I'm afraid nowonder may have tracked wrong from the git-go . . .
"Upconversion"--as the "up" suggests--is the scaling up in resolution of a video signal. It is not the same as transcoding, which is conversion of signal between formats. For example, one can transcode from mpg to avi formats; this changes the format of the data, and may impact quality one way or the other.
But back to upconversion . . . this is again bumping up the resolution of a signal--say from 480i to 1080i--it is an increase in resolution. The process results in a net gain of data; if you compare to still images, it's like upping a picture from 5" x 7" at 72 dpi to 5" x 7" at 300dpi, which is net going from 181,440 pixels to 3,150,000 pixels. Bottom line, it's a gain of data.
So, a component that is capable of upconverting is capable of actually scaling up a signal from one resolution to a higher resolution, period.
Now here's the rub . . . the question is, how well does a given component do the job of upconverting? You see, like with my photo analogy, you're starting with a relatively small amount of data and converting to an image with a relatively large amount of data. If the equipment doing this is using a smart, complex algorithm, then the results can be rather pleasing . . . a standard TV signal can be made to look pretty decent on a nice, big HD screen. OTOH, if the scaling is more rudimentary, the upscaled pic is going to look like an overscaled photo--blocky and unclear. Obviously, the further you try to reach with scaling the harder it is also to maintain a good image; for example, mapping an image from a GameBoy game up to fill a 65" screen is asking a lot of any piece of on-the-fly consumer equipment.
This brings us around to the question of the inherent value of upconversion built into up-and-coming mid-fi receivers. At present, decent standalone scalers cost thousands of dollars . . . so are they overpriced, or are $1500+ receivers more likely to have less-than-spectacular upscaling capabilities? As of yet, we don't know . . . we'll have to wait for the new Denons & Yamahas and the like to hit the market and for capable folks to conduct some meaningful comparative evaluations.
Personally, I've drifted away from all the receiver hype and decided to go the computer route for upscaling/upconversion. There exists tons more documentation and user control of upscaling functions with an HTPC than with a receiver, especially considering the poor quality of documentation that many of the audio equipment manufacturers provide (just finished reading over user gripes about Denon's anemic and frustrating manual for the 3806 over on AVSforum . . . folks are having a time getting their receivers to "run right" because the manual is just this side of useless). HTPC users have been successful in 1:1 pixel mapping signals from an HTPC to HDTV over HDMI (the new HP DLP's are currently the only RPTV's that support this) with amazing results . . . I'd rather invest in a new computer than count on a receiver that I'm not going to be able to perform software/firmware upgrades on . . . but that's me.
Bringing it home, components that are capable of upconversion are different from those capable of simply translating signals from disparate sources and feeding them through HDMI to your viewing device. This is what nowonder was talking about . . . this would be receivers like the Denon 3806 that can feed all your signals from disparate sources to your TV via HDMI, but do nothing to up the signal resolution. These devices can pass a standard TV signal to your super-size screen, but there's no increase in the pixel count. Thus, either you're gonna get a little image with a huge black mask, or (more likely) the existing pixels blown up/stretched to fit the screen dimensions . . . not at all pretty.
Hope this helps . . .