The problem with the argument is that it tries to distort the situation and ignores any useful discussion of the market value of the item in question.
The problem with market value, though, is that the market price of monopoly goods is not naturally regulated. Lots of people use the revenue of music to estimate its value, but monopoly goods are not naturally priced. Copyright is a government created artificial monopoly. It exists for a good reason -- to channel revenue into science and the useful arts -- but the sale price of monopoly goods does not, and cannot, accurately reflect the theoretical market price.
If there is a good way to estimate the value of music, that would be very useful. But it can't be revenue, so it would have to be something like: How much does "Me and Bobby McGee" (Janis Joplin version) make society better compared to "Steer" (Missy Higgins)? How do either of them compare to a table saw? I think those things are inherently difficult to measure, which is why I tend to focus on the resource streams going into production.
The resources going into music are highly mobile, with strong alternative demands, because they are mostly labor that starts at a young age when it can still be shifted into other fields with a low cost of transition. They are also very closely measured by the Department of Labor, so the data we have to work from is pretty solid.
Of course, my approach isn't the only good one -- and more measures are a good thing. Getting multiple estimates of the same market phenomena using different datasets is an excellent way to test for flaws in the measures.