Rubric 1.0.0 · column 11 of 13
Audio sample rate
The sample rate an AudioContext reports, in hertz. Determined by the audio hardware and the OS mixer.
Under rubric 1.0.0 this column is worth between 0.47 and 6.64 bits, and the likeliest outcome the table knows — 48000 Hz — costs 0.47. Those figures are assumptions printed before your run. The counted box beside them is not: it is how many of the devices these sites have read share your value.
How a browser produces the audio sample rate
Constructing an AudioContext makes the browser negotiate an output rate with the operating system mixer, which in turn takes it from the current output device. Plugging in a USB interface, switching to Bluetooth headphones or changing an output device in the system settings can change the number mid-session. It is read without permission and without producing sound, which is why it appears in fingerprinting surfaces at all — nobody needs to hear anything for the value to be available.
How rubric 1.0.0 prices the audio sample rate
Three outcomes. 48000 Hz is assumed at 0.72 and costs 0.47 bits, the second-quietest reading in the table. 44100 Hz is assumed at 0.24 and costs 2.06 bits. Every other rate — 96000, 88200, the 16000 a virtual device might negotiate — shares 0.01 and costs 6.64 bits. Audio is the sole member of its correlation group, so whatever it reads counts in full with no discount protecting the total from it.
The spread between the cheapest and the dearest outcome is 6.17 bits, and that spread is the whole of what this column can ever do to a total. Where the four band edges sit →
Every outcome this rule can produce
This is the printed envelope, written out. The pen cannot land outside it, because these are
all the answers rule audio has.
| Outcome | Assumed p | −log₂p |
|---|---|---|
| 48000 Hz modal outcome — the printed line | 0.720 | 0.47 |
| 44100 Hz | 0.240 | 2.06 |
| 96000 Hz any other rate | 0.010 | 6.64 |
What changing the audio sample rate costs
Normalising an unusual rate to 48000 saves 6.17 bits at full weight, which makes this one of the better-value corrections on the form. The catch is that the rate is strongly implied by the platform — and the rubric does not connect them, so it charges you for both. A declared platform whose mixers do not negotiate the rate you are reporting is a mismatch this form cannot see and a consistency check can.
The rubric's own reasoning, verbatim
Overwhelmingly 48000, with 44100 second. Other values indicate unusual audio hardware or a spoofer.
What moves it
Audio hardware, or an OS-level output device change. Some anti-detect tooling normalises it, which usually means setting it to 48000 and losing whatever it was.
Where this prior is weakest
Grouped with nothing else in practice, so it always takes full weight even though it is heavily determined by the platform — which is already priced by another rule. The correlation model does not currently connect them, and it probably should.
Correlation
The only member of the "audio" group, so it always takes full weight. That is worth knowing when you compare it against a grouped signal: no discount is protecting the total from it.
How the correlation discount works →
Challenge it
If you have a source that contradicts this assumption, the honest thing is to say so in
public against this specific prior. Open a challenge
against audio →
What people ask about the audio sample rate
Answered from rubric 1.0.0 itself. Where an answer quotes a probability, that probability is a constant in the shipped scoring code and is printed in the table above.
- Why does audio sample rate never get discounted?
- Because it is the only member of the audio correlation group, so there is no stronger groupmate to discount it against. It always contributes its full surprisal, which is a stated weakness given how much the platform determines it.
- What makes my sample rate change between sessions?
- The output device. Switching to Bluetooth headphones, plugging in an audio interface or changing the system output can move the negotiated rate, so this column is one of the few that can move without you touching a browser setting.
- Is 44100 Hz unusual?
- The rule assumes it is the second commonest value, at 0.24 against 0.72 for 48000, so it costs 2.06 bits rather than 0.47. Anything outside those two is priced at 0.01 and costs 6.64 bits.
Where to read next
- The complete prior table for all 13 columns — every rule, every outcome, and the rationale the rubric ships with each one.
- Assumed and counted: what a prior is, what the counted column beside it measures, and why the two are never added.
- Compare two passes for the delta on this column — change one setting and read what actually moved, instead of guessing from a total.
Where the pens move
Knowing which prior is weak is only useful if something can actually move the signal.
Engine level, not JS level. Donut Browser spoofs through CDP injection in its Wayfern Chromium fork, so the value a page reads is the value the engine reports — there is no patched accessor to catch.
It refuses to contradict itself. A profile whose proxy exit geography, time zone or language disagrees with its fingerprint does not launch. That is a pre-launch gate in the app, not advice in a manual.
Open, and quiet. AGPL-3.0, desktop app and sync server both in the repository, and no analytics SDK anywhere in the source. macOS, Windows, Linux.