Comment by jmole

Sorry, I think you're reading more into this than I intended to say. My point was that the raw data itself doesn't need noise, but the published data necessarily does.