TTS: Handshake Status 403 Forbidden

Question

TTS: Handshake Status 403 Forbidden

nathannzw opened this issue 5 years ago · 17 comments

voices = service.list_voices().get_result()
print(json.dumps(voices, indent=2))

voice = service.get_voice('zh-CN_ZhangJingVoice').get_result()
print(json.dumps(voice, indent=2))

service.synthesize_using_websocket(#SSML_text,
                                   "你好， 需要什么帮忙吗？",
                                   test_callback,
                                   accept='audio/wav',
                                   voice="zh-CN_ZhangJingVoice"
                                  )

Getting the voice does show that particular Chinese voice model.
Synthesizing causes a handshake error - English voice works.

I assume it's because the Chinese voice isn't showing up from the list voices function.

Why is this so??

mamoonraja commented 5 years ago

E.g, #719

Answer 1 · 2020-04-08T08:11:00.000Z

Can someone else help me verify this?

Answer 2 · 2020-04-08T12:16:58.000Z

Using this voice should work when using WebSocket, can you provide more info on your auth setup? 403 looks like an issue related to the incorrect service endpoints.

Answer 3 · 2020-04-08T13:01:19.000Z

@mamoonraja here:

#text-to-speech    
tts = IAMAuthenticator("xxxxxxxx7pGI7mTrJ0WjMiQi2z9TSj4hGDKei")
service = TextToSpeechV1(authenticator=tts)
service.set_service_url('https://api.jp-tok.text-to-speech.watson.cloud.ibm.com/instances/xxxxxxxxbd8-a574-a1b54d6558c9')

Answer 4 · 2020-04-08T13:01:59.000Z

@mamoonraja If it is related to the endpoints, why does it work for English voice and not the Chinese one?

Answer 5 · 2020-04-08T13:18:48.000Z

oh ok, the same thing works for English. Interesting, I was going that way cause we have seen 403 cause of url issues before.

Answer 6 · 2020-04-08T13:33:43.000Z

I was able to reproduce the issue. Looks like a service issue, we will investigate and get back to you. Meanwhile, you can use recognize instead of recognize_using_websocket because that worked with the Chinese language.

Answer 7 · 2020-04-09T07:02:27.000Z

Ok, good to hear. Thanks!

Answer 8 · 2020-04-09T07:49:35.000Z

@mamoonraja and you mean synthesize right?

Answer 9 · 2020-04-09T12:34:12.000Z

oh yes, my bad.

Answer 10 · 2020-06-23T14:52:11.000Z

@nathannzw were you able to solve the issue with using synthesize?

Answer 11 · 2021-02-15T20:37:59.000Z

Hello! I am also interested in using websocket to get text to speech from IBM with the chinese voices. I want to obtain the timestamps, but I also get a 403 error. I have checked with a wide range of languages and see it supports more common languages like english, french and portugese. Is there any other way to obtain the timestamps?

Answer 12 · 2021-05-06T11:32:46.000Z

Hello I am also having this issue. Has anyone figured it out?

Answer 13 · 2021-06-17T01:30:01.000Z

Looking into this issue again, we have located the source of the error and determined a viable solution. We expect a patch release for this in the coming days

Answer 14 · 2021-06-28T16:58:13.000Z

This issue has been patched in version 5.2.1. Closing as resolved

Answer 15 · 2021-06-29T16:36:51.000Z

This issue has been patched in version 5.2.1. Closing as resolved

Can you please explain what you mean with 5.2.1
I have the same problem and don´t know how to fix it.

Answer 16 · 2021-07-02T14:09:25.000Z

Using version 5.2.1 of the ibm-watson python package has the necessary patch and will resolve this issue for you