The V-Mode-Setting defines, which SMS outputs overall (so all the time, it is running, also when a video is played).
Technically the video-content is not upscaled [atleast, that's what I think, but hw should I/we know if the source is not available anymore...] (and thus the native resolution of the video doesn't change), but the output-resolution is higher and defined by the GS/'Graphic Synthesiser' (PS2 GPU) and this higher resolution could have an effect to the picture-quality.
I guess, it is better visible during movement since the pixel-raster is physically smaller in size.
For example if the resolution is doubled on x&y, the pixel-raster is only a quarter 'per pixel'.
I'm not sure how to explain it in laymen-temrs.
PS2 V7/DMS3 V2 (FW:2.4b7); Seagate Baracuda 200GB
PS2 V7/CC1.0 (FW:34 hacked v2 BM:2.1.6); Maxtor DiamondMAX9 PLUS 160GB
PS2 SCPH-30004R; NoMod+NoLaser
3xSony MC 8MB
MAX/Datel 16MB with Boot-CD
Custom FMCB 1.8b+ Beta-Build, my AIO 0.5, Sony&xRhino-Linux