Just as making a photocopy of a photocopy degrades the quality, so too does a copy of a microfilm copy. The prospect of having access to digital scans of NARA's U.S. census master microfilms is a delightful thought, indeed. If I have not spoken of it before, it is because I was afraid I would awaken myself from this dream! (If that doesn't sound like my typically dull prose, I must beg forgiveness. My daughter is watching a Jane Austen video.)
I've found the chief beneficiary of the first generation scans to be pages cursed with cellophane tape. And the best candidate for cellophane tape is the first page of each county. I thought to give you an example or two from the 1900 census but, alas, Ancestry.com's second generation scans are already gone. Instead, I give you an example from the 1920 census, Illinois, Adams County, Ursa Township, enumeration district 62, sheet 6A.
The left half of the image is the scan of the 1st generation microfilm made by FamilySearch. The right half of the image is the scan of the 2nd generation microfilm made by Ancestry.com. The left is superior in sense and situation.
This is not to say that there will be no glitches. There are bound to be particular images that end up being worse. If image quality is not up to snuff and FamilySearch indexers aren't confident enough to return pages for re-scanning, then they may sneak through into the finished product.
Here is an example from the 1870 census, Utah Territory, Cache County, Wellsville, page no. 1.
A light, blurry image from FamilySearch's pilot site
A clearer, albeit skewed image from Ancestry.com
In this case, it is not the FamilySearch image, but the Ancestry.com image that has the upper hand. I imagine this to leave some of you quite speechless. Yes, quite speechless, and I expect that you have not stopped talking of it since.