I'm an engineering student at HMC (and a life-long rocketeer), so take what I say with a grain of salt:
Flutter (especially in transonic/transitional zones) exists. It is strong enough to destroy just about anything if in a shockwave. I think everyone agrees with this.
However, in this video, the fin flutter you see is an aliased version of mechanical resonance ocurring at 28,32, 58, or 62 (maybe higher) Hz if the video is recorded with NTSC standard. Similarly, if PAL was used, the flutter would be 22,26,46,50 etc.
I'm confident that what you see is NOT what you get for various reasons:
1) Last semester, I build a rocket with some professors with embedded dynamics strain gauges in the fins and airframe of a rocket (with G10 fins) and the resonance modes shifted between 50-60Hz in flight. The data can be found at
https://www.eng.hmc.edu/NewE80/ResultsOverall.htm .
2) I had three professors of mine (one PhD MIT in chemical and materials engineering, second on team that designed SSME and X43 engine, third HMC->Caltech physics) take a good look at that video. They concluded that steady-state or transient resonance as expressed in the video seems unlikely - that is, they all think we are seeing aliased data.
3) The Nyquist-frequency of a standard video device is 14.985Hz (NTSC is 29.97Hz). If you have tapped and audibly been able to hear your fins make a tone, it is likely you are above the Nyquist-frequency and therefore you will experience aliasing/folding in a standard video.
Anyways, what I (and my professors) find very interesting is the distortion that occurs in the transonic region. This may be a digital glitch but it is likely that the compression wave from the boundary layer changed the local index of refraction of the air, (and by the Archimedes Principle) creates an optical illusion.
We'll be launching a rocket, Mudd II on May 12 that will take very careful dynamic measurements of a carbon-fiber rocket's resonance (with 2 video camera to look for flutter). After a quick test flight on an H (rocket weights about 2 lbs), we're planning a J415. In the summer (at LDRS), with a higher-performance fincan (and CFD/nonlinear stress analysis) we'll push it to the breaking point with a full K/baby L... giving us Mach 2.5 (and with a custom 600MHz blackfin flight computer) sampling on 12 sensors at 30kHz.
Given the 200kpsi tensile strength of the airframe/fins, I'm not quite sure an L will destroy it...we only have 3ft x 54mm for motor space....