I've done another slide with that W34/70 tonight, just about to edit the pictures but I might leave the counting until another day. From eyeballing them though there are a lot more dead cells than the first lot I did, I'd estimate viability around 70-75%, but the packet has obviously been sitting unsealed so it's not a fair comparison. It's still in the fridge but rolled up in a rubber band. In any case, the culture is nowhere near completely dead.
I do have a packet each of Nottingham and Windsor yeast that are unopened in the fridge as well and are possibly even older, so I'll do a slide of those next week too.
I think the yeast calculators are based on methylene blue staining, which apparently isn't all that accurate once the viability drops below about 80-85%. I read a Braukaiser post on it where he'd stained a culture with methylene blue that showed 100% stained cells and it still grew when wort was fed to it. Obviously not 100% dead.
From my reading so far, viable or live cells flat out reject trypan blue from passing through the cell membrane, and so remain unstained, whereas the dead ones obviously can't do this and become stained. Methylene blue works differently; it enters all cells, but the blue dye is reduced by an enzyme in the cell to appear clear. The problem is that dead cells can sometimes still have enough of this enzyme left to reduce the dye, and compromised cells may not be able to reduce it, and that's where the inaccuracies occur in regard to viability percentage. When I get a culture that is shown to be largely dead with trypan blue staining, I'll test it with methylene blue as well to compare the results. I'll be interested to find out if there is a difference there.