Hi @Shubham Bajaj Is there anything tha could be done to make the transcriber (Deepgram) only recognize the main speaker? Please check call id 73570840-8d96-4eae-90bc-c3d7d832b2ba As you may see in the recording graph, the secondary voice (a TV) has a much lower amplitude but was still picked by the transcriber. Would any of the secondary Nova 2 models be helpful? https://cdn.discordapp.com/attachments/1313830083022159953/1313830083475017779/image.png?ex=67518f1d&is=67503d9d&hm=f6aab5ffffe0b2299e4bc567dd84df2758154e357a64a8f1d56fca0e2b6ababc&

@Jonas Hello! To address the issue of the transcriber picking up background noise such as a TV, you can indeed explore using different models that may offer better discrimination of the main speaker's voice. Deepgram provides various models, and some are specifically tuned to handle noisy environments or focus on a single speaker. For the call ID you've provided, we can look into the available Nova 2 models to see if there's one that might better suit your needs. For instance, there might be a model optimized for scenarios with background noise or one that is designed to focus on a single speaker. If you'd like, I can guide you through the process of selecting and testing a different model to see if it improves the transcription accuracy for your main speaker. Would you like to proceed with this?