Speech to text portable device?

verilu

New Member
Joined
Nov 7, 2009
Messages
9
Reaction score
0
I know they have a software for that(Draganfly), but don't they have a portable device where they use the device close to their mouth when they're speaking and after finishing it, you can read it?

Or when someone is talking on the phone, you place the device next to the phone and look at it real time.
 
It won't work that way during different pitch low to high and gender's voice. It make words look so wrong.
 
It won't work that way during different pitch low to high and gender's voice. It make words look so wrong.

I know that youtube recently implemented a closed captioning service that translate spoken words into text real time.
 
There is a device that does that - Captel. If you are talking about a cellphone, there is none at the moment I believe, though Webcaptel is catching up very fast on the Iphone. Captels are Speech to Text telephone devices.
 
I know they have a software for that(Draganfly), but don't they have a portable device where they use the device close to their mouth when they're speaking and after finishing it, you can read it?

Or when someone is talking on the phone, you place the device next to the phone and look at it real time.

There is a free speech recognition app for the iPhone and iPod called Dragon Dictate, but it has a lot of errors. Someone with really good speech might be able to figure out how to make it work, I suppose.

Dragon Dictates makes a full-powered speech recognition software program that needs to be trained by the owner of the program. Prices have gone down a lot----I saw it on sale for $39.99 at Costco, I think. The technology is still a very long way from being able to have a small portable device recognize anyone's speech on no notice. That's why we still need to hire providers of CART (Communication Access Realtime Translation) for conferences and teleconference and why television networks have to pay a lot of money to realtime captioners to caption live programs. These services are expensive since it takes years for the providers to get really good, and relatively few people are skilled and experienced enough to do really good work.

There are a few mobile phones that will work with Web CapTel, a specialized kind of VCO relay service that is normally used by people with hearing loss who can speak for themselves. Like any relay service, it's not intended to be used for in-person communication, though. The service is a bit slow, tends to lack a great deal of punctuation, and can fail to produce even approximate text for unfamiliar names and words. It's helpful as a back-up but definitely has limitations at this time.
 
There is a device that does that - Captel. If you are talking about a cellphone, there is none at the moment I believe, though Webcaptel is catching up very fast on the Iphone. Captels are Speech to Text telephone devices.

CapTel (Interpreter) is when the agent typing for you!
 
In time, soon enough, this will one day be a common communication device that's accurate.
 
I know that youtube recently implemented a closed captioning service that translate spoken words into text real time.

I just asked Verzion about getting CC on youtube , and I was told there was no CC !
Do have you sign up this ?
 
I know they have a software for that(Draganfly), but don't they have a portable device where they use the device close to their mouth when they're speaking and after finishing it, you can read it?

Or when someone is talking on the phone, you place the device next to the phone and look at it real time.

I was thinking about an implementation last night after I realized during writing a response what kind of real-time interpreter would suit me best. Ideally, that would be something like a real-time captioning device.

Project Time!

:) I'll do some research and try to get some volunteers together and track progress on a Google Blog. Stay tuned sports fans....
 
There is a free speech recognition app for the iPhone and iPod called Dragon Dictate, but it has a lot of errors. Someone with really good speech might be able to figure out how to make it work, I suppose.

Dragon Dictates makes a full-powered speech recognition software program that needs to be trained by the owner of the program. Prices have gone down a lot----I saw it on sale for $39.99 at Costco, I think. The technology is still a very long way from being able to have a small portable device recognize anyone's speech on no notice. That's why we still need to hire providers of CART (Communication Access Realtime Translation) for conferences and teleconference and why television networks have to pay a lot of money to realtime captioners to caption live programs. These services are expensive since it takes years for the providers to get really good, and relatively few people are skilled and experienced enough to do really good work.

There are a few mobile phones that will work with Web CapTel, a specialized kind of VCO relay service that is normally used by people with hearing loss who can speak for themselves. Like any relay service, it's not intended to be used for in-person communication, though. The service is a bit slow, tends to lack a great deal of punctuation, and can fail to produce even approximate text for unfamiliar names and words. It's helpful as a back-up but definitely has limitations at this time.

