3音 pilot test · February–April 2026 · 36 learners, HSK 0–3
With 3音, beginner learners understood about six times more Chinese content
Each learner tried the same tasks with Microsoft Translator or regular PinYin and with 3音, so every person is their own comparison.
10%→70%content understood, MS Translator → 3音 (median)+51 pp on average · n = 31
6×more content understood with 3音 (median)none understood less
+2.7lines of smaller print readable vs regular PinYin31 of 31 read further
87%would use 3音 · 85% would recommend it0 of 31 knew a similar tool
45%average stated chance of subscribing at $6.99/month20 of 33 said 50% or more
Nobody understood less with 3音
Share of content understood, one line per learner
MS Translator3音
Everyone read smaller print in 3音
Smallest line each learner could read on a 9-line chart
Regular PinYin3音
The big gains start at HSK 1
Extra content understood with 3音, in percentage points. Bar = median.
Liked and helpful; ease of use lags
Average rating out of 10, with 95% confidence interval
What it means
- A large, consistent effect. Understanding rose by about 50 points and legibility by almost 3 lines, and nobody got worse on either task.
- Best fit is HSK 1–3. Complete beginners gained little (median +7 pp, n = 4), while HSK 1–3 learners gained +49 to +60 pp.
- Ease of use is the one to improve. It is the lowest rating (6.4/10), and only 1 in 4 learners rated it 8 or higher.
Read with care
- A one-session pilot. A facilitator estimated understanding while knowing which tool was in use, and the order of the two tools was not recorded.
- A small sample. 36 learners, only 4 of them at HSK 3; the subgroup results are indicative.
- Stated, not actual, payment. People usually overstate how likely they are to buy, so treat 45% as an upper bound.
Files in this pack
These open from the same folder as this page.
