I have used a plain door phone that wasn't designed for SIP but for a standard phone line in conjunction with a SPA2012. It prompts for the person's name, records it, then makes a doorbell sound along with the person's name through my whole-house audio system.
Quick search reveals a few straight-SIP options:
Algo,
Valcom,
2N (the one I wish I had),
ProVu
Though not all of them have video capabilities.
My doorphone specifically is a
Viking E-50E and it has a camera, though it is not network enabled. Uses two pairs of CAT5e wire - one for 12V power, one for phone tip/ring. Sends video signal back via coax to a network video encoder.
This all worked just fine until it got pressure-washed...with bleach. I haven't bought a new one yet, but before it got washed, it worked just fine. I wasn't all too happy with the speaker volume, but the mic picked up decently well. There's probably a way, especially with the new favorite Yealink T-46G to pull a JPG link from any network-enabled camera.
Ward mentioned at some point using a Grandstream GXP2200 with a separate Grandstream network camera. It was either here on the forums or on NerdVittles. Maybe a search is in order?
I have also thought about taking apart the Nortel 1535 and trying to fit it into a doorphone-sized box, but I just haven't had the time.