Hey Tech Buddies! I'm Working On A Simple New App, For Deaf To Non-deaf Communication.

DanTheResearcher

New Member
Joined
Sep 10, 2016
Messages
10
Reaction score
1
Hey all!
My name is Dan. I'm working on a new app, specifically built to help deaf people communicate with non-deaf people.
I wanted to ask you, as individuals and as community. What features would you like to see?

By now, i have cross-language script that takes live-recording voice, from any language, and translating it into text. I.E, if i'm not deaf, i can speak to the phone, and it'll show you what i'm saying, in textual presentation, real time.

Using this key feature, what things you think i can built, to help you, as individuals and community? (I'm not deaf btw).

Would like to hear your suggestions, day to day struggles, etc'.

Another approach, for schools etc'. Would maybe, giving the teacher, the ability to speak to a small device, such a phone etc', while real-timing translating the teacher voice into text, and live streaming it to the students phone/computers. Is it something, as students, that might help? Or there are already other methods i'm not aware of?

Would appreciate any help!
 
Dragon Dictation and CART is what you are describing, although CART uses a live transcriptionist.

There are several deaf apps out like the text one though.
 
Dragon Dictation and CART is what you are describing, although CART uses a live transcriptionist.

There are several deaf apps out like the text one though.
Hey! Thank you for replying. I saw this apps, tho i thought they deliver poor user experience, and less focused aim,specified to the dead community. Do you use does apps? If so, can you please elaborate under what scenarios you'll be using it?
 
Hi Dan! If you can provide user experience with functionality, I'm sold. That is a harder bar than it looks, though.

Competitors I've experimented with have all had, per my experience, so-so user experience. They can be inconvenient and wieldy in different situations. Seriously, I love that you are attempting to microtarget and go for the needs of the Deaf community. Awesomely badass. :)

If you need beta testers to give you honest feedback, I'd be glad to volunteer and try it out.

As for your question, I would use it in scenarios such as the below (a small sample!):
  • Impromptu social situations
  • Formal work situations, especially meetings
  • While watching a movie that is not captioned
  • To listen to the radio
 
Dragon already exists. And no app has been able to translate text to speech accurately. It takes training, like the dragon app provides. I really wish these students would do their research first. Even on this forum, if nothing else. But look at all the websites out there that offer deaf and speech products. Sigh. This gets so old.
 
Hi Dan! If you can provide user experience with functionality, I'm sold. That is a harder bar than it looks, though.

Competitors I've experimented with have all had, per my experience, so-so user experience. They can be inconvenient and wieldy in different situations. Seriously, I love that you are attempting to microtarget and go for the needs of the Deaf community. Awesomely badass. :)

If you need beta testers to give you honest feedback, I'd be glad to volunteer and try it out.

As for your question, I would use it in scenarios such as the below (a small sample!):
  • Impromptu social situations
  • Formal work situations, especially meetings
  • While watching a movie that is not captioned
  • To listen to the radio
First of all, thank you for your kind words!
Secondly, and for our subject, this is exactly why i wanted real life scenarios, would never ever thought about this. If possible, i'd like to request as much as possible scenarios, This is the main resource for our UX design(i.e, if you are watching a movie, we should increase the size of the translated text etc'). It would be greatly appreciated if you can add more stuff to the list :)
 
Dragon already exists. And no app has been able to translate text to speech accurately. It takes training, like the dragon app provides. I really wish these students would do their research first. Even on this forum, if nothing else. But look at all the websites out there that offer deaf and speech products. Sigh. This gets so old.

Hey, Dragon uses machine learning in order to improve their algorithm. Obviously, there will always be misspelled words etc'. This is exactly where user experience takes place. Tho, if for example only 80% of the words is given correctly. It's not enough, in your perspective, to be used regularly?
 
@DanTheResearcher

If you have managed to find a tool that accurately captures spoken text and puts it into text, I'd love to test it.

But with different accents and languages this seems to be a real hassle.
 
