[Feature request]a good direction

[LY131022]

In fact, live caption_ app has done a good job, but the translation accuracy is not as good as AI，and it cannot convert the translated text into speech.. When you are talking, it will keep translating and displaying the words it hears and the translation progress, instead of stopping every time.

You know, it is difficult to communicate with the other party in the real world and tell him that my app did not listen to the whole thing this time. The other party said a long paragraph, and I have to wait for the app to translate it all before I can understand what the other party said. Real-time translation is the key, right? I think this is a good direction.
In fact I tried to watch a video with your app and I had to pause the video because the app kept picking up the audio. Then the app needed time to translate and during this time it couldn't pick up the audio if someone spoke again. So, maybe, you need to start multithreading.

Otherwise, it will be difficult to achieve the purpose of daily use. Of course, I am not a technician. I don't know what difficulties will be encountered in developing a software. I just provide some personal needs as a user. Thank you again for your open source app and your efforts for it.

[niedev]

Hi, thank you for your feedback, I will respond point to point:

In fact, live caption_ app has done a good job, but the translation accuracy is not as good as AI，and it cannot convert the translated text into speech.

I don't know what you mean with live caption app, if you refer to RTranslator, the text should be converted to speech in the Walkie-Talkie and Conversation mode unless you have the indication (no TTS) when you select a language, in that case you should set a TTS that support that language, see here for more details.
For the quality of the translation I will try to embed other models with higher quality in the future.

The other party said a long paragraph, and I have to wait for the app to translate it all before I can understand what the other party said. Real-time translation is the key, right?

The Walkie-Talkie mode isn't the real time mode of RTranslator, if you want a real time translation, you should convince the other party to download the app and use the Conversation mode to connect with the other phone, connect Bluetooth headphones to your phone, and start a real time (or almost real time) conversation (plus you can change some settings in the app to reduce the time of silence required to send the message in order to have a faster translation, sacrificing some quality).

In fact I tried to watch a video with your app and I had to pause the video because the app kept picking up the audio.

From what I know it's impossible to not pick the audio of a video, the mic of you phone will pick every audio it hears, even if it comes from you speakers.

Then the app needed time to translate and during this time it couldn't pick up the audio if someone spoke again.

Yeah that's a problem I tried to solve, but I can't, because some phones will limit the use of the CPU for background apps, and RTranslator needs a lot of power to run, so when you exit from RTranslator leaving it running in the background the performance will be reduced so much that it will practically stop working until you reopen it again 😢

I hope this helps. Thanks again for the feedback.

[pioneerAlone]

Hi niedev,

Converting the translated content back to TTS output is sometimes not necessary. For example, when a user watches a video and wants the app to transcribe subtitles and translations in real-time, there is no need to output TTS. However, the current app defaults to outputting TTS, unless there is no corresponding TTS set for that language.

[niedev]

Hi, RTranslator has already the option to turn off TTS. You just need to click on the speaker button in the top right corner of the WalkieTalkie or Conversation mode.

[pioneerAlone]

Thanks, I got this feature now.