Xiaoyi: Text to Speech

Xiaoyi voice

The voice Xiaoyi is available in the Azure Text-to-Speech service for the Chinese language.

How to use Xiaoyi voice in your videos

To use Xiaoyi voice in your videos, you can use the following JSON2Video code:

{
    "type": "voice",
    "model": "azure",
    "voice": "zh-CN-XiaoyiNeural",
    "text": "\u5728\u6625\u5929\uff0c\u82b1\u56ed\u91cc\u5145\u6ee1\u4e86\u4e94\u989c\u516d\u8272\u7684\u82b1\u6735\u548c\u9e23\u79bd\u3002 \u53e4\u8001\u7684\u6a61\u6811\u4e3a\u8bbf\u5ba2\u63d0\u4f9b\u9634\u51c9\uff0c\u800c\u8774\u8776\u5728\u73ab\u7470\u4e1b\u4e2d\u8df3\u821e\u3002 \u4e00\u4e2a\u5c0f\u5c0f\u7684\u55b7\u6cc9\u53d1\u51fa\u5b81\u9759\u7684\u58f0\u97f3\uff0c\u4f7f\u8fd9\u91cc\u6210\u4e3a\u653e\u677e\u5e76\u4eab\u53d7\u5927\u81ea\u7136\u7f8e\u4e3d\u7684\u5b8c\u7f8e\u573a\u6240\u3002"
}

$scene->addElement([
	'type' => 'voice',
	'model' => 'azure',
	'voice' => 'zh-CN-XiaoyiNeural',
	'text' => '在春天，花园里充满了五颜六色的花朵和鸣禽。 古老的橡树为访客提供阴凉，而蝴蝶在玫瑰丛中跳舞。 一个小小的喷泉发出宁静的声音，使这里成为放松并享受大自然美丽的完美场所。'
]);

scene.addElement({
	"type": "voice",
	"model": "azure",
	"voice": "zh-CN-XiaoyiNeural",
	"text": "在春天，花园里充满了五颜六色的花朵和鸣禽。 古老的橡树为访客提供阴凉，而蝴蝶在玫瑰丛中跳舞。 一个小小的喷泉发出宁静的声音，使这里成为放松并享受大自然美丽的完美场所。"
});

Xiaoyi supports SSML

SSML stands for Speech Synthesis Markup Language. It's a way to add instructions to your text so that a Text-To-Speech (TTS) system knows how to read it aloud.

You use SSML like HTML, but for controlling speech. It helps you adjust things like: Pronunciation, Pauses, Pitch and Volume, Emphasis, Speaking Rate.

{
    "type": "voice",
    "voice": "zh-CN-XiaoyiNeural",
    "text": "<speak>Hello, <break time="500ms"/> how are you today? <emphasis level="strong">This is important!</emphasis></speak>"
}

$scene->addElement([
	'type' => 'voice',
	'voice' => 'zh-CN-XiaoyiNeural',
	'text' => '<speak>Hello, <break time="500ms"/> how are you today? <emphasis level="strong">This is important!</emphasis></speak>'
]);

scene.addElement({
	"type": "voice",
	"voice": "zh-CN-XiaoyiNeural",
	"text": "<speak>Hello, <break time="500ms"/> how are you today? <emphasis level="strong">This is important!</emphasis></speak>"
});

Xiaoyi supports different voice styles

As part of SSML, you can use the style tags to change the voice style.

Xiaoyi supports these styles: affectionate angry cheerful disgruntled embarrassed fearful gentle sad serious

{
    "type": "voice",
    "voice": "zh-CN-XiaoyiNeural",
    "text": "<whispering>I have a secret for you</whispering>"
}

$scene->addElement([
	'type' => 'voice',
	'voice' => 'zh-CN-XiaoyiNeural',
	'text' => '<whispering>I have a secret for you</whispering>'
]);

scene.addElement({
	"type": "voice",
	"voice": "zh-CN-XiaoyiNeural",
	"text": "<whispering>I have a secret for you</whispering>"
});

Xiaoyi is a neural voice

In Azure Cognitive Services, a Neural voice refers to a voice generated using neural network technology. This means the Text-To-Speech system uses advanced machine learning models to create more natural, human-like speech compared to traditional methods.

Key characteristics of Neural voices:

• More expressive and realistic
• Better at handling pitch, tone, and rhythm variations
• Sounds closer to how humans naturally speak