it is probably the pixel aspect ratio…. the dv footage you capture from your camcorder is nonsquare (.9 pixel aspect raition). for some reason ae is not interpreting it as such, but you can tell ae the correct pixel aspect by selecting the footage in the project window and right-clicking (or file menu)>interpret footage>main… here you can set the correct interpretation for pixel aspect (and other settings like field dominace–should be lower for ntsc dv). then when you put these into comps, ae will make the aspect correction for you, based on the comp’s pixel aspect.
you can do the same for images. you can also create or convert nonsquare pixel images in photoshop cs2. when you create a new file in photoshop, the file parameter setting window that first opens will have a preset for dv and d1 standard (and others) that are .9 pixel aspect. ps will then handle the conversion of images that you drag into the file (say a square pixel jpeg from a digital camera).
note that if you work with 720×540 sqr pixel images in ps (which is a very common and often preffered sive to work in for video), when you put those in a d1 or dv standard comp in ae, you will also scale the image down to 90% (or right-click (or layer menu)>fit to comp). that is normal.
Kevin Camp
Designer – KCPQ, KMYQ & KRCW