x264 largely requires the source video to be in the YV12 format. If you're using an Avisynth script and it is being processed by an x264 binary that was built with AVS support then that is also true. If you're using other sources (either FFMS2 or libavcodec support, assuming it was compile with such) it'll handle that internally.
Sometime between now and tomorrow I'll try encoding your source video through x264 (libavcodec), ffmpeg (using libx264), and mencoder (again using libx264). Odds are I'll try to get at least one lossless version, and I'll have them both scaled and unscaled (since colors are being "made up" to fill in the additional pixels and may make it look a bit off).
It is entirely possible that the slight color differences you're seeing may be due to chroma subsampling. The human eye is less sensitive to changes/movement of color than luminance, and a variety of methods exist to store image information with luminance data than color. This is normal, and most media you're exposed to (DVDs (MPEG2), Blu-ray (H.264), streaming HD video, JPEGs, etc.) do this.
EDIT:
Here is the first set.