Hey, Dragon uses machine learning in order to improve their algorithm. Obviously, there will always be misspelled words etc'. This is exactly where user experience takes place. Tho, if for example only 80% of the words is given correctly. It's not enough, in your perspective, to be used regularly?
No. What if you're calling your doctors office to schedule an appointment and an app gets that messed up? Or to pay a bill and the dollar amount gets messed uo? This is the 20 percent you're not taking into account. Right now using the relay and other sources accomplish exactly what I need. I won't settle for an 80 percent accuracy. Nobody should.
 
No. What if you're calling your doctors office to schedule an appointment and an app gets that messed up? Or to pay a bill and the dollar amount gets messed uo? This is the 20 percent you're not taking into account. Right now using the relay and other sources accomplish exactly what I need. I won't settle for an 80 percent accuracy. Nobody should.

I think i'm a bit confused. The app answers the need to translate a non-deaf person voice into text, for "lightweight" scenarios like @Muse wrote. Even tho i'm pretty sure it's above 80%, from my check, it's doing just fine when speaking clearly and even if a bit not. But again, i'm coming to solve daily lightweight scenarios. :)
 
@DanTheResearcher

If you have managed to find a tool that accurately captures spoken text and puts it into text, I'd love to test it.

But with different accents and languages this seems to be a real hassle.

It does, But it keeps learning. And as well i need to think exactly where to target the app. As for situations with "controlled" talk. Or with something not related exactly to the users, as for radio,tv. Where the tone does not necessarily meant to be loud and clear. That's why i ask you guys to give me list of scenarios where you need this kind of app :)
 
A non-deaf person still has to train their voice to an app. My hearing husband, with perfect speech, is a perfect example.
 
Just my 2 pennies , from a hearing perspective, but hey I think what I would want from a speech to text tool, goes for anyone.

It should be able to capture new voices without problems, multiple speakers, without a real learning curve.

Let's assume you meet someone for the first time, this would be one of the situations where an app like this could add value.

The need for accuracy is higher when targeting deaf/ hoh community
 
Both @Mieke and @AlleyCat , the algorithm doesn't need to "train" to a specific voice. I.E Alley, if your husband will use the app for the very first time. it'll get recognize almost perfectly. let's say that the algorithm is very good and accurate, yet, there is no reason to "test" his limits :) As far as i understand, if this tool needed. Why aren't you using a tool like google speak etc'? It might not built to serve the deaf community, but it does what it does. Would love to hear your respond :)
 
Because it's not accurate. Been there and done that. Know what you write of before you post. I've been deaf for all my life (I'm 46) and everything out there has been tried for us for speech-to-text with unsuccessful results. Good luck.
 
Last edited:
Both @Mieke and @AlleyCat , the algorithm doesn't need to "train" to a specific voice. I.E Alley, if your husband will use the app for the very first time. it'll get recognize almost perfectly. let's say that the algorithm is very good and accurate, yet, there is no reason to "test" his limits :) As far as i understand, if this tool needed. Why aren't you using a tool like google speak etc'? It might not built to serve the deaf community, but it does what it does. Would love to hear your respond :)

As Alleycat said like google speak is just not accurate. But if like you say, you have a better algorithm, please do share. I offered to test for you since I'm Flemish, so wondering about the different languages comprehensiveness and would love to help you build a tool that is better.

Not to test the limits but to let you see if your tool really does what you have in mind.
 
Because it's not accurate. Been there and done that. Know what you write of before you post. I've been deaf for all my life (I'm 46) and everything out there has been tried for us for speech-to-text with unsuccessful results. Don't belittle what we know. Good luck. I'm done here.

Didn't meant to disrespect, I'm using the forum BECAUSE of your knowledge, not otherwise. Thank you for everything, as well @Mieke . I'll be sure to test out and post some result via YouTube because the app is not ready yet(development etc')
 
I don't use those apps, I write or talk (I can use voice to speak sometimes), but I'm not sure if there's already an app that does recognize voices in real-time? This is ideal or essential especially in meeting, something like that.
 
Back
Top