IOS Chrome中SpeechSynthesis eloquence语音额外朗读内容问题咨询

问题背景 #

核心朗读代码如下：

const synth = window.speechSynthesis;
const utterance = new SpeechSynthesisUtterance(text);
utterance.voice = selectedVoice;
synth.speak(utterance);

该代码在Mac Chrome上运行正常，但在iOS Chrome中，不同语音的表现出现明显差异：

• 含compact的语音（如com.apple.voice.compact.en-GB.Daniel）：仅朗读传入的text内容，符合预期。
• 含eloquence的语音（如com.apple.eloquence.en-GB.Grandma）：会额外播报调试类信息，顺序为：'speak. Voice name equals "com apple eloquence en-GB Grandma"' → 目标text内容 → 'Voice. speak.'。

iOS Chrome下en-GB语言的可用语音列表：

SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Rocko", name: "Rocko", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Shelley", name: "Shelley", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.voice.compact.en-GB.Daniel", name: "Daniel", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Grandma", name: "Grandma", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Grandpa", name: "Grandpa", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Flo", name: "Flo", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Eddy", name: "Eddy", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Reed", name: "Reed", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.eloquence.en-GB.Sandy", name: "Sandy", lang: "en-GB", localService: true, default: true}
SpeechSynthesisVoice {voiceURI: "com.apple.voice.super-compact.en-GB.Daniel", name: "Daniel", lang: "en-GB", localService: true, default: true}

差异根源 #

1. 语音引擎的设计定位不同：eloquence系列是苹果早期的语音合成引擎，带有调试播报特性——最初为了便于开发人员确认语音服务是否正常启动，会在朗读前后额外输出语音标识信息。而compact系列是苹果后续推出的轻量化生产级引擎，移除了这类调试逻辑，仅专注于目标文本的朗读。
2. 浏览器与系统语音的桥接逻辑：iOS Chrome没有自研语音合成服务，完全依赖iOS系统原生的SpeechSynthesis接口。对于eloquence这类遗留引擎的额外播报内容，Chrome的桥接代码未做过滤，导致调试信息直接被输出；而compact引擎本身无此类输出，因此表现正常。
3. 平台兼容性优先级：eloquence语音主要是为了兼容旧版iOS应用保留的，苹果并未针对现代浏览器场景做适配优化；而compact语音是面向现代应用设计的，在跨浏览器调用时行为更统一。

内容的提问来源于stack exchange，提问作者Travis Chan