How do we solve an AXSpeechAssetDownloader error on iOS?

Question

We have an iOS app that uses AVSpeechSynthesizer to speak. It works on the iPad and other devices - but we notice it is not working on our iPhone 6 Plus.

When examining the console output we see this error:

|AXSpeechAssetDownloader|error| ASAssetQuery error fetching results Error Domain=ASError Code=21 "The operation couldn’t be completed. (ASError error 21 - Unable to copy asset information)" UserInfo=0x174a7e100 {NSDescription=Unable to copy asset information}

The device on which the app is running does have a network connection.

Any ideas how to even begin solving this?

Answer 1

I was having the same issue. I couldn't find any help, so, I had to debug it by trying various things.

What I found is that if you access Settings -> General -> Speech, and enable "Speak Selection" and muck about with the English voice, it will just start working for you.

Let me know if this helps.

Answer 2

This means that the resources required to speak the required languages have not been downloaded, and the app failed to do that automatically. as @softwarenerd mentions, you can go to Settings -> General -> Speech, and then go to Voices and download whatever voices you need.

But then, this is not really a good solution if you are developing the app for the app store. There must be a way to handle the download automatically and on demand.

Answer 3

I see the exact same problem on iOS9. Here is my work-around, which is not perfect but at least avoids the crash.

let voices = AVSpeechSynthesisVoice.speechVoices()
for voice in voices {
    if lang == voice.language {
        utterance.voice = voice
        break;
    }
}

Answer 4

You may see this error also if you are passing string into AVSpeechUtterance constructor in language A, but asking to speech it using language B.

Example:

let utterance = AVSpeechUtterance(string: "Hello")
utterance.voice = AVSpeechSynthesisVoice(language: "pl-PL")

In this case "Hello" is not a Polish language (pl-PL). App shows such Error and may (in some cases) speeches text.

Answer 5

For me, I was able to resolve this by setting the utterance.voice after the utterance string.

speechUtterL1 = AVSpeechUtterance(string: ("أتمنى لك نهارا سعيد")) 
speechUtterL1.voice = AVSpeechSynthesisVoice(language: "ar-SA")

The reverse, where I set the voice only once results in an error:

[TTS] Error running custom voice query Error Domain=ASError Code=15 "Unable to copy asset information from https://mesu.apple.com/assets/ for asset type com.apple.MobileAsset.VoiceServices.GryphonVoice" UserInfo={NSDescription=Unable to copy asset information from https://mesu.apple.com/assets/ for asset type com.apple.MobileAsset.VoiceServices.GryphonVoice}