Darrieus designs are more efficient, but suffer from insanely high cyclic loads that tend to destroy them after a short time. I don't know of any but maybe some Gorlov helical designs that have survived for long enough to provide decent power. (I know the catenary "eggbeater"s, the only darrieus that was tried in California's Livermore area, were shut down and torn down years ago, leaving nothing but a plethora of three-blade horizontal-axis designs in commercial service.)
Savonius based designs are less efficient, but seem to me somewhat underrated, for a couple of reasons.
One is that there's a chart circulating - of efficiency of different turbine designs at various tip speed ratios - that had the labels of the savonius and the american multiblade ("patent") windmills swapped. The basic two-hemicylinder savonius is more efficient than the patent, and the patent windmill is eminently practical and has pumped water - and sometimes generated electricity - across the great US great plains for 170 years or so.
Another is that the Savonius presents a rectangle, not a circle, to the wind. So its swept area is greater for a given diameter. I once worked out that, using the Sandia Savonius rotor's claimed efficiency, a rotor only 12% taller than its diameter, would collect the same energy as a good horizontal axis machine of the same diameter.
The Ugrinsky rotor is claimed to have an efficiency about as good as a good horizontal axis machine - which would be the ideal, as they are approaching the Betz limit. But I have yet to see any serious analysis of the design and (as our member Adriaan Kragten points out) wind tunnel tests can easily overrate horizontal axis designs. They tend to restrict the spreading of the airflow, forcing more through the rotor than they'd experience in free air.
An upside to Savonius designs is that they're very sturdy and, at their low speed, don't tend to fly apart (or chop up birds). A corresponding downside is that they're HEAVY. On the other hand, they're simple and can be made with easily available materials. (Auto/trailer wheel bearings can support a LOT of spinning weight.)
For converting the slow rotation to a fast spin of a generator shaft, I've always considered v-belts to be a viable option. They're easy to obtain in many sizes, and only lose a few percent of the energy if they're kept properly adjusted. You may be able to find a giant pulley on a salvaged large-buiding blower.