Update README.md

README.md

@@ -9,10 +9,9 @@ The ABX-accent project is based on the preparation and evaluation of the Accente
 
 What is ABX Evaluation?
 -----------------------
-The ABX metric evaluates whether a representation X of a speech unit (e.g., the word “bap”) is closer to a same-category example A (also “bap”) than to a different-category example B (e.g., “bop”). The ABX error rate is calculated by averaging the classification errors over all minimal phoneme trigrams in the corpus.
-
-This benchmark focuses on the more challenging ABX across/within speaker task, where the X example is spoken by a different speaker than the ones in pair (A, B), testing speaker-invariant phonetic discrimination.
+The ABX metric evaluates whether a representation X of a speech unit (e.g., the triphone “bap”) is closer to a same-category example A (also “bap”) than to a different-category example B (e.g., “bop”). The ABX error rate is calculated by averaging the discrimination errors over all minimal triphone pairs (ie., differing only by the central phoneme) in the corpus.
 
+This benchmark focuses on the more challenging ABX across speaker task, where the X example is spoken by a different speaker than the ones in pair (A, B), testing speaker-invariant phonetic discrimination.
 About the Dataset
 -----------------