hi all,

I was wondering what kind of algorithm use the "regular" standalone players (DVD / SD Divx-Xvid) use for resizing the video?
Is it a high quality algorithm? I have two players based on mediatek MT1389xx chip if that matters. I live in a PAL region but I don't think this really matters in this thread's context.
When transcoding from DVD to xvid, due to cropping you end with lower dimensions. We can either encode at this lower SAR such as 688x something or 704x or even 640, or we can upscale during encoding to 720p, with two possible benefits:
1. reduce the "load" on the standalone during play
2. apply a slow and "good" (as in advanced, high quality) resample algorith ONCE, while encoding, instead of the (a assume) fast and lower quality algorithm used by the standalone. I am aware that the resample algorith could be done also by a hardware chip, but somehow I fail to convince myself (or find information on the web) if this is really the case and its quality is better than slow and "good" algorithms such as lets say Spline36 or Lancrosz4.

In my oppinion, I think it's best to upscale once, while encoding and not let the stand-alone player do it.
Also, when I hear some advices such as resize your 688-width crop to 640 width... I can't stop wondering how good or bad such an advice can be...
I mean, any kind of rescaling is a lossy process. Why DOWNscaling to 640 if the player will have to upscale to display 720? This is twice worse than leaving the crop as is (such as 688).
So, what's your take?

Tank you in advance for the feedback.