by mirkosp » Sat Mar 10, 2012 8:33 pm
No, that's exactly the point. If your source is YV12, converttoyv12 is useless, if it isn't, then it has to convert to yv12 which means it'll have to harshly reduce the chroma resolution, which can cause even sensible quality drops (though it's shouldn't be looking as brad as you seem to imply by saying shit, heh)?
In order to keep quality to its best you should avoid colourspace conversion by all means, which means that, as your original source is most likely yv12, you should keep it that way from the preprocessing, through the editing, to the export and the final encode in order to ensure the quality stays the same. Converting back and forth from yv12 to other colourspaces is gonna cause visible colours alterations due to how the subsampling has to work.
Addborders does not do anything to the quality as it simply adds black bars around the video, so either the issue stems from a bunch of colourspace conversions, or you're doing something wrong when encoding again in vdub.