As alluded in a previous post, the current versions of Adobe Reader have a text to speech feature that takes any paper that they recognize as text, and read it. So, what I often do is to output the file as PDF, and use this feature to have the computer read my paper back to me. This helps me tremendously as I often find this way far easier to pick up grammar mistakes and find sentences and paragraphs that either sounds long and boring or makes no sense at all.
If I am working with LaTeX, I would output the paper as a PDF. If I am working with Microsoft Words, I will print it out as a PDF. To do this, I use PDF Creator, which is a free open source program. The program takes anything you print to it and output it as a PDF. Many similar programs exist, and the two that comes immediately to mind are CutePDF and PrimoPDF.
The controls for using the text-to-speech feature are located under the "View" menu, and the "Read Out Loud" sub menu. Here is a screen shot of the options for Adobe Reader 8:
The controls are rather crude. After opening the PDF in the Adobe Reader , you first activate the feature by selecting "Activate Read Out Loud." Once activated, Adobe could read portions of the text you click with your mouse. When I click on a paragraph, it often reads the word I click instead of the paragraph. As result, I can never quite control what it reads, and therefore use the "Read This Page Only" option, and go have Acrobat Reader read one page at a time. The issue might have to do with the way the PDF is generated, because I get the read a paragraph behavior sometimes when I use PrimoPDF. As I listen, I would note the sentences that need changes in my draft, and go back and look over the trouble area after the page is read. Alternatively, you can pause the reading and edit your paper mistake by mistake.