I've been trying to build this branch however I am coming into an error:
System.IO.FileLoadException was unhandled
Message=Could not load file or assembly 'PlayerServer,
Version=1.0.1.39032, Culture=neutral, PublicKeyToken=null' or one of
its dependencies. The application has failed to start because its
side-by-side configuration is incorrect. Please see the application
event log or use the command-line sxstrace.exe tool for more detail.
(Exception from HRESULT: 0x800736B1)
Source=Kinovea.ScreenManager
FileName=PlayerServer, Version=1.0.1.39032, Culture=neutral,
PublicKeyToken=null
FusionLog=""
StackTrace:
at Kinovea.ScreenManager.ScreenManagerKernel..ctor()
at Kinovea.Root.RootKernel.BuildSubTree() in
D:\Programming\Kinovea\Root\Kernel.cs:line 180
at Kinovea.Root.RootKernel..ctor() in
D:\Programming\Kinovea\Root\Kernel.cs:line 133
at Kinovea.Root.Program.Main() in
D:\Programming\Kinovea\Root\Program.cs:line 69
at System.AppDomain._nExecuteAssembly(Assembly assembly,
String[] args)
at System.AppDomain.ExecuteAssembly(String assemblyFile,
Evidence assemblySecurity, String[] args)
at
Microsoft.VisualStudio.HostingProcess.HostProc.RunUsersAssembly()
at System.Threading.ThreadHelper.ThreadStart_Context(Object state)
at System.Threading.ExecutionContext.Run(ExecutionContext
executionContext, ContextCallback callback, Object state)
at System.Threading.ThreadHelper.ThreadStart()
InnerException: System.Runtime.InteropServices.COMException
Message=The application has failed to start because its
side-by-side configuration is incorrect. Please see the application
event log or use the command-line sxstrace.exe tool for more detail.
(Exception from HRESULT: 0x800736B1)
ErrorCode=-2147010895
InnerException:
I'm using C#2010 with the projects set to 3.5 like before. When I use
the old playerserver.dll it will load (however of course it is missing
the required features for playing the video), so something in the next
configuration is not working. Would you have an idea what is
missing? (I'm not a .net programmer normally ;)).
*From:*kinovea-dev-bounce@xxxxxxxxxxxxx
[mailto:kinovea-dev-bounce@xxxxxxxxxxxxx] *On Behalf Of *Joan
*Sent:* Tuesday, 22 March 2011 5:28 AM
*To:* kinovea-dev@xxxxxxxxxxxxx
*Subject:* [kinovea-dev] Re: Direct2D
Yes, the reverse operation. After the frame is decoded by FFmpeg, it
now goes straight to a Direct2D bitmap. But I couldn't find an easy
way to get a regular GDI Bitmap out from it, to apply image filters
for example.
I think one have to go through something called WIC or something, but
that is not implemented in SlimDX yet.
Maybe there's another solution... Or maybe asking FFmpeg again and
getting a GDI+ Bitmap just for the operation...
If you can think up of something...
Le 21/03/2011 18:47, Hugh a écrit :
Nice!
By getting a GDI image, do you mean the reverse of this temporary
conversion function that I was using before?
These days extra memory like this is not an issue, and if in the
future the whole process is faster, the buffer does not need to cover
so much. So I think this is not an issue.
Hugh
*From:*kinovea-dev-bounce@xxxxxxxxxxxxx
<mailto:kinovea-dev-bounce@xxxxxxxxxxxxx>
[mailto:kinovea-dev-bounce@xxxxxxxxxxxxx] *On Behalf Of *Joan
*Sent:* Saturday, 19 March 2011 10:08 PM
*To:* kinovea-dev@xxxxxxxxxxxxx <mailto:kinovea-dev@xxxxxxxxxxxxx>
*Subject:* [kinovea-dev] Re: Direct2D
Hi,
Update on the Direct2D front.
First I made some changes to the trunk to keep the diff with the
branch to a minimum.
The most important is decoding directly to BGRA pixel format (r406 and
r407).
This means a 25% increase in memory requirements (32bpp against
24bpp), so for similar memory settings in preferences, the working
zone will be 25% shorter in duration.
On the other hand, it gives about 10% speed boost on the rendering
part, depending on image size. (because it's the .NET internal format
so an extra copy is avoided).
This should be most visible for big image size when images have been
extracted to memory already.
Then I went on to clean up the Direct2D sandbox code I had to create
the branch.
- After decoding, the image is stored by the Renderer itself (in its
preferred format to avoid any further conversion).
- Then we call Render() which use the stored image.
However, several problems:
1. There is a flickering issue.
2. I don't know how to get back a GDI Bitmap from the Direct2D code.
There 31 instances of code where this is needed (to create the key
image thumbnail, to save to disk, etc.)
