TTS audio level inconsistent especially when switching across APIs

The information in this article applies to:

• SpeechKit 3

Symptoms

The audio level for text-to-speech playback is inconsistent especially when switching between SAPI 4 and SAPI 5 voices.

Cause

SAPI 4 audio is managed by a vendor supplied SAPI 4 audio object where SAPI 5 audio is managed by SAPI 5.

This is why you will notice the SAPI 4 synthesizer volume setting has less impact on the volume than does the mixer audio level. This is not the case with SAPI 5. SAPI 5 volume scales the audio playback volume of the voice.

The ChantTTS component exposes the SAPI 4 audio object setting via the AudioLevel property. However, depending on the synthesizer and the system audio, too soft or too loud conditions may occur when switching voices or setting the audio levels.

It is recommended that audio level setting be done via the system audio mixer and not the SAPI 4 audio object. To support this and eliminate possible side effects, the AudioLevel property will become read only and the SpeechKit console Destination Audio and Source Audio sliders will be replaced with a button that launches the system audio mixer for adjusting levels.

Status

Changed in SpeechKit 3 SR5 build 3.0.5.6.

Component Formats Impacted

All formats.