Enable high-fidelity voice support

AI-generated summary

Genesys Cloud enables high-fidelity voice support through a 16 kHz audio capability for native and third-party text-to-speech (TTS), virtual and human agents, recordings, and Architect pre-recorded prompts. Administrators with Telephony > Settings > All permissions can activate this feature via the "Prefer Hi-Fi TTS" option in Global Telephony Settings, which generates clearer, more natural-sounding voice prompts compared to standard 8 kHz telephony audio. When enabled, call recordings are stored and downloaded at 16 kHz, enabling higher-quality audio analysis, though this increases file sizes and associated storage costs. Key limitations include: the setting only affects TTS providers supporting 16 kHz output; the Genesys TTS Connector always outputs 16 kHz PCM regardless of underlying provider formats; high-fidelity TTS is supported exclusively in cloud-only infrastructures, excluding Local Deployment Model (LDM) and BYOC Premises Edge architectures. Audio quality depends on end-to-end 16 kHz support across the entire call path, and mixing TTS prompts with 8 kHz pre-recorded prompts creates noticeable quality differences unless pre-recorded prompts are reuploaded at 16 kHz.

Feature coming soon: High-fidelity voice support.

PrerequisitesClick to expand

Administrators can enable 16 kHz high-fidelity audio across Genesys Cloud native and third-party text-to-speech (TTS), virtual and human agents, recordings, and Architect pre-recorded prompts through the Prefer Hi-Fi TTS option in Global Telephony Settings. When enabled, supported TTS providers and Architect voice flows generate speech at a higher audio quality than standard 8 kHz telephony audio, resulting in clearer, more natural-sounding voice prompts.

For more information about high fidelity audio and text-to-speech support, see Text-to-speech FAQs.

To enable high-fidelity TTS.

Click Menu > Digital and Telephony > Telephony > Global Telephony Settings.
Under High-fidelity text-to-speech, select Prefer Hi-Fi TTS.
Click Save.

Note: When the high-fidelity (Prefer Hi-Fi TTS) setting is enabled for your organization, call recordings are stored and downloaded at 16 kHz, allowing for higher-quality analysis of recorded audio. Because 16 kHz recordings contain more data than standard 8 kHz recordings, file sizes increase, which may also increase associated storage costs.

Warnings:

Actual audio quality depends on end-to-end support for 16 kHz audio across the entire call path. If any part of that path — such as a PSTN connection — does not support wideband audio, the audio is downsampled accordingly.
If your flow mixes TTS prompts with pre-recorded audio prompts, pre-recorded prompts still play at 8 kHz, regardless of the Prefer Hi-Fi TTS setting. This creates a noticeable difference in audio quality between TTS and pre-recorded prompts within the same flow. To avoid this, reupload pre-recorded prompts at 16 kHz; when the setting is enabled, they then play at 16 kHz.

Limitations

This setting only takes effect for TTS providers that support 16 kHz audio output. Enabling it with an unsupported provider has no effect, so confirm provider support before relying on this setting.
Genesys TTS Connector always plays audio as 16 kHz PCM, regardless of the audio formats the underlying TTS provider supports.
High fidelity TTS is supported exclusively in cloud only infrastructures. Local Deployment Model (LDM) recording and BYOC Premises Edge architectures are not supported.

[NEXT] Was this article helpful?

Get user feedback about articles.