Writing product requirements documents is an art with a dabble of science. Lately I've seen two extremes: If we had all these features we could sell....... and if we don't have actual market research we shouldn't spec it. Innovation requires the need to break out from these trains of thought. Let's face it, sales people cannot see beyond the current quarter, and analytical marketing people cannot see the truck until it hits them.
The first thing I do is examine the market research very closely. Then I talk to people who are the thought leaders in the market area and others in the channel arena (which is where I found out about recycle laws in Germany and packaged everything in cornstarch plastic and cardboard). And yes I talk to the sales group.
You also need to figure out what comes next. Because this product should provide a foundation to lead to the next product and the next market segment. Then I like to take a walk on the technical side to minimize the amount of absolutely pure invention needed to finish the product.
But I do throw in everything including the kitchen sink, if I think it will make the best product. Knowing full well that the product definition will change. But engineering always suprises me by what they can do. If I can convey the big vision, many times you do get big results. As long as you are willing to kill off features / functions which are not essential for success.