Understanding Video Renderers

[jameson_uk]

At a high level I understand the technical differences between the various renderers but I am struggling to understand what this means in practice.

I know Overlay is the old renderer and is quite simple. VMR7 is the default renderer in XP and uses directdraw. VMR9 extends and changes VMR7 slightly. Both VMR have windowed, windowless and renderless modes. EVR is the default renderer in Vista and can be got to work in XP (with a little jiggery pokery) but there are issues with DXVA and EVR in XP?

Next I see that DXVA2 is only supported in EVR and Vista and a lot of hardware de-interlacing seems to require DXVA2??

So first question is whether anyone can explain the differences in renderers in terms of actual playback.

[JohnnyMalaria]

Enhanced Video Renderer:

http://msdn.microsoft.com/en-us/library/ms694916(VS.85).aspx

Using the Video Mixing Renderer:

http://msdn.microsoft.com/en-us/library/dd407294(VS.85).aspx

Answering the question you pose isn't straightforward since it depends a lot on what is going on upstream in the filter graph. The above MSDN references provide the technical detail on what they do. The EVR page has a small diagram that shows that it uses the Direct3D pipeline and not the DirectDraw pipeline (the classic 'overlay').

[jameson_uk]

Answering the question you pose isn't straightforward since it depends a lot on what is going on upstream in the filter graph. The above MSDN references provide the technical detail on what they do.

I have been through the tech specs and understand it from that point of view but in terms of actual impact on playback they might as well talk about how they use the flux-capacitor

There must be some real world examples of when to use the different renderers and how they differ in terms of real world usage.

DXVA2 for example sounds very nice but is this in practice any better than DXVA1 ?