There's no shipping or officially-announced product that I'm aware of to do it, but I'll be absolutely SHOCKED if SOMEBODY doesn't make a device that can take 480i video as input, and modulate it to something a TV with ATSC-compatible tuner can handle within the next 3-5 years. If you're planning for the future, I'll predict that a device that uses QAM will probably hit the market at least a year or two before a device that uses 8VSB, simply because QAM is easier and cheaper to implement a minimal subset that's "good enough" for in-home distribution (assuming you had to implement the whole thing from scratch with discrete components; the moment someone rolls out a nice ASIC to do most of the work behind 8VSB modulation, the whole equation changes).
Unlike NTSC/PAL modulators, an ATSC modulator will be neither simple nor cheap, because it needs to do three things:
* Decode the analog video and audio signals to bitstreams.
* Encode them to primitive MPEG-2 video and audio streams, optionally combine them with one or more additional sets of streams (if you want to have two or more subchannels, like 3.1, 3.2, etc), and bundle them into an ATSC-compliant ~19.2mbit/sec bitstream
* modulate them to channel 3, 4, or some other channel via QAM or 8VSB
Right now, an engineer with background in mixed-signal design could probably build a box capable of handling 1-4 480i video streams as a personal hobby project for about $1,200-1,800... and build subsequent ones for about $500. If a real company with the resources of Cisco or D-Link were behind it, the cost would probably fall to $199-299. Add Chinese companies, and whack another hundred off within a year, with ~$99 probably being the "island of stability" before the manufacturers lose interest in 480i and go for newer, more expensive models that can do 480p60 from component video, then maybe 1080i with 640 or 720 horizontal pixels, then 960. Beyond that point, the legal problems are harder to overcome than the technical ones thanks to the nice folks behind Macrovision and HDCP (whose viral licensing terms infect everything they touch, as well as everything made by someone infected by the Macrovision/HDCP legal virus).
The good news is that we ARE likely to see at least one solution that involves a capture+encode+transmit box at one end of an ethernet cable, and one or more decode+output-to-tv boxes at the other end, by Christmas. The hardware to get the video into MPEG-4 and shoot it down the ethernet already exists -- HAVA, Slingbox, and others. All that's missing is the box to plug into the network, grab the video stream, decode it back to video, and output it to an attached TV without involving a computer as an expensive faux cable box. HAVA boxes can already output straight MPEG-4 via RTSP that's compatible with open-source playback software, so it's only a matter of time before they (or someone with a background in embedded hardware design) makes a box to do it.