Re: Online tracking: A 1-million-site measurement and analysis

On 26 May 2016 at 09:04, Joseph Lorenzo Hall <joe@cdt.org> wrote:
> a fascinating use of AudioContext to
> fingerprint the device without doing any actual audio output (one
> method does an FTT and sets gain to zero (mute) before playing, the
> other copies to a OfflineAudioContext buffer).

The problem here is that it is highly correlated with other
information (like the UA string) since it's based on implementation
variation and not machine-based.  It will be easily defeated by the
planned addition of dithering.

Received on Friday, 27 May 2016 07:02:41 UTC