speech-translator-ext / speech-translator-readme

This repository outlines the procedures and general information for the Speech Translator project.
https://st.kappaflow.dev
47 stars 2 forks source link

About the volume level #28

Open HilbertMaximov opened 5 months ago

HilbertMaximov commented 5 months ago

Hello, I set it according to realtek's method, but only in the case of a loud volume of my speaker, the voice will be recognized and translated, and the loud volume will affect the viewing, how can I make the speech translator run even when the volume is low?

kappaflow commented 5 months ago

Hello @HilbertMaximov, First of all, please make sure you have set the audio level for "Stereo Mix" device to 100% (like in the instruction). Usually if you have done so, then the audio level is enough to recognize the speech. In case you are using external speakers with an amplifier in additional to that level settings, you might want to set a higher volume for the output device in Windows Audio settings, at the same time make it lower on the speakers/amplifier itself. It should fix your issue, in addition to that it should improve the audio quality, because of a better SNR.

Please let me know if it helped, or maybe I didn't understand correctly your setup.

HilbertMaximov commented 5 months ago

Hello @HilbertMaximov, First of all, please make sure you have set the audio level for "Stereo Mix" device to 100% (like in the instruction). Usually if you have done so, then the audio level is enough to recognize the speech. In case you are using external speakers with an amplifier in additional to that level settings, you might want to set a higher volume for the output device in Windows Audio settings, at the same time make it lower on the speakers/amplifier itself. It should fix your issue, in addition to that it should improve the audio quality, because of a better SNR.

Please let me know if it helped, or maybe I didn't understand correctly your setup.

Thanks! I used the speakers that come with my laptop and set them to 100% according to your instructions. I would like to ask if the speech translator can work normally if my volume is turned down very low under such settings? And on the sound settings page of Windows, do I need to change the input device to stereo mix? I found that there is one more input device after using stereo mix.

kappaflow commented 5 months ago

Basically, "Audio Mix" take the audio from your output (playback) devices, but "Audio Mix" is also input (recording) device, which make it possible to choose in the web browser as a device.

So because Audio Mix takes the audio from the output (playback) devices, the level set there (Windows Audio settings) affects the audio capturing for the recognition, at the same time you can safely control the audio level further along the pipeline, for example on your speakers or amplifier, or headphones themselves. Just don't set too low level for the playback/output device in Windows Audio settings.

kappaflow commented 5 months ago

Also, if you can't control the audio level "externally", then you may need some more complex solution like VoiceMeter... where you can pipe audio and adjust levels individually for speakers and for the recording device selected in the web browser.