Call for tester for 1.4.0 features and fixes (1 Viewer)

[Owlsroost]

Pure device was enabled to get a really tiny performance boost. We should either fix the code to detect the GPUs that won't support hardware vertex processing or remove the Pure device + HardwareVertex change (only the flags since the DirectX debug logging will blow up again and fill the whole world with warnings if we remove the state manager ). Hopefully VertexProcessingCaps will tell enough info to make the sw vs. hw vertex flag selection possible.

A quick comment - there are probably a lot of laptops still in use with older Intel GPUs, which people run MP on mostly for TV use - we would lock all these users out for (if I understand the comments correctly) only a very small performance gain for other users. Also it's not unusual for potential new users to try out MP on an older system - if it won't run on what is still a reasonably modern laptop (or SFF system with no expansion slots) they'll probably move on to other HTPC software....

(Although I wouldn't run MP on it for real use, I installed Win 8 on an old Core2Duo 2 GHz + 945G chipset laptop recently - the video playback performance is actually not that bad - much better than I remember it on it's original Vista OS, it can even manage 1080i video at sensible CPU usage )

 

[Atomic7431]

has anyone tested this with the Djblu sky plugin? and when will all the new changes since the 18a build be made into another build so i can help test this amazing work

 

[AtTimesPerplexed]

OSD progress bar problem is solved in the latest branch. Thanks.

 

[Scythe42]

made tests with your new mediaportal.exe! we are getting closer i think

Looks like it. Attached is the next one.

has anyone tested this with the Djblu sky plugin? and when will all the new changes since the 18a build be made into another build so i can help test this amazing work

New build will come when the current known issues are sorted out.

Hopefully VertexProcessingCaps will tell enough info to make the sw vs. hw vertex flag selection possible.

Yes, that can be detected and the creation flags adjusted accordingly.

@Scythe42, it's working now but i can see a lag when switching from fullscreen to windowed (log attached) when playing video too (it's no as smooth as before lol)

That's Windows itself. Now that some multi display related handling is fixed it goes through a few processing steps on the WinForm size probably before displaying it to user. Not something I intend on working on. I think we can live happily when switching between full screen and windowed mode takes a frame or two longer compared to seconds as before.

The branch does not merge cleanly to master now.

Merged master to my branch and resolved the conflict.

 

[Scythe42]

I found another strange behavour

In my Environment ISO-Videos are played with TMT5. The remote control is done by IRSS. I use the Standard Video plugin.
With the final 1.3 everything is fine.
TMT5 is started, gets the focus and the controll is done by IRSS TMT settings.
IF I close TMT, MEPO gets again the focus.

Do you have always on top enabled?

With build 18 the focus remain at MEPO if TMT is started.
Indepent from the Setting "...minimize to tray."
I need to press alt-tab so that TMT gets the focus and I can controll it with IRSS TMT Settings.

I can reproduce it by changing the files of final or build 18.

Can you try without the refresh rate changer? Just trying to limit possibilities down here.

[2013-04-14 15:37:11,012] [29884  ] [MPMain  ] [DEBUG] - W7RefreshRateHelper.SetDisplayConfig(...): SDC_VALIDATE of 23/1 failed
[2013-04-14 15:37:11,015] [29887  ] [MPMain  ] [DEBUG] - W7RefreshRateHelper.GetRefreshRate: QueryDisplayConfig returned 74175824/3093750
[2013-04-14 15:37:11,015] [29887  ] [MPMain  ] [INFO ] - CycleRefreshRate: successfully changed refresh rate to 23.976Hz (23Hz requested)
[2013-04-14 15:37:11,016] [29888  ] [MPMain  ] [DEBUG] - CycleRefresh: DWM Detected, performing shenanigans
[2013-04-14 15:37:11,017] [29889  ] [MPMain  ] [INFO ] - g_Player.Play(G:\BDMV\index.bdmv Video)
[2013-04-14 15:37:11,019] [29891  ] [MPMain  ] [INFO ] - start process C:\Program Files (x86)\ArcSoft\TotalMedia Theatre 6\uMCEPlayer6.exe "G:\"
[2013-04-14 15:37:11,393] [30265  ] [MPMain  ] [DEBUG] - Main: WM_ACTIVATE
[2013-04-14 15:37:11,394] [30266  ] [MPMain  ] [INFO ] - Main: Deactivation Request Received
[2013-04-14 15:37:11,394] [30266  ] [MPMain  ] [INFO ] - Main: Refresh rate changer running. Ignoring deactivation request
[2013-04-14 15:37:11,394] [30266  ] [MPMain  ] [DEBUG] - D3D: OnLostFocus()
[2013-04-14 15:37:16,898] [35771  ] [MPMain  ] [DEBUG] - Main: WM_DISPLAYCHANGE
[2013-04-14 15:37:16,899] [35772  ] [MPMain  ] [INFO ] - Main: Resolution changed to 1920x1080x32 or displays added/removed
[2013-04-14 15:37:16,901] [35774  ] [MPMain  ] [DEBUG] - Main: WM_DEVICECHANGE (Event: 32772)

Windows sends out a WM_DISPLAYCHANGE message. That's strange. It happens roughly five seconds after the external player was started.

