From cb721a0549bd183d8795cfb3f8b1d76893c1e00e Mon Sep 17 00:00:00 2001
From: xiaolai <lixiaolai@gmail.com>
Date: Wed, 27 Mar 2024 11:18:28 +0800
Subject: [PATCH] keep updated

---
 1000-hours/sounds-of-american-english/2.4.2-cv.md | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/1000-hours/sounds-of-american-english/2.4.2-cv.md b/1000-hours/sounds-of-american-english/2.4.2-cv.md
index 22d4ea7e..c02108f8 100644
--- a/1000-hours/sounds-of-american-english/2.4.2-cv.md
+++ b/1000-hours/sounds-of-american-english/2.4.2-cv.md
@@ -38,4 +38,10 @@ In the audio clip above, the phrase *goal~of*<span class="speak-word-inline" dat
 
 Beginners often overcorrect when learning English, trying to merge all "consonant + vowel" combinations they encounter. Remember, it's *possible* to connect sounds, but it's not always *necessary*. This largely depends on personal speaking habits. More importantly, when speaking, one should divide their speech into *thought groups*, pausing between each group. Do not blend the ending consonant of one thought group with the beginning vowel of the next.
 
+> [!Note]
+>
+> 目前（2024.03） OpenAI 的 TTS 引擎生成的语音中，*`l` + vowels* 的组合总是没有连读效果 —— 也许将来的版本会改进？
+>
+> As of now (March 2024), the TTS engine from OpenAI doesn't seem to blend the sound of ending `l` with following vowels smoothly. But who knows? Future versions might just improve on this.
+
 [^1]: This audio clip is extracted from the Audible edition of Knowledge and Decisions by Thomas Sowell.