这里讲iSphinx on CocoaPods.org

iSphinx 是基于 Pocketsphinx 引擎的离线语音识别 iOS 库。通过 Cocoapods 将语音识别功能添加到您的 iOS 应用程序中。iSphinx 为您的应用程序提供简单配置和实现，无需处理 Pocketsphinx 资产和配置。

功能

预览

Cocoapods

如何使用

添加请求权限

添加监听器

import iSphinx

class ViewController: UIViewController, iSphinxDelegete {
    var isphinx: iSphinx = iSphinx()

    override func viewDidLoad() {
        super.viewDidLoad()
        isphinx.delegete = self
    }
}

func iSphinxDidStop(reason: String, code: Int) {
    if code == 500 { // 500 code for error
        print(reason)
    } else if code == 522 { // 522 code for timed out
        print(reason)
    } else if code == 200 { // 200 code for finish speech
        print(reason)
    }
}

func iSphinxFinalResult(result: String, hypArr: [String], scores: [Double]) {
    print("Full Result : \(result)")
    // NOTE :
    // [x] parameter "result" : Give final response with ??? values when word out-of-vocabulary.
    // [x] parameter "hypArr" : Give final response in original words without ??? values.

    // Get score from every single word. hypArr length equal with scores length
    for score in scores {
        print(score)
    }

    // Get array word
    for word in hypArr {
        print(word)
    }
}

func iSphinxPartialResult(partialResult: String) {
    print(partialResult)
}

func iSphinxUnsupportedWords(words: [String]) {
    var unsupportedWords = ""
    for word in words {
        unsupportedWords += word + ", "
    }
    print("Unsupported words : \(unsupportedWords)")
}

func iSphinxDidSpeechDetected() {
    print("Speech detected!")
}

准备语音识别

在使用之前，您需要准备语音识别。您可以在 onPreExecute 中添加新参数以增加准确度或性能。

isphinx.prepareISphinx(onPreExecute: { (config) in
    // You can add new parameter pocketshinx here
    self.isphinx.silentToDetect = 1.0
    self.isphinx.isStopAtEndOfSpeech = true
    // config.setString(key: "-parameter", value: "value")
}) { (isSuccess) in
    if isSuccess {
        print("Preparation success!")
    }
}

更新语言模型/语法

您可以在不中断的情况下即时更新词汇表、语言模型或 JSGF 语法。确保在更新词汇/语法之前删除标点符号。iSphinx 中默认要移除的标点是 (.), (,), (?), (!), (_), (-), (\), (:)

// Update vocabulary with language model from single string
isphinx.updateVocabulary(text: "YOUR VOCABULARIES!", oovWords: ["WORDS DISTRUBER", ...]) {
    print("Vocabulary updated!")
}

// Update vocabulary with language model from array string
isphinx.updateVocabulary(text: "YOUR VOCABULARIES!", oovWords: ["WORDS DISTRUBER", ...]) {
    print("Vocabulary updated!")
}

// Update vocabulary with JSGF Grammar from string
isphinx.updateGrammar(text: "YOUR GRAMMAR", oogWords: ["WORDS DISTRUBER", ...]) {
    print("Grammar updated!")
}

开始语音识别

// Start speech recognition with Timeout in seconds
isphinx.startISphinx(timeoutInSec: 10)

// Start speech recognition without Timeout
isphinx.startISphinx()

iSphinx 1.2.0

iSphinx 1.2.0

iSphinx 1.2.0

iSphinx

功能

预览

Cocoapods

如何使用

添加请求权限

添加监听器

准备语音识别

更新语言模型/语法

开始语音识别

停止语音识别

播放音频记录

注意：请参阅 iSphinxDemo 以获取详细信息。

MIT 许可证