I know that there has been a lot of people talking about Voice recognition, and I was wondering does anyone know when the Voice recognition is coming out? Because I just keep seeing that it will be released in scratch 3.0

There was a Voice Recognition extension available in some of the beta builds of Scratch 3, however it seems to have gone missing in later beta builds and it hasn't been in Scratch 3. I do support the introduction of voice recognition into Scratch due to there being various uses for it and to provide extra accessibility options, similar to the other controller options and the text to speech extensions.

Negative support.

Seriously, I don't fancy getting tracked on this site. If it were possible to detect and isolate specific voices this is basically asking for an invasion of privacy.

In general, conditional support only if alternatives were presented.

i agree that voice commands would help a lot though

ResExsention wrote:

Negative support.

Seriously, I don't fancy getting tracked on this site. If it were possible to detect and isolate specific voices this is basically asking for an invasion of privacy.

Nothing was said about recording other people's voice and tracking people.

I support, since voice commands would be neato burrito.

Mr_PenguinAlex wrote:

ResExsention wrote:

Negative support.

Seriously, I don't fancy getting tracked on this site. If it were possible to detect and isolate specific voices this is basically asking for an invasion of privacy.

Nothing was said about recording other people's voice and tracking people.

I support, since voice commands would be neato burrito.

Recording people's voices is basically a requirement with speech recognition. Every single speech recognition program we use today record and track voice recordings to better the system (and other motives as well). I'm not sure, but I think adding that into Scratch would be a violation of COPPA, since Scratch is sold as a children's tool (and maybe that's why it was removed). Unless we have Scratch create their own speech recognition algorhythm that doesn't record users, and overall just exists in a vaccum, I don't think this is a good idea.

Mr_PenguinAlex wrote:

ResExsention wrote:

Negative support.

Seriously, I don't fancy getting tracked on this site. If it were possible to detect and isolate specific voices this is basically asking for an invasion of privacy.

Nothing was said about recording other people's voice and tracking people.

“Voice recognition” often implies identifying a user based on their voice, such as for authenticating themselves or having personalized responses from a smart home voice assistant. The OP probably means speech-to-text

They should just not be able to record the voice. All it does is hear the voice, but the Camera sensing block is fine!. And seeing people faces are worse than hearing the voice? So I am completely unsure why it was removed. (It could have been removed due to different problems with languages.

Doughnutking06 wrote:

They should just not be able to record the voice. All it does is hear the voice, but the Camera sensing block is fine!. And seeing people faces are worse than hearing the voice? So I am completely unsure why it was removed. (It could have been removed due to different problems with languages.

The video sensing blocks can be calculated in the browser because it's just pixels moving around and its accuracy isn't very noticeable. However, speech-to-text is extremely complicated, having to deal with different accents, languages, and background noise. The technology is constantly being improved upon, yet it often does not correctly identify some words

The extension was never removed; the Scratch team just put it off to prioritize other things

Doughnutking06 wrote:

They should just not be able to record the voice. All it does is hear the voice, but the Camera sensing block is fine!. And seeing people faces are worse than hearing the voice? So I am completely unsure why it was removed. (It could have been removed due to different problems with languages.

That's different. The camera block doesn't do face recognition, it's just pixels on a screen. We can already use
raw microphone data in projects (the loudness reporter), what OP wants is recognition.

Fupicat wrote:

Doughnutking06 wrote:

They should just not be able to record the voice. All it does is hear the voice, but the Camera sensing block is fine!. And seeing people faces are worse than hearing the voice? So I am completely unsure why it was removed. (It could have been removed due to different problems with languages.

That's different. The camera block doesn't do face recognition, it's just pixels on a screen. We can already use
raw microphone data in projects (the loudness reporter), what OP wants is recognition.

Precisely. And I don't want to be recognized.

Perhaps I might tolerate voice recording that doesn't recognize specific voices, though.

Also, I don't do projects with camera sensing.

Sheep_maker wrote:

Mr_PenguinAlex wrote:

ResExsention wrote:

Negative support.

Seriously, I don't fancy getting tracked on this site. If it were possible to detect and isolate specific voices this is basically asking for an invasion of privacy.

Nothing was said about recording other people's voice and tracking people.

“Voice recognition” often implies identifying a user based on their voice, such as for authenticating themselves or having personalized responses from a smart home voice assistant. The OP probably means speech-to-text

Then why can't OP be added to scratch. And simply make it impossible for the user to store the actual voice recording, so they are forced to store it as text without accent or anything else?!

All I would like from voice recognition is if, we can pick up the player saying certain words. No need to record it! That's all I want, well the camera sensing would have been worse really. Voices sound similar. It would be cool to see a Voice Recognition come back! I honestly can't think why voices would need to be recorded.

Doughnutking06 wrote:

All I would like from voice recognition is if, we can pick up the player saying certain words. No need to record it! That's all I want, well the camera sensing would have been worse really. Voices sound similar. It would be cool to see a Voice Recognition come back! I honestly can't think why voices would need to be recorded.

Speech-to-text works by recording your voice and sending it over to a server to be recognized. Recording your voice is pretty much required. None of the current blocks that use computer features (like the camera and loudness blocks) have to record or send anything to a server.

Honestly Scratch could probably do that if they implemented Google's voice-to-text service as an extension. I assume that is what the original extension was supposed to be during the beta. If it was removed, it probably had a reason, be it cost or maybe legal concerns about recording children's voices and sending them to Google. I really wished we could get an explanation by the Scratch Team as to why the extension was removed so this suggestion could go somewhere. It was a feature a lot of people were excited for.

support

scratchastroLOL wrote:

Dupe

did anybody even read this?

If word/s [...] said :: #8ecc37 hat

<word/s [...] said? :: #8ecc37>

this is what the blocks should look like

There was a video that was demonstrating it! But it is very old, I will try to find it. It was obviously in scratch, but they must have run into a problem!