AtmoLight 2.1.0.0 for MediaPortal2 [2015-01-21] (2 Viewers)

[kilik360]

Just to let you know that I use Beta 2 with MP2 10th AE Update 1, and so far so good Lightning303 !

Thanks a lot !

 

[Lightning303]

The MP2 version of AtmoLight gets an update also, of course .

With this update, we close the gap to AtmoLight 1.x, as now everything the MP1 version can do, the MP2 version can do aswell.

Information about now supporting Boblight, Hue and Hyperion, aswell as some other changes can be found here: https://forum.team-mediaportal.com/threads/atmolight-1-14-0-0-2014-11-23.125633/page-6#post-1111698


On top of that, these are some of the changes exclusive to the MP2 version:

You can now change the settings for AtmoLight without having to restart MP2. Before lots of the settings would not be considered until MP2 got restarted. This also means that you can add or remove targets (e.g. AtmoWin) on the fly.

We added the VUMeter and VUMeter Rainbow effect from the MP1 version. This effect will light up your leds according to the music you are listening to.

Blackbar detection and removal has been added. If you are watching an old tv show or a bluray, and the blackbars are encoded in the video, then AtmoLight can now remove them, so that the leds light up even with the bars. Note that this will not change how the video looks on your screen.

We fixed a bug where AtmoLight would not change the effect if a video ended on its own.

Finally, we added Picture-in-Picture support. The leds will always light up according to the primary videoplayer, if you switch the player, then the lights will switch aswell. Also AtmoLight now handles the effects correctly when you start or stop a player in PiP mode.

 

[morpheus_xx]

@Lightning303 can you please check this error? The latest release has changed player code (using Texture directly, not a copied Surface):
Methode nicht gefunden: "SharpDX.Direct3D9.Surface MediaPortal.UI.SkinEngine.Players.ISharpDXVideoPlayer.get_Surface()".
at AtmoLight.Plugin.UICapture(Object sender, EventArgs args)
at MediaPortal.UI.SkinEngine.DirectX.GraphicsDevice.Fire(EventHandler eventHandler)


Besides this error also the plugin.xml needs to be adjusted to match new compatibility management. Can you check this, please?

 

[morpheus_xx]

@Lightning303 did you find a moment to update the plugin for last SR'15?

I can offer a patch if you like. Changes are minimal. We plan to release an Update build soon, would be great if the plugin works again

 

[Lightning303]

Hey,
no i didnt take a look at it yet. If you have a patch and would provide it, i could give it a try and make a new version after that. Will probably be easier and faster .

 

[morpheus_xx]

Before I checkout source, I can post the changes:

https://github.com/ambilight-4-medi...er/AtmoLight.MediaPortal2/Plugin.cs#L442-L444
old: surfaceSource = player.Surface;
new: surfaceSource = player.Texture.GetSurfaceLevel(0);

I think you need to Dispose the returned Surface by GetSurfaceLevel, but only the one from player, not the RenderTarget.

A conceptional question: why do you take the video surface if it is present? Isn't it better to capture the finished rendered backbuffer? It should give you better results, i.e. if there are 2 PiP video players, the actual screen position of video frames need to be considered (fullscreen background+PiP window overlay).

 

[Lightning303]

why do you take the video surface if it is present?

I dont even know what that means , "if it is present". As said before, everything directx and so on is pretty much just me trying everything i can think of until it works. MP1 uses surfaces, so thats what i was looking for in MP2. If you know of a better solution (im sure you do), please, go for it. Implement it and either provide me with a patch, or i can give you the rights on GitHub so you can just push it. But giving me hints on how to make it better is not going to give us any results, as i dont have the background in all this render stuff and currently using my free time for other things .
If you want to leave it up to me, thats fine aswell, and im going to do it, at some point. But i have no idea when.

 

[morpheus_xx]

I dont even know what that means , "if it is present".

I was refering to the existing code parts.
Here you take the VideoPlayer's frame surface as source: https://github.com/ambilight-4-medi...er/AtmoLight.MediaPortal2/Plugin.cs#L438-L445.

And if no player is active you take the back buffer (full rendered frame): https://github.com/ambilight-4-medi.../master/AtmoLight.MediaPortal2/Plugin.cs#L448

A generic solution in all case (no video, 1 or 2 video players) is the latter one, the finished rendered back buffer frame.

 

[Lightning303]

I dont even know what that means , "if it is present".

I was refering to the existing code parts.
Here you take the VideoPlayer's frame surface as source: https://github.com/ambilight-4-medi...er/AtmoLight.MediaPortal2/Plugin.cs#L438-L445.

And if no player is active you take the back buffer (full rendered frame): https://github.com/ambilight-4-medi.../master/AtmoLight.MediaPortal2/Plugin.cs#L448

A generic solution in all case (no video, 1 or 2 video players) is the latter one, the finished rendered back buffer frame.

I see, but my comment in the code should then show why i used 2 different approaches. The backbuffer was the first approach i ever did. However using the frame surface is faster, or atleast it was faster when i implemented it, and it has the nice effect that blackbars that are not encoded into the video will not be in the frame, as far as i can remember this also is true for osd and so on.
So whenever a frame surface from the videoplayer is available, it gets used, because its more desireble. Otherwise we fall back to the backbuffer which gives us everything we can see on screen (e.g. while browsing menus, looking at pictures and so on).

 

[Lightning303]

Found some time to have a look. Here is an intermittent version which works with the snapshot from 2015-08-01.
As soon as we are ready to release a new AtmoLight version for MP1, we will also update this one properly.