and the "VideoAspect" ratio comes from a mean error (if I remember correctly) distance from standard values.

The data that feeds that comes from the database but it is fed into the database by the background thumb loader which uses values that come from the player cores. When a video is played and there are no StreamDetails, the data from the active player is stored to the database, which used to happen in Application.cpp but it may have moved since I wrote it like... years ago.