No, rewards are not reduced by a reduction in votes (the number of votes will already are reduced by the fee), the distribution includes the whole pool and whatever was voted on gets that part of the pool that is momentarily available to assign to rewards.
Eliminating curation share means 100% of the vote-assigned rewards go to the authors. What I am proposing is to replace it with non-voting stake type that pays liquid rewards (probably weekly) based on the share of stake that would have gone to the 50% split author/curation, and that leaves the other 50% to be added to the rewards pool.
Author rewards go way up, voting is only for philanthropic purposes and participation in deciding what is the best content at any given time, and whales can focus on their fancy hobbies instead of managing their voting to maximise their reward for staking, and not on manipulating the feed, chasing the best curation rewards, and still make the same profit.
I don't see what downside there is to this, though it's a fairly significant change in the protocol.
your proposal is interesting and i see upside where it appears to encourage concious voting, while (as you explain) still maintaining staking and posting well rewarded. up-voting may not only be a philantropic act rather -indirectly- serving the blurt-community with conciousness (whether artificial or human - that is for another discussion?). this in turn by association serves all of humanity, returning these intangible benefits to the up-voter (curator) as well. nice feedback loop u have going there! to be up-blurted! Downsides anyone?