How to setup call recording for high quality speech-to-text

If you're going to transcribe phone call recordings with speech-to-text or speech recognition service (Google Speech-to-Text, Amazon Transcribe, OpenAI Whisper, Azure Cognitive Services), you will likely use a third party speech recognition system. These systems usually recommend to use stereo recordings in which the voices are recorded to different channels. You may use mono recordings, but the speech recognition quality will be lower.

You can configure Softphone.Pro to record calls in stereo. To achieve that please do the following:

1. Go to Settings - SIP settings page and find Call recording section:

Open SIP settings

2. Make sure that Start call recording automatically option is enabled.

3. Select the type of recording in the Channels parameter:

Pick stereo or mono recording

Available options:

  • Stereo - each side of the call is recorded to a separate channel, the resulting recording file occupies more space on the disk;
  • Mono - both sides of the call are recorded to one channel, the resulting recording file is smaller.

4. Select the file format of the recordings in the Format parameter:

Pick recording format

Available options:

  • wav - uncompressed recording file which occupies more space on the disk;
  • mp3 - compressed recording file which occupies less space on the disk.

5. Click Save.