Update
I installed the iOS 13 GM on a different device, and did not encounter this issue. So I'm no longer sure if this is specific to the iOS 13 GM, or something about the way I installed it. On the first device (an iPhone X), I did an upgrade (via the "Option Key + Check for Update Button" in iTunes). On the second device (an iPhone SE), I did a full restore. Not sure yet what caused it to get into this state.
Original Question
AVSpeechSynthesisVoice.speechVoices()
reports that there are no voices available when I run it on my device with the iOS 13 GM. And indeed, when I try to load a voice using AVSpeechSynthesisVoice(identifier:)
it always returns nil. It works in the simulator. I don't believe this was an issue with the previous betas, and it seems to work in iOS 13.1 beta 2 and 3.
Has anyone encountered this, and figured out any sort of workaround? I don't see any documentation of an intentional change to this behavior, so I suspect this is a bug. I intend to file a radar. But that won't help me for the impending release of iOS 13.
This behavior is reproducible in a newly created Single View project, with the following trivial viewDidLoad
function added to the included view controller.
override func viewDidLoad() {
super.viewDidLoad()
let voices = AVSpeechSynthesisVoice.speechVoices()
print("Voice Count: \(voices.count)")
}
The output is very long:
2019-09-11 10:43:09.370992-0400 SpeechTest[1617:412065] Creating client/daemon connection: C591307C-313E-4B5F-91A8-184B7E662819
2019-09-11 10:43:09.396250-0400 SpeechTest[1617:412065] Got the query meta data reply for: com.apple.MobileAsset.MacinTalkVoiceAssets, response: 0
2019-09-11 10:43:09.397519-0400 SpeechTest[1617:412065] Consumed extension
2019-09-11 10:43:09.401982-0400 SpeechTest[1617:412065] Got the query meta data reply for: com.apple.MobileAsset.MacinTalkVoiceAssets, response: 0
Voice Count: 0
2019-09-11 10:43:09.435890-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServicesVocalizerVoice, response: 0
2019-09-11 10:43:09.445269-0400 SpeechTest[1617:412066] Consumed extension
2019-09-11 10:43:09.468849-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServicesVocalizerVoice, response: 0
2019-09-11 10:43:09.499573-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServicesVocalizerVoice, response: 0
2019-09-11 10:43:09.520244-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServices.GryphonVoice, response: 0
2019-09-11 10:43:09.545243-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServices.CustomVoice, response: 0
2019-09-11 10:43:09.549941-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServices.CombinedVocalizerVoices, response: 2
2019-09-11 10:43:09.550075-0400 SpeechTest[1617:412066] [AXTTSCommon] Error running custom voice query XML not present
2019-09-11 10:43:09.569114-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServicesVocalizerVoice, response: 0
2019-09-11 10:43:09.599459-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServicesVocalizerVoice, response: 0
2019-09-11 10:43:09.630763-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServicesVocalizerVoice, response: 0
2019-09-11 10:43:09.650161-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServices.GryphonVoice, response: 0
2019-09-11 10:43:09.666885-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServices.CustomVoice, response: 0
2019-09-11 10:43:09.671279-0400 SpeechTest[1617:412066] Got the query meta data reply for: com.apple.MobileAsset.VoiceServices.CombinedVocalizerVoices, response: 2
2019-09-11 10:43:09.671412-0400 SpeechTest[1617:412066] [AXTTSCommon] Error running custom voice query XML not present
...
It repeats those last 7 lines 39 times, presumably corresponding to the 39 voices it's supposed to know about.