Fact · Languages and accents
7.3times
The best model's average word error rate on EdAcc conversational speech was about 7.3 times its word error rate on US English clean read speech.
Checked against its source on October 17, 2026
- What it measures
- A ratio of two averages from the same paper. It compares two test conditions (accented conversation and US clean read speech), so it is a size of the gap, not a measure of accent alone.
- How it is calculated
- 19.7 / 2.7 = 7.3
- What it does not show
- The two conditions differ in speaking style as well as accent, so the ratio mixes both effects. Use it as a rough indicator only.
Calculated from
- On the EdAcc accent test set of conversations between friends, the best-performing English speech recognition model, trained on 680 thousand hours of transcribed data, averaged a word error rate of 19.7 percent, against 2.7 percent on US English clean read speech.
- The same best English speech recognition model, tested by the EdAcc authors on US English clean read speech, had a word error rate of 2.7 percent.