Why not use Red October HQ without any modifications?
madVR provides a number of minor picture quality improvements, that - when they accumulate - produce a significant net improvement.
Each minor improvement can be quantitatively increased in the settings - but almost always that uses more GPU. Even very recent GPUs do not have enough power to use all the madVR settings at maximum - because madshi is always coming up with another feature.
Certainly GPUs prior to 2013 have only enough power for certain features and not others.
Different types of video benefit from differing amounts of improvement, so to maximize the use of one's GPU, one needs to use the Profiles in madVR. The main way to do this is varying setting of "image upscaling" because that has the most different in Performance between some methods and others.
For example:
* On SD (usually from DVDs from pre-HD TV shows, like The Wire), "softcubic 80" produces excellent results in making artifacts less noticeable.
* On 59 or 60 fps, one wants to use a less GPU-intensive method like "Bilinear"
* On 1072 or 1076 line videos, one wants to use a less GPU-intensive method than normal because of the GPU power needed to scale from 1072 to 1080.
IMHO, the biggest benefit of madVR is chroma upscaling. Attention is always focusing on resolution, but I find that "three dimensionality" is actually relative to the number of colors used. Chroma upscaling thus makes video look more realistic. With my GPU, I use Bicubic 75 w/ anti-ringing.
( Recently, I happened to use VLC player to play a TV video to test something else, and was struck at how much less realistic it looked than when I had played the same video using MC20 and Red October HQ/madVR. )
