Introduction

In social and entertainment scenarios, users often want various voice effects to improve their interactive experiences. For example, in chat rooms, a user can select a voice effect to add a virtual stereo effect to their voice.

To accomplish this, Agora RTC SDK provides preset voice effects. You can also dynamically change the users' voices, such as adjusting the pitch, setting the equalization, and reverberation modes. Try out the preset voice effects on the online Demo provided by Agora.

Sample project

Agora provides an open-source sample project (BeautyAudio) that implements the voice beautifier and audio effects functions on GitHub. You can download it to try out this function and view the source code.

Implementation

Before proceeding, ensure that you have implemented the basic real-time communication functions in your project.

Use preset voice effects

To meet the requirements for voice effects in different scenarios, Agora provides setVoiceBeautifierPreset and setAudioEffectPreset. You can use the following SDK preset voice effects:

Chat beautifier

Chat beautifier refers to beautifying the characteristics of male or female voices without altering the original voice beyond recognition.

You can implement the chat beautifier by the enumerations in setVoiceBeautifierPreset as follows:

VOICE_BEAUTIFIER_PRESET preset;
// Beautifies the local voice by making it sound more magnetic (for male only).
preset = CHAT_BEAUTIFIER_MAGNETIC;
rtcEngine.setVoiceBeautifierPreset(preset);
// Disables voice effects.
preset = VOICE_BEAUTIFIER_OFF;
rtcEngine.setVoiceBeautifierPreset(preset);

Timbre transformation

Timbre transformation changes the timbre of a voice in a specific way. Users can choose the most appropriate effect for their voice.

You can implement the timbre transformation by the enumerations in setVoiceBeautifierPreset as follows:

VOICE_BEAUTIFIER_PRESET preset;
// Beautifies the local voice by making it sound more vigorous.
preset = TIMBRE_TRANSFORMATION_VIGOROUS;
rtcEngine.setVoiceBeautifierPreset(preset);
// Disables voice effects.
preset = VOICE_BEAUTIFIER_OFF;
rtcEngine.setVoiceBeautifierPreset(preset);

Voice changer effect

Voice changer effects adjust the voice to make it sound different from the original.

You can implement the voice changer effect by the enumerations in setAudioEffectPreset as follows:

AUDIO_EFFECT_PRESET preset;
// Presets the local voice by making it sound like the Hulk.
preset = VOICE_CHANGER_EFFECT_HULK;
rtcEngine.setAudioEffectPreset(preset);
// Disables voice effects.
preset = AUDIO_EFFECT_OFF;
rtcEngine.setAudioEffectPreset(preset);

Style transformation

A style transformation makes singing more harmonious for specific styles of songs.

You can implement the style transformation by the enumerations in setAudioEffectPreset as follows:

AUDIO_EFFECT_PRESET preset;
// Presets the local voice to the style of pop music.
preset = STYLE_TRANSFORMATION_POPULAR;
rtcEngine.setAudioEffectPreset(preset);
// Disables voice effects.
preset = AUDIO_EFFECT_OFF;
rtcEngine.setAudioEffectPreset(preset);

Room acoustics

Room acoustics refers to the spatial dimension added to a user’s voice for making the voice seem to come from a specific type of enclosed place.

You can implement the space construction by the enumerations in setAudioEffectPreset as follows:

AUDIO_EFFECT_PRESET preset;
// Presets the local voice by making it appear to be moving around the user.
preset = ROOM_ACOUSTICS_3D_VOICE;
rtcEngine.setAudioEffectPreset(preset);
// Changes the cycle period to 30 seconds.
rtcEngine.setAudioEffectParameters(preset, 30, 0);
// Disables voice effects.
preset = AUDIO_EFFECT_OFF;
rtcEngine.setAudioEffectPreset(preset);

Pitch correction

Pitch correction refers to the effect for correcting a singing voice to make it match the song based on the basic mode and the tonic pitch.

You can implement the pitch correction by the enumerations in setAudioEffectPreset as follows:

AUDIO_EFFECT_PRESET preset;
// Presets the local voice by making it match the song.
preset = PITCH_CORRECTION;
rtcEngine.setAudioEffectPreset(preset);
// Changes the basic mode to natural minor scale and the tonic pitch to D.
rtcEngine.setAudioEffectParameters(preset, 2, 6);
// Disables voice effects.
preset = AUDIO_EFFECT_OFF;
rtcEngine.setAudioEffectPreset(preset);

Customize voice effects

If the preset effect does not meet your requirement, you can also customize the voice effects by adjusting the voice pitch, equalization, and reverberation settings through setLocalVoicePitch, setLocalVoiceEqualization, and setLocalVoiceReverb.

The following sample code shows how to change the original user's voice to the Hulk's voice.

// Sets the pitch. The value ranges between 0.5 and 2.0. The lower the value, the lower the pitch. The default value is 1.0, which is the original pitch.
int nRet = rtcEngine.setLocalVoicePitch(0.5);

// Sets the local voice equalization.
// The first parameter sets the band frequency. The value ranges between 0 and 9. Each value represents the center frequency of the band: 31, 62, 125, 250, 500, 1k, 2k, 4k, 8k, and 16k Hz.
// The second parameter sets the gain of each band. The value ranges between -15 and 15 dB. The default value is 0.
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_31, -15);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_62, 3);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_125, -9);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_250, -8);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_500, -6);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_1K, -4);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_2K, -3);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_4K, -2);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_8K, -1);
nRet = rtcEngine.setLocalVoiceEqualization(AUDIO_EQUALIZATION_BAND_16K, 1);

// The level of the dry signal in dB. The value ranges between -20 and 10.
nRet = rtcEngine.setLocalVoiceReverb(AUDIO_REVERB_DRY_LEVEL, 10);

// The level of the early reflection signal (wet signal) in dB. The value ranges between -20 and 10.
nRet = rtcEngine.setLocalVoiceReverb(AUDIO_REVERB_WET_LEVEL, 7);

// The room size of the reverberation. A larger room size means a stronger reverberation. The value ranges between 0 and 100.
nRet = rtcEngine.setLocalVoiceReverb(AUDIO_REVERB_ROOM_SIZE, 6);

// The length of the initial delay of the wet signal (ms). The value ranges between 0 and 200.
nRet = rtcEngine.setLocalVoiceReverb(AUDIO_REVERB_WET_DELAY, 124);

// The reverberation strength. The value ranges between 0 and 100. The higher the value, the stronger the reverberation.
nRet = rtcEngine.setLocalVoiceReverb(AUDIO_REVERB_STRENGTH, 78);

API reference

Preset voice effects

• setAudioEffectPreset

• setVoiceBeautifierPreset

• setAudioEffectParameters

• Customized voice effects*

• setLocalVoicePitch

• setLocalVoiceEqualization

• setLocalVoiceReverb

Considerations

• Do not set the profile parameter of setAudioProfile to AUDIO_PROFILE_SPEECH_STANDARD(1) or AUDIO_PROFILE_IOT(6); otherwise, setAudioEffectPreset, setAudioEffectParameters or setVoiceBeautifierPreset method call fails.
• All methods mentioned in this article work best with the human voice. Agora does not recommend using them for audio containing both music and voice.
• Agora recommends not calling the following methods together, because the method called subsequently overrides whichever one is called before it:
  • setAudioEffectPreset
  • setVoiceBeautifierPreset
  • setLocalVoicePitch
  • setLocalVoiceEqualizationOfBandFrequency
  • setLocalVoiceReverbOfType