Originally Posted by frankcastle1975
Little overwhelmed and before I go studying these things I just want to wrap my head around what they do and is it for me.
xmbc from my understand is a way to display and organize your stored media in an attractive way..... do you need software to play media or does xmbc take care of that as well?
If you have more than one pc sharing files from one central source do you need xmbc installed on the server and each front end pc?
madvr..... googling what it does I get a bunch of mumbo jumbo on rendering...... is this just for people creating high end graphics or do they mean rendering from the viewer at home as well? what this thing does is completely unclear to me. what does it do (in layman's terms) and why would someone want to do it?
interframe creates 60 fps files I get that I think.... I mean if I have dvds and plan on storing them digitally I might want to create new files at 60 fps and watch them later
svp does interframe in real time
is it correct to say that going from 24 to 60 fps is upsampling? I sometimes see this term being used but I thought it meant improving the resolution of a video..... are there two meanings? Or is that how upsamping is achieved.
I'm sorry these are probably basic questions but I'm finally nearing the htpc I want to build (see below).... although might want to make changes if going to madvr or svp or interframe are things that are of interest to me.....
XBMC is the all in one application, download install and run. It needs to be installed on every PC that is connected to a TV or Display.
You are correct about Interframe and SVP, both use the underlying AviSynth engine.
A movie is typically encoded at 24fps, modern TVs refresh the screen at 60fps, so the work is fill in the missing frames by interpolating across the existing frames. to create the extra 36 frames in each second.
This is a lot of work and SVP used the GPU to help with the work.
The nice thing about SVP is that I do on my computer what used to take a broacast frame rate converter like a Teranex, these things cost $50K or more.
I use realtime conversion that is SVP, if you do the conversion with Interframe, the output has to be re-encoded and it must be a lossy encoding like H264 to save space.
The result is a 60fps version that is not as sharp as the original.
MadVR's job is to scale to the display resolution and do the color space conversion.
Digital video is not usually sored in RGB values like PC images instead it is broken up into 1 Intensity component (Y) and 2 color subcomponents (U and V)
There are formulas that convert between YUV and RGB, as there are several standards BT601 for Std Def and BT709 for HD
The math is here if you are interested https://en.wikipedia.org/wiki/YUV#BT.709_and_BT.601
This is quite a compute intensive exercise and MadVR uses the GPU to off load the work converting YUV to RGB for the PC to display.
The other thing MadVR does is scaling.
On codecs like MPEG2 or AVC/H264 the Intensity component and color components are encoded and stored separately.
Moreover the resolution of the color information is stored at a lower the resolution than Intensity. (4:2:0 Y'CbCr encoding)
MadVR restores this thru the YUV mapping described above. We are not done yet as this only bring us to the original resolution
MadVR then has to adjust this to the final size of the display such as 1920x1080
The built in DXVA subsystem in Windows does the same thing if you use the EVR renderer, MadVR does the job with higher precision.