I have recorded my voice to generate an AI-voice. I am from northern england and say words such as glass, bath, castle differently from those in southern england. In my voice training I uploaded a file of my voice with these types of words in it, however, when I upload a script that has these words they are pronounced in a southern english style. This is know as the TRAP-BATH split. Can you please try (with Eleven labs) to fix this.