For the purposes of calculation of magnification ratio, it doesn't really matter whether you compare the size of the image on film to the size of the object in the world being photographed, or the width of the image field (the width of the film frame) to the width of the subject field (the whole area that the lens takes in at the subject distance). These ratios should be identical.

Methods like the "QuikDisk" available for free online and Calumet's device use the former approach, where an object is put into the scene, and a ruler marked in stops is used to measure the image of the object (the QuikDisk or Calumet's square card) on the groundglass, but this is only really practical with large format, where you have easy access to the groundglass. With a 35mm or medium format camera, I find it easier just to use the width of the frame as a known quantity for this purpose and to measure the subject field with a ruler or tape measure.