MP reacts to it and resizes itself as instructed by Windows (needed for the general HDMI fix, that's why these builds behave different than 1.3.0), probably bringing it into focus again.

Do you have Always On Top Enabled?

Are you doing anything related to turning on/off screen or changing the resolution? Configuration inside TMT that does something to the display?

Can you try with a different player than TMT? Maybe TMT is doing something here to the display it should not or is configured in that way. My best guess would be it changes the screen properties when playback begins 5 seconds after the start.

Same when it stops:

[2013-04-14 15:38:17,220] [96093  ] [MPMain  ] [INFO ] - Main: Resolution changed to 1920x1080x32 or displays added/removed
[2013-04-14 15:38:17,222] [96095  ] [MPMain  ] [DEBUG] - Main: WM_DEVICECHANGE (Event: 32768)
[2013-04-14 15:38:18,160] [97033  ] [MPMain  ] [DEBUG] - Util: External player stopped on C:\Program Files (x86)\ArcSoft\TotalMedia Theatre 6\uMCEPlayer6.exe

@Testers: anyone else out there with TMT?

 

[boldie]

It is the change of the Resolution !

The reason why I use an external Player is to watch 3d BlueRay.
Therefore TMT (and I think all other external Players) changing the resolution for 3d bluerays.

If I Change the Default refresh rate of MEPO to 24 HZ (the same as the 3D blueray Resolution) it works.

 

[Scythe42]

Sure, full understand why you use an external player for BluRays as libbluray does not fully support menus on blurays.

The problem here is the not matching refresh rate. MP sets it to one with the refresh rate changer and then TMT resets it again to a different refresh rate. Or MP does not set it all, and the player itself does it, which for 3D content is even more important.

This change causes Windows event that the display changed. In this specific scenario it is only the refresh rate. Accordingly MP reacts to this and as a result become the top most form as it is this still the parent of the external player.

I could catch this in code with some dirty hack, but I think it is not worth the effort and should better be properly implemented.

The problem is caused by the fact that refresh rate changer and external player are doing to different things here. The external player resets the refresh rate set by MP because it says it is wrong for the content it is trying to play. Probably the Movie itself has a different RR than what is detected by mediainfo for the BluRay in general.

@Testers: Can someone please Mantis these issues for future reference? Tweaking the refresh rate changer is out of scope for this work.

Mantis issues:
- Make MP aware of changed refresh rates not changed by the refresh rate changer
- Keep external players in focus when they change the refresh rate

Both childs of #4083

Why does it work with stock MP 1.3.0: Because 1.3.0 is not aware of changes to the display. This is one of the causes for the initial "half-full-screen" issue. And when a display change is detected, MP will resize itself. And therefore the parent window will overlay the child window that runs the external player.

Edit:
I am adding a small workaround in WM_DISPLAYCHANGE checking if the actual resolution changed. The problem is that WM_DISPLAYCHANGE does not report the screen's current refresh rate.

I hope this does not break the Half-Full-Screen issue but could solve at least the focus change. Still MP does not know of the overridden refresh rate that might cause subsequent problems to the refresh rate changer.

@Testers: Please check current GIT (if you can compile MP yourself), if this workaround breaks the half-full-screen fix. Already pushed to GIT.

Change itself: https://github.com/MediaPortal/MediaPortal-1/commit/e36c58fbbd6f7cc201bfac437cc2d586d9a1f155

 

[Scythe42]

So this leaves us with the following:
- Verifying that all scenarios for multiple displays work properly
- Add device capability checking to decide if we can use a pure device or not - and if not log a warning about the GPU and create a compatible device.

Maybe I look into the PS issue and this "Away Mode" in regards to properly unmuting sounds.

But I don't know yet. It's really out of scope. And to be frank: this special "Windows Media Center" mode should not be part of MP in the first place. It is not like it is an official power saving state or anything. And from my understand it brings nothing to the table. It might have been OK at the time XP was introduced as hardware power management was really limited and also XP itself is.

 

[azzuro]

- 0004352: [Dynamic Refresh Rate] Keep external players in focus when they change the refresh rate - new.
- 0004351: [Dynamic Refresh Rate] Make MP aware of changed refresh rates not changed by the refresh rate changer - new.

 

[mbuzina]

And to be frank: this special "Windows Media Center" mode should not be part of MP in the first place. It is not like it is an official power saving state or anything. And from my understand it brings nothing to the table. It might have been OK at the time XP was introduced as hardware power management was really limited and also XP itself is.

I would not say so. The Away Mode makes MP behave like an appliance and that is what it "should" be for me. The basic thing is that a shutdown request can not be aborted by normal software since vista (or 7). So the only thing we can do, when recording is to register for away mode. This way even if the user presses the power switch, the PC will stay awake (just shut off sound & display) and continue recording.