There are some people who are able to use Dragon Dictate with great results. HOWEVER, speech recognition software is not ever going to replace CART providers and relay typists. The problems with speech recognition software come from the ambient sounds around the user and the variances in pronunciation. These ambient sounds interfere with the ability of the computer to recognize the words spoken into the system. Even the same user will have things mis-translate if they are not in a quiet environment. Most speech recognition software is built and functions best with ONE user voice. It "learns" that person's voice and speech patterns.

There are many engineers working on the problems with speech recognition software, but they have been working on it for years. There is only so much that they can do with the computer programs. There is no replacement for the human brain as of yet. Until they can replicate the way the human brain functions speech recognition software will be stuck as a single user application. The problems of everyone pronouncing the language with a different accent are too great to over come at this time.

I am working on a system to use CART providers over the phone for incidental communications, but have not gotten it up and running yet. When I am in the process of testing it, I will let you know!

In the mean time, insist on quality NCRA certified CART providers for your meetings, appointments, trainings, and any other time you need them.
 
There are some people who are able to use Dragon Dictate with great results. HOWEVER, speech recognition software is not ever going to replace CART providers and relay typists. The problems with speech recognition software come from the ambient sounds around the user and the variances in pronunciation. These ambient sounds interfere with the ability of the computer to recognize the words spoken into the system. Even the same user will have things mis-translate if they are not in a quiet environment. Most speech recognition software is built and functions best with ONE user voice. It "learns" that person's voice and speech patterns.

There are many engineers working on the problems with speech recognition software, but they have been working on it for years. There is only so much that they can do with the computer programs. There is no replacement for the human brain as of yet. Until they can replicate the way the human brain functions speech recognition software will be stuck as a single user application. The problems of everyone pronouncing the language with a different accent are too great to over come at this time.

I am working on a system to use CART providers over the phone for incidental communications, but have not gotten it up and running yet. When I am in the process of testing it, I will let you know!

In the mean time, insist on quality NCRA certified CART providers for your meetings, appointments, trainings, and any other time you need them.

After reading a lot last night, yeah that pretty much sums up the difficulty of getting it to work properly. Basically, the technology hasn't been "perfected" so you get captioned YouTube videos that are mistranslated, etc.

I have some ideas and gathered some resources. It'll take quite some time and money to develop an "easy" system for users, but I'm intending on developing a prototype device within 6 months.

gstreamer [CMUSphinx Wiki] seems to be a good open source package to consider.

I'll probably start a blog soon on it shortly.
 
I read something the other day about a product of SoundAssociates.com called I Caption, which is a small hand-held box to be used in theatres, and one can read the captions during the play. Currently only available for a small number of plays which makes me think that they are not run by speech recognition programs, but captioned before the curtain goes up.
 
If you want to see a lot of different technologies that can be used for deaf/hoh, check out WDW's website:
Hearing Disabilities | Plain Text | Walt Disney World Resort
I'm sure their "captioning" device(s) are pre-recorded.
I tried one at Hollywood Studios last fall and found the text was either a little ahead or a little behind "the action" (can't remember now). I did not try any of the other devices while I was there.
 
Voice to text phones

I have been using the CapTel phone for some time and it is very good. Just recently, I acquired a CaptionCall phone, which I was provided free of charge (purchase $149). It does the same thing as CapTel,but is more thoroughly computerized and has a few added features. CaptionCall was provided after I responded to an offer by E-mail and it is being paid for by the FCC, supposedly. Both systems work the same, with a 6-8 second delay for the display. If it is a Recording /Spam call, it tells you up front.
It is the only way I can communicate on the phone.
Now looking for a portable device to do the same thing fairly reliably.
 
I would love a device that could translate speech to text, something to carry with me. I keep waiting for technology to catch up. If anyone knows of any portable options, I would love to know about them (I almost said 'hear' about them).
 
I would love a device that could translate speech to text, something to carry with me. I keep waiting for technology to catch up. If anyone knows of any portable options, I would love to know about them (I almost said 'hear' about them).

There are lots of Android and iPhone apps that do this now.
 
Back
Top