I'm trying to better understand singular homology groups. Standard examples like spheres and tori make it seem like $H_i(X)$ is generated by a subset that is in bijection with $i$-dimensional holes in $X$. Even more, they make it seem like those groups are freely generated by such subsets, but I guess it would be naive to think THAT might hold generally.
Could someone throw some light on the way these groups relate to holes in a topological space? Also a short comment on what changes when we use a ring of coefficients different than $\mathbb{Z}$ would be appreciated.