I think that by fixing common problems before launching, you get more meaningful feedback later from your users. They won't concentrate on the obvious problems that could have easily been fixed.

The benefit of launching earlier far outweighs the annoyance of trivial feedback. If someone is genuinely interested in the product they'll be willing to overlook critical show-stopping bugs that destroy data in order to help you get the product to a point where they can start paying to use it. A few missing alt tags won't make the slightest bit of difference to the user, and that's the person who matters.

But failing to launch of a month while you 'polish' things will make a difference. You'll lose out on feedback, potential customers, and give more foothold to any competition. Plus it's boring and expensive to not launch.

People who'll walk away because there's a spelling mistake on the contact page would never become paying users regardless of how great your product is. Ignore those people.