does recoverPose() return up-to-scale or correct translation?

edit

Refer to the literature, for instance: Multiple view geometry in computer vision.

9.6.2 Extraction of cameras from the essential matrix

The essential matrix may be computed directly from (9.11) using normalized image coordinates, or else computed from the fundamental matrix using (9.12). (Methods of computing the fundamental matrix are deferred to chapter 11). Once the essential matrix is known, the camera matrices may be retrieved from E as will be described next. In contrast with the fundamental matrix case, where there is a projective ambiguity, the camera matrices may be retrieved from the essential matrix up to scale and a four-fold ambiguity. That is there are four possible solutions, except for overall scale, which cannot be determined.

Also, to estimate correctly the Essential matrix, the points should be in general configuration (planar points are degenerate case, see Two-View Geometry (Course 23, Lecture D), Jana Kosecka).