There's no way to tell the difference from a paused frame, a still mage, or a frame as seen through a media player instead of a direct frame copy. You appear to have disturbingly soft video playback, if all you consider are the posted images. Why are you capturing at 640x480 (or has this image been resized by your video player, which would be another reason your posted test is invalid).
Next time, create your images directly from the video using something like
VirtualDub's ability to copy unaltered frames directly from the original video. And when you're checking for noise, use moving video rather than stills. I still see a lot of bad noise remnants in both images and detail lost and distorted through jpg compression. Use lossless png next time. IS this tape a 6-hour extended play recording played thrugh a JVC with JVC noise reduction engaged? The detail is really poor. Leaving the Panansonic NR on causes motion smearing and posterization effects, but your images are so processed and multi-generational that it's difficult to detect differences.
By the way, you should post images directly to the forum. As soon as your images are removed from the offsite location, your forum post and your question will be meaningless.