Top
Best
New

Posted by emctech 12 hours ago

AliExpress runs silent WebAudio fingerprinting that breaks Bluetooth multipoint(blog.laserphile.com)
785 points | 269 commentspage 3
fuzzy2 4 hours ago|
When I visit an article on a popular German tech news website, it interrupts music playback on my iPad (the website takes audio focus). I bet they do something similar.
ninalanyon 4 hours ago||
Why are web pages allowed by default to do such things? Browsers should give the user the ability to forbid all sorts of thing and have them forbidden by default.
admax88qqq 6 hours ago||
Somebody (Mozilla?) should make a browser that just proactively blocks shit like this

I’m sure some Adblock addon could do it but at the browser level would be preferred. A browser vendor that just proactively does security and “correctness” tweaks to live sites would actually be in my interests as a user

nkjoep 11 hours ago||
JS enabled by default seems every day less secure.
emctech 10 hours ago||
So many website break completely with JS disabled and you end up having to enable it half the time anyway.
ruuda 9 hours ago|||
About half of the time, when a website doesn't work with js disabled, I realize that I didn't want to see the page that badly anyway, and I close the tab.
masfuerte 10 hours ago|||
It was pretty good until about six months ago. Since then loads of sites have added a js requirement to try to stop the AI bots.
MisterTea 10 hours ago||
IMO web browser have been enabling all sorts of obnoxious behavior since before JS. One of my all time favorites were the sites that opened pop-ups in a loop faster than you could close them while an audio clip of a guy yelling "Hey everyone! I'm looking at gay porn!" You had to hit reset. Fuck the Web.
afandian 10 hours ago|||
The web around the late 90s and early 2000s had some really sketchy stuff. I think the difference is that it used to be the sleazy underbelly. Now it's accepted as mainstream.

My local 'newspaper' website is chock full of scam adverts. The print version is dignified. The website people, somehow, turn a blind eye.

And I got an advert on Youtube this week using sexually explicit language to sell pills.

Feels like standards, and expectations, have really slipped.

grishka 8 hours ago|||
Opening a popup needs JS though.
nozzlegear 6 hours ago||
Not with the new popover API.
grishka 4 hours ago||
This doesn't look like it gives websites something they couldn't do before with CSS and JS.
nozzlegear 3 hours ago||
I'm pretty sure it does but maybe I'm wrong, I had never tried to create popovers/popups using CSS. What's neat[†] about the popovers API is that it doesn't need any JS, you just write a couple of HTML attributes and voila, a popover/popup. They will run with JS disabled as well.

https://chrisburnell.com/html-can-do-that/

[†] Neat for me building things like tooltips and dropdowns into my web apps without using JS; not neat for annoying people with popups.

barrystaes 7 hours ago||
Aha this would explain. I have seen similar behaviour with a news website trying DRM requests (has no reason to ask this info) resulting in stopping playback.. did not consider the impact of multipoint here. Interesting, might be worth looking into if i see this "bug" again.
ibaikov 9 hours ago||
I had this (?) happen. I have a soundbar hooked up through spdif in my pc. It automatically switches sources, so I can play music through airplay and then have it play sounds from pc when I open youtube etc. So it switches from airplay music to pc even when nothing is playing on pc. This was happening on some websites and it is extremely annoying.
grishka 8 hours ago||
Is there any particular reason these kinds of APIs are not behind permission prompts?
sillyboi 8 hours ago||
I thought the App Store review guidelines explicitly prohibit hidden features and using public APIs outside their intended purpose. Is audio-based fingerprinting just not something review can realistically catch?
ValdikSS 8 hours ago||
They probe all audio devices, including microphones, which probably temporarily switches Bluetooth devices into HPF mode due to how Bluetooth duplex audio works.

I'd argue it's "silent" though: aliexpress wakes up my audio card if nothing plays, which results in a very faint "pop" sound every time I open the tab.

It's been this way for ~3+ years at least.

lapcat 8 hours ago||
It's the website.

The title of the article literally mentions "WebAudio", and the first paragraph states that the author is using a PC. The second paragraph mentions Chrome and Firefox.

Apple and the App Store have zero involvement here.

hunter2_ 7 hours ago||
This is a huge stretch, but if this problem exists in not only the PC versions of Chrome/Firefox but also the Android/iOS versions, then theoretically the app store reviewers could flag the browsers for facilitating this behavior against app store guidelines. In practice, apps of such caliber as popular browsers might be a bit above such reviewers' pay grade, so to speak.
lapcat 7 hours ago||
> This is a huge stretch

This is nonsense. Safari also supports Web Audio. Safari does not, however, support Microsoft Windows, which is why the article author didn't mention it.

Moreover, all web browsers on iOS have to use Apple WebKit, so Web Audio support is not actually the fault of the non-Apple browser vendors.

ngl999 11 hours ago||
Just curious, why silent sound would allow fingerprinting? What are they sampling if it can't be heard?
emctech 11 hours ago|
The script generates a known waveform, it is passed through the browser's audio implementation and then the script analyses the result after. Based on your devices settings and hardware the output will be different, e.g. a PC with analog output might have 44KHz audio output bandwidth, but a bluetooth headset might have a lower, different audio bandwidth. That is a datapoint that can be used in device fingerprinting alongside screen and viewport dimensions, device pixel ratio, browser plugins, etc.
hunter2_ 7 hours ago||
On the one hand, I wouldn't expect too many variations here (the vast majority of devices probably use 48 kHz and 24-bit output, a few use 44.1 kHz and/or 16-bit, etc.) but just like DPR and all the other properties with a very small set of popular values in practice, you only need a bit or two from each measurement to eventually have a high quality fingerprint.
dzonga 2 hours ago|
I think x.com does this too - haven't been able to dig deeper.
More comments...