11-cis-retinal can be found as a part of rhodopsin and it is the non-excited form of retinal. By absorbing light, retinal goes to a higher energy state and converts to 11-trans-retinal. At least I understand it so. If all above is correct, I just don't get one thing. Shouldn't the cis form be less stable than trans (due to steric hindrance)?
Cis-retinal is less stable than the trans isomer. The carotenoids ingested by the body are all trans. Enzymes in the body convert them to trans-retinoic acid, then turn them into a fatty ester which is then converted by the enzyme retinoid isomerohydrolase to cis-retinal. The higher energy state of the cis isomer makes it prone to photoisomerization back to the lower energy trans isomer.
The answer below is in very simplified terms and does not take the enzyme into account.
The photoexcitation will effectively move an electron from a π to a π* orbital. Thereby, a π bond is effectively broken and the molecule can rotate around what is now effectively only a σ bond. Thus, this rotation will lead to a thermodynamic equilibrium between cis and trans until the electron relaxes back to the ground state.
We do indeed expect the trans form to be more stable; therefore it is more likely for a cis-retinal molecule to be converted to trans-retinal than vice-versa. It is entirely possible that nature found this out as well — through random evolution — and selected the process that would yield a slightly higher probability of a photon inducing isomerisation. This process would lead to a slightly higher light sensitivity or the ability to better see in the dark.