[Simondm]

Reasons why demagnetization is quite hard:
1. The tape is high coercivity (hard to magnetize/demagnetize).
2. It's a digital recording. The system doesn't use bias to work in the linear part of the hysteresis curve, it saturates the medium.
3. I'm not completely familiar with the DV format (although i have an ancient semi-pro DV camera), but it probably uses a PRML detection system. Advanced versions of these are capable of retrieving data when the s/n ratio is <1, i.e. more noise than signal.
4. It might also use trellis coding of the symbols sent to tape, such that the error correction is good (comes from modem technology but was adapted for tape in the early 1990s).

Put that lot together and add-in that it's looking for magnetic transitions, rather than absolute magnetic levels, and it's extremely hard to erase.

You might contrive something mechanical involving a powerful permanent magnet, but you'd need to expose the tape to both poles separately and to saturation, and then reduce the field to zero progressively. I don't see that as practical.

And before anyone mentions cooking the tape, the Curie point of those magnetic media is above the melting point of the substrate, so that won't work, either.

My background was in DAT tape for computer purposes (DDS), but there are many similarities and DV format was developed around the same time.