I would suggest experimenting with a BPW34 photo-diode for a number of reasons. Firstly they are widely available and at a reasonable cost. Their response time is something like 20ns, circuit arrangement allowing rather than the lethargic response of solar cells or LDRs.
The thing to consider is whether they would fit into your optical arrangements. The sensitive area is 7.5sq.mm and it is square (approx. 2.8mm x 2.8mm). So if you can arrange for your projected image to fall within these bounds it may be worth considering.
In addition Chris Long has used them for his long-range optical communications, it's worth a look at his (and others) site, there are pre-amp circuits which are developed for extreme low-noise without requiring liquid Nitrogen...
http://www.modulatedlight.org/
These should be adaptable to NBTV requirements.
If you do consider using these make sure you get the clear-packaged version, the plain BPW34, not the black IR version, BPW34F or BPW34FA. There is also a blue-enhanced version, BPW34B or BPW34BS.
As for Steve's question r.e. signal levels, it depends on the configuration. Used with reverse bias a photo-diode produces a current output. The higher the load resistance the greater the voltage out....but watch stray capacitance. This gone into detail on Chris's site.
Datasheet for the visible version attached.
Steve A.