Security quotes of the week

Posted Mar 14, 2017 6:49 UTC (Tue) by jmspeex (subscriber, #51639)
In reply to: Security quotes of the week by andrey.turkin
Parent article: Security quotes of the week

Yeah pretty much. A lot of these "attacks" like turning a speaker into a mic make a nice paper, but are often not practical. In this case, the amplifier is indeed the thing that would kill the idea in practice. At least until the power amps are on the SoCs, which I don't think is the case right now if only because of thermal issues.

Security quotes of the week

Posted Mar 14, 2017 14:40 UTC (Tue) by farnz (subscriber, #17727) [Link] (2 responses)

Choosing a random reconfigurable codec shows that you have a 2.3W per channel amp embedded in the codec. So, you've got "jack detection" (which means that it can reconfigure pins to select whether they're mic in, line in, line out, headphone out or speaker out pins), plus a 2.3W per channel amplifier for the speaker out.

That type of chip can easily be reconfigured to use a speaker as a microphone from software - it won't be as good as a dedicated digital microphone, but it'll be as good as any cheap analog microphone.

Security quotes of the week

Posted Mar 15, 2017 11:10 UTC (Wed) by cladisch (✭ supporter ✭, #50193) [Link] (1 responses)

Such HDA codecs are designed specifically for PCs; the corresponding HDA controller can be found only in PC chipsets.

And while it could be imagined that someone builds a TV on top of PC hardware, it would not be possible to rely on the built-in amp (2.3 W is "for mini-speakers").

Security quotes of the week

Posted Mar 15, 2017 15:23 UTC (Wed) by farnz (subscriber, #17727) [Link]

I've found equivalent reconfigurable I2S codecs in the past, inside smart TVs, driving the speakers directly. It's just that even getting a feature list for the I2S codec version is next to impossible if you haven't already already signed an NDA with the codec maker, whereas the HDA codec makers will give out a feature list but not a datasheet without NDA.