In reply to Tall Clare:
Yes, woven fabrics do stretch slightly on the bias, which is why bias cut dresses hug the figure nicely. It's also why, twenty years ago, I considered modelling tent fabrics as a square grid mesh of four inelastic beams, jointed at the corners, and with springs across the diagonal corners. And why, when cutting panels for a tent, you can't simply try to achieve the best fit of panels from the roll, since you have to take into account the bias stretch on the way the tent behaves...
In a woven fabric, there's usually a bit more give in the weft direction, since the warp fibres are held tight in the loom, and the weft fibres are threaded through them, so end up having more slack. This give is used in cross grain designs.
But I think what I'd call four-way stretch is achieved by knitted, rather than woven fabrics. Since these are formed by loops of thread, rather than a rectilinear array, they can stretch in all directions more easily. That's why a lot of stretch soft shell fabrics are actually knits, rather than woven. And I suspect that coverings for foam seating that Nicola mentions is probably a knit, too.
As you know, I'm only an amateur at this stuff, and what little I know is down to observation, rather than formal study. So it's quite likely my understanding is wrong; looking up the MEC T3 hoody linked in a recent thread led me to a video by their clothing designer, and she seemed to use 'two-way stretch' to be in one direction, and 'four-way stretch' to mean two directions.