Hey,I had the same problem.Do you have a solution?Looking forward to your answer.
Hello,
Quick query regarding the Azure Unreal plugin for speech (azspeech): Is there any plan to expand its support to macOS in the near future? If so, any tentative timeline for this? And if not, are there any recommended workarounds to get it running within macOS?
Thanks in advance for any insights!
when I deploy the demo to my device ,but the blueprint had some problems.can you help me solve the problem.
I cant get speech recognition with task queue to work. aren’t there any examples or docu?
Hey! : )
Just submitted an update fixing some things, including a check to avoid this crash.
Will update this project soon!
AzSpeech v1.6.13
- Release: Release AzSpeech v1.6.13 · lucoiso/UEAzSpeech (github.com)
- Marketplace: AzSpeech - Voice and Text in Code Plugins - UE Marketplace
Changes
- Remove unnecessary references
- Adjust string conversion
- Set tasks as ready to destroy
- Force thread destruction when a task finishes
- Set runnable as unique pointer and reset it when the task is marked as ready to destroy
- Add check to avoid crashes due to null audio config
- Update SDK to v1.32.1 (What's new - Speech service - Azure AI services | Microsoft Learn)
Related Issues
- plugin crash · Issue #238 · lucoiso/UEAzSpeech · GitHub
- Azspeech crashes while running on ue · Issue #234 · lucoiso/UEAzSpeech · GitHub
- Add subsystem to manage tasks + the possibility to execute tasks in a queue · Issue #221 · lucoiso/UEAzSpeech · GitHub
- STT task sometimes causes ue editor to crash · Issue #220 · lucoiso/UEAzSpeech · GitHub
- TTS Crash · Issue #219 · lucoiso/UEAzSpeech · GitHub
- Crash after TTS sound · Issue #215 · lucoiso/UEAzSpeech · GitHub
- How should I check if the memory occupied by the audio has been released? · Issue #233 · lucoiso/UEAzSpeech · GitHub
hi, I’m a beginner in unreal and I’m trying to detect the input text from the user’s voice. I have tried to download the examples in version 5.3 from the github releases and using the git bash commands, but none of them work, I get an error that tells me that plugins are missing and if I want to rebuild, and then I get many more errors. I would like to know if you could help me and tell me how I can join the blueprints to make it work, or otherwise, help me to get the examples open. Thank you all in advance !!
I am trying to use the speech to text node on UE 5.1. When voice recognition ends, I notice a scene freeze. The phrase is recognized correctly, I have enabled logs, and before the freeze, the phrase is recognized in real-time. The freeze occurs, I believe, at the moment of ending the connection. Is it a known issue? I even tried on a blank project.
+1
I’m also having this issue, but it’s not really consistent; it seems to trigger mostly one time (first or second time) and after that it is a bit smoother. I’ve noticed in larger levels it tends to hang for a longer time.
Working in UE 5.2
I’m not certain yet as I haven’t had the chance to try it, but I read that it might depend on the type of Azure account you have. During the test, we had a standard free account, and now we are enabling a “premium” account. It seems there is a limit on the resources Azure provides if you have a free account, and it could be quite impactful. I will check in the coming days and, if necessary, provide an update here.
Hello, have you resolved this issue? Is it related to the services on Azure? I have tried the Azspeech plugin and it can be used for all versions of UE 5.3. Only the speech to text will have a freeze time of 8-10 seconds. There is no problem with other nodes. I am currently reviewing the source code of the speech to text
hi Sometimes when using the speech to text node, it may suddenly lag. How can I solve this problem? The engine I am using is 5.3 plugin version 1.6.15
Sorry for the delay. Otherwise no, the premium Azure account didn’t solve the issue. I still have a freeze.
hi I have resolved the issue of Unreal5.3 lagging
I have ported the non lagging Azspeech version applicable to Unreal 5.1 to Unreal 5.3 and supported Android packaging
How did you do it? Great!
Hi, thank you for the complement.
I have a project in 4.27 that uses this plugin and when I want to do the spech to text it works without problems. But I have updated the project to 5.1, but now whenever the spech to text is called, my app closes. From the editor it works without problems.
In the device log it gives me the following message
Operation not started: uid=10321 pkg=com.c.m(null) op=RECORD_AUDIO
If you want to use version 5.1, you need to download version 1.6.8 from GitHub. There is no lag in this version