Update record-custom-voice-samples.md

Author: sally-baolian

articles/cognitive-services/Speech-Service/record-custom-voice-samples.md

@@ -109,7 +109,7 @@ Below are some general guidelines that you can follow to create a good corpus (r
 
 ### Difference between script for voice talent and script for training
 
-The sample scripts we provided on [GitHub](https://github.com/Azure-Samples/Cognitive-Speech-TTS/tree/master/CustomVoice/script) are just defined for voice talent. If you use the sample scripts to upload for traning, you must normalize them in their spoken form. The scripts prepared for voice talent need to follow the native reading conventions, such as 50% and $45, while the scripts used for training need to be normalized to make sure that the scripts match the audio content, such as *fifty percent* and *forty-five dollars*. Make sure the scripts used for training match the recordings of your voice talent, especially scripts contaning digits, symbols, abbreviation, date, and time. We provide a few examples of text normalization rules and explain the difference between script for voice talent and script for training.
+The sample scripts we provided on [GitHub](https://github.com/Azure-Samples/Cognitive-Speech-TTS/tree/master/CustomVoice/script) are just defined for voice talent. If you use the sample scripts to upload for training, you must normalize them in their spoken form. The scripts prepared for voice talent need to follow the native reading conventions, such as 50% and $45, while the scripts used for training need to be normalized to make sure that the scripts match the audio content, such as *fifty percent* and *forty-five dollars*. Make sure the scripts used for training match the recordings of your voice talent, especially scripts contaning digits, symbols, abbreviation, date, and time. We provide a few examples of text normalization rules and explain the difference between script for voice talent and script for training.
 
 | Category |Script for voice talent<br> (non-normalized) | Script for training <br> (normalized) |
 | --------- | --------- | --------------------------- |