Does Skype use audio watermarking?

Question

Today on CNN they conducted an interview with a career criminal, who spoke by Skype on condition of anonymity. It occurred to me that, if Skype uses digital watermarking to embed identifying information in the audio stream, then this criminal has potentially blown his cover, at least to Skype and to any other entity with knowledge of the watermarking technique.

Skype is, of course, a closed-source program that is known to surreptitiously access third-party configuration files (for reasons that may or may not be nefarious), and it is claimed that law enforcement agencies have the ability to eavesdrop on Skype conversations. However, I don't know whether anyone has ever alleged or investigated claims that Skype uses imperceptible audio watermarking to covertly embed identifying information, such as the originating Skype account ID, MAC address, IP address, operating system details, etc.

Have there been any leaks or admissions by Skype confirming their use of such watermarking? If not, has anyone attempted to use technical methods to detect the presence and/or contents of such watermarking?

Answer 1

As far as I know, there haven't been leaks or admissions that Skype uses this kind of technology. However, other techniques can be used.

First, law enforcement agencies can ask Skype to give them certain data about a Skype account:

In response to a subpoena or other court order, Skype will provide:

• Registration information provided at time of account registration

• E-mail address

• IP address at the time of registration

• Financial transactions conducted with Skype in the past year, although details of the credit cards used are stored only by the billing provider used (for instance, Bibit, RBS or PayPal)

• Destination telephone numbers for any calls placed to the public switched telephone network (PSTN)

• All service and account information, including any billing address(es) provided, IP address (at each transaction), and complete transactional information

Note that billing addresses and data related to financial transactions can be more useful than IP and email addresses.

In addition to that, there is an interesting technique for forensic analysis of audio and video recordings, which doesn't rely on watermarking: electrical network frequency analysis.

The basic idea is that you can detect the fluctuations of the frequency of power distribution networks (usually, around 50 or 60 Hz) in audio and video recordings. Since the fluctuations are known by the power distribution company and vary according to the power grid, it is possible to geo-locate any speaker, potentially including a Skype user, at least in an approximate way. These are two sample papers on the topic.

Since Skype is a (closed-source) software, it is possible that it sends system information such as the OS and the processor being used by the user , but I haven't found any evidence of this so far.