Hello
I need to prepare voice recordings for my android project. As mentioned in the tutorial I need to prepare those files in 8kHz, 16bit, mono for telephone applications. But when I convert my files to this format the sound quality decreases drammatically. In 16kHz there's no such issue but as soon as the kHz drops to 8 the voice turns into something that's not too clear.
I know that's how it sounds like when you're calling someone on the phone, but while using voice recorders of the mobile phones, it's clear enough. I feel like I should be using 16kHz sound files instead of 8kHz even though the tutorial tells the opposite.
So should I still use 8kHz speech recordings despite the low quality? Let me mention that I'm working on an android app, once again.
Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
All telephone speech is encoded with telephone codecs which only support 8khz. If you want to recognize callcenter data, 16khz is not much useful. At the same time you can still record at 16khz, then you will be able to use such data for both telephone and wideband training.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It won't be a callcenter app, it will be a reading app for the kids. So should I use 16kHz? I have nothing to do with phone calls, we're just using the smartphones for mobility. Kids will see the text and try to read it. That's when the app will recognize their speech and turn into text.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello
I need to prepare voice recordings for my android project. As mentioned in the tutorial I need to prepare those files in 8kHz, 16bit, mono for telephone applications. But when I convert my files to this format the sound quality decreases drammatically. In 16kHz there's no such issue but as soon as the kHz drops to 8 the voice turns into something that's not too clear.
I know that's how it sounds like when you're calling someone on the phone, but while using voice recorders of the mobile phones, it's clear enough. I feel like I should be using 16kHz sound files instead of 8kHz even though the tutorial tells the opposite.
So should I still use 8kHz speech recordings despite the low quality? Let me mention that I'm working on an android app, once again.
Thanks in advance.
All telephone speech is encoded with telephone codecs which only support 8khz. If you want to recognize callcenter data, 16khz is not much useful. At the same time you can still record at 16khz, then you will be able to use such data for both telephone and wideband training.
It won't be a callcenter app, it will be a reading app for the kids. So should I use 16kHz? I have nothing to do with phone calls, we're just using the smartphones for mobility. Kids will see the text and try to read it. That's when the app will recognize their speech and turn into text.
Yes
Thank you so much