Branch is created here :
http://svn.codingteam.net/kinovea/branches/Direct2D/
Check the new folder ScreenManager/PlayerScreen/Rendering/...
Some experiments to do :
0. Switch Renderers in FrameServerPlayer constructor.
1. Painting on the Panel or on the PictureBox ?
The control to paint on is passed to
m_FrameServer.Renderer.Initialize() in the PlayerScreenUserInterface
constructor.
Then a call to .Render() must be done depending on the paint
invocation method (see below).
2. Calling the rendering directly or through Control.Invalidate() ?
(See DoInvalidate())
3. in GDI+, Getting the Graphics object from the Paint event or from
the handle of the control we draw on (flickering issue ?).
Note that tests must be done without triggering the memory caching of
frames. (use long video or low values in preferences)
Any function needing a Bitmap copy of the current frame will likely
crash for now :-) (key images, saving, etc.)
Let me know if you try some experiments or if anything seems unclear
or strange.
Thanks
joan.
Le 10/03/2011 10:40, Hugh a écrit :
Cool! Sounds like you are making some serious in roads to getting it
working smoothly.
If the images are held in the render I think it is ok, as long as the
different display classes can tell it how to treat this image. I
think the best is when they are all rendered in one single output
controller and it is fed a list of classes to go through. Then the
images can more easily be rendered on top of each other, rotated
around etc. Sometime a class could be made for image morphing, and
then it would be the one passed to this output controller. When it is
called, it can then hold the child classes and be an intermediate that
does the work.
I hope you can understand what I mean.
I can't wait to see the code in the tree so I can mess a little with
it ;) However time is still my enemy for the next few weeks.
Good work with the progress! For me this is the biggest thinking
holding the program back, as I really would love to be able to play
videos from the cameras I use directly. But as they get higher
quality, the program needs to keep up.
*From:*kinovea-dev-bounce@xxxxxxxxxxxxx
<mailto:kinovea-dev-bounce@xxxxxxxxxxxxx>
[mailto:kinovea-dev-bounce@xxxxxxxxxxxxx] *On Behalf Of *Joan
*Sent:* Sunday, 27 February 2011 9:17 PM
*To:* kinovea-dev@xxxxxxxxxxxxx <mailto:kinovea-dev@xxxxxxxxxxxxx>
*Subject:* [kinovea-dev] Re: Direct2D
Follow up...
I managed to remove the temporary Bitmap conversion :-)
In the end it was mostly a matter of finding the right pixel format to
go along.
For some reason, I had always been converting from FFmpeg to a Bitmap
in Format24bppRgb, which is not optimal since the internal Bitmap
format seems to be Format32bppPArgb.
So the first change was to convert from the internal FFmpeg format
(YUV420P) to BGR32 (the naming of the Drawing.Bitmap enum is reversed)
instead of BGR24.
This will also positively impacts the GDI+ code.
The D2D code is expecting data in a similar format than Drawing.Bitmap
(here it's called DXGI_FORMAT_B8G8R8A8_UNORM) - so I could use the
existing snippet for Bitmap conversion, providing the raw data from
the FFmpeg decoding.
The architecture is still a bit tentative (now it's the Renderer that
holds the image to be drawn)
I'm having rendering issues though, there are black frames once in a
while. I don't know where this is coming from.
(I went with the asynchronous .Invalidate() call on the background
panel for these tests.)
Anyway, I'll create a branch next week and push what I have so far.
joan.
Le 22/02/2011 14:20, Hugh a écrit :
Cool. Sorry for the late replay, I've been crazy busy recently. I
was wondering if you had any luck getting the decoding straight onto
the directX style bitmap?
*From:*kinovea-dev-bounce@xxxxxxxxxxxxx
<mailto:kinovea-dev-bounce@xxxxxxxxxxxxx>[mailto:kinovea-dev-bounce@xxxxxxxxxxxxx]
*On Behalf Of *Joan
*Sent:* Wednesday, 16 February 2011 6:25 PM
*To:* kinovea-dev@xxxxxxxxxxxxx <mailto:kinovea-dev@xxxxxxxxxxxxx>
*Subject:* [kinovea-dev] Re: Direct2D
Just a quick note to say that I'll try to focus on the Direct2D
rendering stuff for a while.
Hopefully, to get some sort of a working base that we can build upon
and refactor later...
joan.
Le 14/01/2011 23:42, Hugh a écrit :
I'm in a rush, so just quickly:
As long as there is a sound way to destroy it and any threads it may
create, then all is good! I just came from outside C# so not so use
to it all.
