Hi Maiklas - and season's greetings to you too.
This is the gif that you linked to:
Let's first compare it to the phenomena observed in the Naudet video:
Here is a cropped view of the building at right:
And here is, most importantly, the same video-segment extracted directly from the DVD of the Naudet's movie "911" (which should be available in your local video store):
http://www.septclues.com/ANIMATED%20GIF ... quake1.wmv Please download it and watch it at leisure in your video player.
Yes, that short clip has of course been through a .wmv conversion, so you have only 2 options: 1.Take my word for the fact that it's not significantly different from the original DVD source material. 2. Go and buy the DVD (I think it's still available for purchase on the internet too), upload it to your video editor and check it out for yourself. For now, please download this clip which certainly has 'suffered' less downgrading than the .gif file above.
You may know that this phenomena seen in the Naudet video is by no means the only problem with their "AA11 first WTC hit" footage. But since your specific query has to do with this 'wobbling' effect, let's stick with this issue right now. So firstly, there is no reason to believe that the original video material has undergone the sort of noise reduction mentioned in the article you linked to. As far as we know, the Naudet DVD is the result of professional video tape converted into DVD format at top quality. Secondly, if - for some reason - the Naudet 's source material had some inherent, random defects, we would observe similar 'wobbling' effects elsewhere in their 129min movie - which is not the case (and there's no shortage of motion views with buildings in their film). Thirdly, and to get back to your linked article, the Naudet phenomena affects a wide range of areas in the frame (entire 'blocks' of buildings/color zones) and appears quite different from the example you linked to, where we only have a distinct, white pixel area (the white sliver of window) 'bobbing' in realtion to adjacent, much darker pixels.
In conclusion, the Naudet 'phenomena' cannot be explained - neither technically nor visually - by the noise reduction artifacts described in the article you linked to.