Another different reconstruction of Betelgeuse on January 9, 2014. This time the inversion with ℓ_max = 3, 4, and 5 are shown from left to right, but the reconstructions have been made using only the spherical harmonics coefficients up to ℓ = 3. This means that the left figure with ℓ_max = 3 is final for that inversion, but in the centre and right images the high-order coefficients are missing from the reconstruction. The images ℓ_max = 4 and 5 show only the biggest granule, the fine details being missed with the high-order coefficients. The ℓ_max = 3 case however tries to reproduce the profiles by adding large granules on the edges, thus making them small by projection. This shows the necessity for ℓ_max > 3 to correctly invert the observed profiles.