The DrawVideoBitmap() was very much a hack to try and get it to work
in the current program. It would be good if it can draw straight onto
a DirectX style object, and do all rendering on it (there are many
functions for it). Obviously the more it is in the hardware (and not
just copied over), the best results will occur.
I don't know enough about FFMPEG to know if it is possible to draw
from it onto these surfaces, but I'm sure it must be since VLC is very
quick at rendering.
The whole class was designed to have a module feel to it, but also to
fit into the current program. I did something close to what you have
to do the rendering, but I hope the class can easily be changed to fit
any new designs. As I did more work on it, I realised there were
other bottle necks in the program that would need to be fixed get it
fully working (like drawing onto DirectX objects, new rendering
process etc).
I left the other classes (like DirectX 9 and 10) mainly because
Windows XP does not support Direct2D. However I guess there is a time
people must decide to move on. They would be much harder to work with
as you must draw the shapes in the perspective, not simply simple
objects. Direct2D also has many other small advantages like for
remote desktop, integration etc. I never tested them, but didn't want
to remove the options in case someone wanted to go down this path in
the future. Personally I favour Direct2D the most.
At the moment I am rather limited for time (sorry!), but this is
something I would love to see in the program.
Hugh
*From:*kinovea-dev-bounce@xxxxxxxxxxxxx
<mailto:kinovea-dev-bounce@xxxxxxxxxxxxx>[mailto:kinovea-dev-bounce@xxxxxxxxxxxxx]
*On Behalf Of *Joan
*Sent:* Thursday, 13 January 2011 10:19 PM
*To:* kinovea-dev@xxxxxxxxxxxxx <mailto:kinovea-dev@xxxxxxxxxxxxx>
*Subject:* [kinovea-dev] Re: Direct2D
Hi,
I have tried the Direct2D / SlimDX code you sent.
I wasn't too sure as to how exactly use the control, here is what I did:
- Added a Direct2DControl member to PlayerScreenUserInterface.
- In the PlayerScreenUserInterface constructor I call
Direct2DControl.SetControl() passing in the center panel (to test on
the back panel for now)
- immediately call Direct2DControl.Run() to trigger initializations.
- At each new frame, (in PlayerScreenUserInterface.ShowNextFrame), I
call Direct2DControl.DrawVideoBitmap() passing in the decoded image,
- and Direct2DControl.Update() to actually draw on the panel.
Let me know if I did anything wrong. Did you also modify
PlayerScreenUserInterface in this way for your tests ?
It seems to work pretty well.
I've tried to instrument and get some figures regarding speed, but it
proved harder than I thought to really pinpoint the rendering time.
(Also I've recently changed machine and it's like everything is
running super fast :-))
The part taking the most time (about 10 to 20ms each frame on my
machine) would seem to be DrawVideoBitmap().
This function involves copying the whole GDI+ bitmap into a Direct2D
bitmap for later use. This is a step we can probably avoid completely
by decoding directly into a Direct2D bitmap from the FFMpeg code.
This is maybe what you suggested in the forum thread... In the case of
pre-extraction to memory, we can also probably keep an in-memory array
of Direct2D bitmaps instead of an array of GDI+ bitmaps. I'm sure
we'll find a way to avoid this extra copy.
About your concern to release resources, I don't think we need to use
a separate form for rendering and react to the OnClose() event of it.
Screens must implement a BeforeClose() method (see AbstractScreen)
which is triggered when the screen is closed (close button, changing
screens configuration, etc.)
We should be able to use it to ask the PlayerScreen and the renderer
to dispose any outstanding resources.
We'll probably need to transform and simplify the DirectXInterface class.
Also, I don't know if we need the other classes, (Direct3D10Context,
etc.) ? If not, we might even merge the two for simplicity.
I haven't looked at the direct2D drawing primitives and how we might
implement the drawing tools visuals.
The architecture discussion will probably depend on how these can be
abstracted between Direct2D and GDI+.
I will create a topic on the wiki to push some schematics of possible
architectures so we are clearer about what we discuss.
Let me know of your experiments.
Thanks,
joan.
Le 10/01/2011 21:34, Joan a écrit :
Hi!
I am super sorry for the delay.
Unfortunately, I haven't had much time to look into this, but I will.
(I have some issues at work with the company wanting to close the
local office, this has somewhat prevented proper focus on the project,
but it's going to resolve one way or the other in the next few weeks.)
Thanks
joan.
Le 05/01/2011 17:34, Hugh a écrit :
Hey Jean,
Was just wondering how the Direct2D code was working for you? Did you
have any luck moving the drawing onto a new form? Just didn't see any
recent changes in the code and was hoping it is working!
Thanks,
Hugh
