edge_tts.exceptions.NoAudioReceived: No audio was received. Please verify that your parameters are correct.

[kamimachi]

When I call the speech generation function multiple times with the code, I get the error: edge_tts.exceptions.NoAudioReceived: No audio was received. Please verify that your parameters are correct.， last week all is ok， but today is have the error

[rany2]

How are you calling it? Can you show code that has this issue? Maybe the voice you are using got deprecated

[syh-667]

Mine is also this problem. Sometimes it works, sometimes it doesn't.

[guoguozhang]

Mine is also this problem. Sometimes it works, sometimes it doesn't. I use it in multiprocessing.

[guoguozhang]

edge_tts.exceptions.NoAudioReceived: No audio was received. Please verify that your parameters are correct.
Error: Command 'edge-tts --voice zh-CN-XiaoyiNeural --text "hello world" --write-media ./output/56EBE590-8671-457A-9299-D9453B52CB8F.mp3' returned non-zero exit status 1.

[rany2]

@guoguozhang I'm not entirely sure what the issue is, but seems to be working reliably on my end. Perhaps their service is under some load/maintenance?

[guoguozhang]

This command runs fine in the command line, but when executed by multiple threads in Python, it usually results in an error on Windows.

[rany2]

@guoguozhang Then Microsoft increased their protections against abuse. That is the only explanation.

[guoguozhang]

But I don't make frequent calls, maybe once an hour, and the first call fails because this feature is still in the testing phase.

[zixingonline]

Now it's zh-CN-XiaoyiNeural, this sound is gone forever, right? . .

[kamimachi]

When I receive a sentence, I split the sentence with commas and periods, divide the sentence into several paragraphs, and call the speech-generated function once each time I split. The purpose is to reduce the delay, and I can generate speech content while receiving sentences sent to me by others, rather than waiting for me to send them and then generate speech.

This feature was completed last week and was tested without any issues. But when testing today, he made an error, when splitting the sentence, the first split content into the function is no problem, but immediately after the second speech into the speech generation function, there will be no reply, an error will appear.

Eventually I replaced the original edge_tts package with a VITS model to generate speech.

[rany2]

I'll study this, hopefully it's a service degradation issue and not an attempt to block the library.

[rany2]

Is anyone still experiencing this issue? I can't trigger it and I've been running hundreds of threads in parallel just constantly launching TTS generations.

[rany2]

I presume this issue no longer occurs or still relevant? @kamimachi @guoguozhang

[kamimachi]

I presume this issue no longer occurs or still relevant? @kamimachi @guoguozhang

Thank you very much for your reply, I just went to test the code that used edg_tts for speech generation before and found that there were no problems. But I decided to use the VITS model anyway because he is not prone to accidents.

[rany2]

@kamimachi Fair enough, so I'll close this because it appears to have been a server issue on Microsoft's end.