R.i.p Edge TTS

[alfamus]

Doesn't work for the last two days. 403 Bad gate. Forbidden error.

[samsonsin]

It worked for me last night, but I have been getting this error all day:

go.Universe$proxyerror: websocket: bad handshake: 403 Forbidden
	at tts_server_lib.EdgeApi.getEdgeAudio(Native Method)
	at com.github.jing332.tts_server_android.model.speech.tts.h.getAudio(MsTTS.kt:385)
	at com.github.jing332.tts_server_android.model.speech.tts.e.getAudioWithSystemParams(ITextToSpeechEngine.kt:23)
	at com.github.jing332.tts_server_android.model.speech.tts.e.getAudioWithSystemParams$default(ITextToSpeechEngine.kt:14)
	at u5.w0$i$b.invokeSuspend(SysTtsListItemHelper.kt:45)
	at cc.a.resumeWith(ContinuationImpl.kt:9)
	at tc.j0.run(DispatchedTask.kt:112)
	at yc.h$a.run(LimitedDispatcher.kt:4)
	at ad.k.run(Tasks.kt:3)
	at ad.a$a.run(CoroutineScheduler.kt:96)

[alfamus]

It worked for me last night, but I have been getting this error all day:

go.Universe$proxyerror: websocket: bad handshake: 403 Forbidden
	at tts_server_lib.EdgeApi.getEdgeAudio(Native Method)
	at com.github.jing332.tts_server_android.model.speech.tts.h.getAudio(MsTTS.kt:385)
	at com.github.jing332.tts_server_android.model.speech.tts.e.getAudioWithSystemParams(ITextToSpeechEngine.kt:23)
	at com.github.jing332.tts_server_android.model.speech.tts.e.getAudioWithSystemParams$default(ITextToSpeechEngine.kt:14)
	at u5.w0$i$b.invokeSuspend(SysTtsListItemHelper.kt:45)
	at cc.a.resumeWith(ContinuationImpl.kt:9)
	at tc.j0.run(DispatchedTask.kt:112)
	at yc.h$a.run(LimitedDispatcher.kt:4)
	at ad.k.run(Tasks.kt:3)
	at ad.a$a.run(CoroutineScheduler.kt:96)

Same for me. Trying to fix it for the last hour. No luck so far

[papewas]

Succesfully tried this one to add in TTS-Server for extra fine tuning: https://github.com/ag2s20150909/TTS

[samsonsin]

Succesfully tried this one to add in TTS-Server for extra fine tuning: https://github.com/ag2s20150909/TTS

I got this to work, thank you!

To add more detail; install the app. Then you can add it as a "Local TTS" within TTS Server. Took a few minutes to replicate my setup with multiple voices, but it works - so far - as well as my previous setup using the voices provided by the TTS Server app.

[alfamus]

Succesfully tried this one to add in TTS-Server for extra fine tuning: https://github.com/ag2s20150909/TTS

Succesfully tried this one to add in TTS-Server for extra fine tuning: https://github.com/ag2s20150909/TTS

I got this to work, thank you!

To add more detail; install the app. Then you can add it as a "Local TTS" within TTS Server. Took a few minutes to replicate my setup with multiple voices, but it works - so far - as well as my previous setup using the voices provided by the TTS Server app.

YOU GUYS ARE AWESOME! It works now!!!!! I am so happy I can listen books and listen ai chat again! MUCH love

[alfamus]

edit: okay I enabled sentence processing fom top right setting and it works most of the time
Ah, there is an issue. It repeats same sentences twice. Do you have any idea what could be the reason.

[samsonsin]

edit: okay I enabled sentence processing fom top right setting and it works most of the time
Ah, there is an issue. It repeats same sentences twice. Do you have any idea what could be the reason.

I don't think I understand your issue? Everything works fine on my device. If you'd like help troubleshooting, you'll need to provide as much information as possible / you think is prudent.

Edit: will just show my setup to clarify. I made a new group, and added the other TTS app's TTS Engine as "Local TTS". I have many voices since I have a custom rule manager that uses regex for different voices.

[alfamus]

@samsonsin Thank you so much for the screenshots!

I think it's the Rikkahub app I'm using has the issue. It didn't have an issue before I switched to your new method but there seems to be some conflict with this app.

See here if you would like for my problem: rikkahub/rikkahub#405

[alfamus]

update: Other apps stop working after 4 sentences. The error i get is 'sent ping but didnt recieve pong in 20000ms

[alfamus]

@papewas could you please share your finetuned version

[samitooooo]

I have the same issue(403 Handshake). And when using a different tts inside of the server it stops after 5 sentences

[samsonsin]

Quite strange. Using the engine recommended by @papewas, every few sentences there's a massive delay. I got frustrated enough to just use the Google TTS via TTS Server. It's actually quite a lot better in comparison to using it directly as a TTS Engine (as I'd assume most have done and immediately dismissed). It's an obvious downgrade, but will hold me over the few days I'll need to figure out a replacement.

If anyone would like to try offline engines out, I found this page which have some models that are actually very impressive. However, the engine doesn't support multiple voices / only exports the one you've selected in its interface. Moreover, the models I'd call good have an unacceptable latency. The only models that were fast enough to use (in my limited testing) were of such low quality that I'd rather use the Google engine. That said, there are some extremely interesting voices that I could totally see myself using if latency was better. There are even voices that whisper and such, which is very neat. Though the UI makes it very hard to find these

[alfamus]

@samsonsin I switched to terrible (compared to beautiful azure) google tts too, let us know if you find a good latency better than google tts alternative please. Thank you for your link and answer. Wish you a nice week!

[AbsoluteConfig]

go.Universe$proxyerror: websocket: bad handshake: 403 Forbidden
	at tts_server_lib.EdgeApi.getEdgeAudio(Native Method)
	at com.github.jing332.tts_server_android.model.speech.tts.h.getAudio(MsTTS.kt:385)
	at com.github.jing332.tts_server_android.model.speech.tts.e.getAudioWithSystemParams(ITextToSpeechEngine.kt:23)
	at com.github.jing332.tts_server_android.model.speech.tts.e.getAudioWithSystemParams$default(ITextToSpeechEngine.kt:14)
	at u5.w0$i$b.invokeSuspend(SysTtsListItemHelper.kt:45)
	at cc.a.resumeWith(ContinuationImpl.kt:9)
	at tc.j0.run(DispatchedTask.kt:112)
	at yc.h$a.run(LimitedDispatcher.kt:4)
	at ad.k.run(Tasks.kt:3)
	at ad.a$a.run(CoroutineScheduler.kt:96)

I have relied on this for years. I feared this day would come. I guess Microsoft killed it? It didn't stop functioning when this issue was first posted nor the last two days as stated. It was 12 hours after this issue post date that it stopped working for me. That makes it seem like an intentional change on Microsoft's part.

I really can't go without this app. I have looked for countless alternatives and never found anything better. I am looking for any possible alternatives, even paid through Azure.

Succesfully tried this one to add in TTS-Server for extra fine tuning: https://github.com/ag2s20150909/TTS

I tried this along with instructions posted and it just hangs. When I enabled notifications, I figured it was getting the same 403 error as this app. The local AI voice models are both inferior and have extreme latency.

[Nyubug]

For anyone who needs an alternative I found this which seems to work for now.
https://github.com/yynag/edge-tts-android