There is no comparison. BD looks a hell of a lot better.
Play a DVD in a DVD player. Play that same DVD in a Blu-ray player. When I do this at my house it looks much better when played on the Blu-ray player
A DVD has highly compressed 480i, a BD has less compressed 1080p (with very few 1080i). Now if you have a digital TV or projector (plasma, LCD, SXRD, HD-ILA...) you have a fixed resolution and none are 480. So what ever happens at some point it is progressive and not 480 rows. Now cheap DVD player will not up convert and most TVs have real mediocre up converters. Now More expensive up converters will have better algorithms. So what could be happening is that your BD player is better at up converting then what you where using with your DVD player. But either way it has nothing to do with BD (except possibly that due to BD a more powerful chip was needed and so they can afford to make a better up converter)
As to the original question and topic this is what happens. The world is not discrete when it becomes digital (scanning of film or filming with a digital camera) it is divided into little squares and the whole area of the square has the exact same value. Now let's say you had a triangle's, you can't get a diagonal line so you could have something that looks like
when digitized.
Ten you need to compress it, so it becomes even more messy.
Now if your resolution and that on the disk are exactly the same, no processing is needed, you just show it as is. But lets take a simple example, let's say that your resolution is 4x (i.e. 2x height and 2x length- for example going from 480 to 960) then one choice to represent that image would be to have every pixel on the disk be 4 pixels (2x2) on the display. On the other hand if we look at the triangles above, that would mean that they will look the same except that the black lines consisting of a simple black box diagonal will now be black boxes that are 2x2. An other thing (and more correct in this case) is to make each 2x2 black box into two triangles (i.e. R1=black, blue, R2=green, black). The issue is that after compression a lot of information is lost, not to mention . What most up scaling algorithms do is analyse the surrounding pixels and then extrapolate what this pixel should represent. In essence (because of over compression and lack of knowledge) sharpening the pic (to try and determine the shpes) and then blurring it (to get rid of EE caused by over sharpening and give a smoother look).
Anyone that says up scaling looks as good needs his eyes examined because up-scalling cannot add missing detail.
