UHD592 LEXI Voice Quick Start Guide

LEXI Voice is a LEXI add-on that produces a synthetic, translated audio track from a LEXI Translate caption stream and returns it to the encoder over iCap. On the UHD592, the translated audio is tied to a caption service you select (S1–S6) and embedded on the SDI output.

This guide follows one worked example end to end: English primary captions on Service 1, with a Spanish LEXI Voice output on Service 2. 

Note: The UHD592 supports a maximum of one concurrently enabled LEXI Voice output (mono or stereo).

Before you begin

LEXI Voice sits at the end of a chain: source audio → LEXI Live captions (S1) → LEXI Translate with Voice (S2) → audio returned to the encoder over iCap.

Confirm all nine conditions below before you configure anything. 

Important: LEXI Translate and LEXI Voice will not activate unless the S2 access code is shared with EEGASR and linked to the same encoder as S1. Source audio must also be present and not clipping on the SDI video input. Confirm this on the Main page under Audio Meters. The Input Programs meters should be moving and not pinned at the top of the scale.
 

Step 1: iCap Audio Settings

1. Select the Settings icon next to the encoder name.
   
    
2. In the navigation bar, select iCap Audio Settings. This defines the audio the encoder sends up to iCap for LEXI to transcribe.
   
   Input Select chooses between Embedded (SDI embedded audio) and Analog. Embedded exposes four input programs (Program 0–3) and Analog exposes Program 0 only.
   • Program 0 is the main audio LEXI listens to and what gets transcribed.
   • Programs 1–3 are additional input channels. They provide the audio input for LEXI transcription jobs on S2–S4 respectively, and the base audio a voice job mixes into.
     
     In this example, we will be setting up Embedded with Program 0 configured.
     
     iCap Audio Settings with Embedded input
     
     
     iCap Audio Settings with Analog input
      
3. Select either Embedded or Analog as your Input Selection.
4. Configure Program 0:
   • Audio Group — Select which Audio Group (1–4) carries your source audio.
   • Audio Mix — The encoder will create a mono track for LEXI to transcribe, mixing (if necessary) down from Surround or Stereo. Set this based on how audio tracks are arranged in your SDI video input.
   • Scale Audio — Trim the input level if needed, from +12 to −12 dB. Leave at 0 dB if the level does not require adjustment.

Step 2: LEXI Voice Settings

Each voice output is tied to a caption service (S1–S6); enabling voice on a service drives audio for that same caption path, and all enabled services must share the same Audio Group.

Note: The UHD592 supports a maximum of one concurrently enabled LEXI Voice output.

1. In the navigation bar, select LEXI Voice Settings.

2. Set Service to the service that matches your Translate service (S1–S6) set up in iCap/EEG Cloud. In this example, Service 2 (Spanish).

   

3. Set Enable Voice to On. This will allow you to set up your Voice Service configuration.

   

4. Configure the Voice Service and fine tune your LEXI Voice output:

   1. Audio Group — Select the Audio Group (1–4) the voice output is placed on.
      • In this example, we will set this to Audio Group 1.
   2. Channel Config & Channel — Set Mono or Stereo, then choose the Mono Channel or Stereo Pair the voice lands on.
      • In this example, we will set this to Stereo and Channels 3 & 4.
   3. Mix Mode — Choose how the voice is combined with the input audio; Off keeps it on its own channels.
      • Off: Only Voice audio is used.
      • Background: Input channels are mixed with Voice audio.
      • Ducking: Background audio is reduced by a specified amount when Voice is present.
   4. Audio Delay — Add 0 to 20 seconds of delay to keep the voice in sync with the background.
   5. Voice Level — Set the voice loudness, from −20 to 0 dB.
   6. Voice FX — Shape how the voice sounds:
      • Reverb — Set the amount of reverb added to the voice.
      • Spatialization — Choose a room preset (e.g. Studio) for a more natural, with consideration of space.

   

5.  Confirm and watch the Voice Amplitude meter register level.
   • Note: Ensure that your EEGCloud Translate instance is toggled On. This is required before you can see the Voice Amplitude value reflected. 
6. Click Confirm to save.

Validate your voice output: Audio Meters

You have now configured LEXI Voice with the UHD592 Encoder. Let's confirm the job is live:

On the Main page, navigate to the Audio Meters tab. 

The voice appears in two places: under Voice (via iCap) — the voice arriving from EEG Cloud on your chosen service — and on the SDI output, where that voice is embedded on the channels you set in LEXI Voice Settings (the Audio Group and Channel Config selections). Input Programs 0–3 confirm the source audio.

To hear the result, select the preview icon on the chosen Service meter. The encoder plays that channel back in real time, so you can listen to the synthesized voice as it is generated and confirm it sounds the way it will be heard. To fine-tune the output, return to LEXI Voice Settings in the Edit modal and adjust from there.

The VU meters show how close the voice is to its peak level. The meters are also color-coded: a meter reads blue at a healthy level and turns yellow once the signal rises above about −19 dBFS, showing it's getting close to the top of the scale. Use this to confirm the voice is reaching a good, strong level without peaking too high.

In our example, the Voice output is set to Audio Group 1, Stereo on Channels 3 & 4. On the SDI Output meters below, Channels 3 & 4 show activity — the VU meter reading the output level confirms the voice is landing on the correct channels. 

Audio Meters — Input Programs and Voice

Audio Meters — SDI Output