Ksimond is one of the tool but i haven't used it ....you can check this on net..
Printable View
Ksimond is one of the tool but i haven't used it ....you can check this on net..
Speech-to-text is very different, and much more difficult to do, than text-to-speech, and it might be a good idea to address the two separately.
There is a lot of text to speech out there, probably because it got done sooner than speech-to-text, meaning that there are more people in the Linux development community that are dependent on text-to-speech, which means that there are more people developing speech-to-text for Linux, which means... etc. That hasn't happened with speech to text yet, but it bears mentioning because it is possible that if you build it, they will come. I am a big believer in the open source development model, in part because user-developed software is a whole lot better at meeting users' needs, but I think that in this case, somebody needs to get the snowball rolling down the hill.
The degree to which the major software distributors (most notably, Microsoft and, I think Mac too) seem to have gotten their *cough* fundaments handed to them by Nuance when they tried to do speech to text would seem to indicate that it is a heavier lift than text-to-speech, but the demand is definitely there.
Among other things, speech-to-text is, to a certain degree, irreducibly resource hungry, so an operating system that is extremely resource hungry for no good reason (e.g. Windows) would really not be my first choice to run Dragon overtop of. For example, Dragon crashed while I was editing my post. I can only speak to my personal experience, but the only reason that I haven't made the switch to Linux is Dragon's notorious finickiness.
Incidentally, has anybody had any luck developing a Mac emulator-- it seems to me like it would be easier, because Mac, the last I remember ( admittedly, that was almost 10 years ago) uses a UNIX or Linux kernel---and does Dragon for Mac work any better with that than Dragon for Windows has worked with wine?
I do have a message in to Nuance, asking them when they are going to start supporting Linux, and I will certainly keep you posted if I get a reply, but I'm not holding my breath. Nuance seems pretty cocky.
I'll have to add a "me too" here: the apparent lack of speech-to-text solutions (not limited to English) is what has ultimately held me back from Linux.
I haven't contacted Nuance because I think that:
a) It's much more likely for Wine to support Dragon than for Dragon to support Linux;
b) A free native solution would be nicer, and could control the entire system instead of just writing what I say in compatible applications.
I'm lucky enough to be able to use normal keyboards and mice, but I can't write fast due to a spinal cord injury.
Thanks for the reply. Nice to know that there are other people interested in this, because that's what we need. I also can type some (partly because I used to type really fast), but writing is what I do, so I need Dragon.
I can definitely see your point about wine supporting Dragon versus Dragon supporting Linux. I bug Nuance about it on general principles, but I don't seriously expect them to do anything.
I had a thought on Dragon and wine:
I think that those of us who use Dragon on a regular or semiregular basis might have some new information to contribute to that discussion, and that project. For example, the COM error message that people mention getting in wine has happened on every Windows box that I've ever put Dragon on too. It might also be useful to take a second look on what programs Dragon needs the dictation box for in Windows. If it will, tell me and I will keep track. Basically, I think the wine/Dragon project needs a control for comparison, because some of these issues are Dragon issues, not compatibility issues.
Windows emulators are starting to look like a very promising direction for speech to text for Linux because that's likely to happen a lot more quickly and if there is a solution in the short term, there will be a lot more speech to text dependent people in the Linux community shortly thereafter, which probably means an influx of developers who are really interested in speech to text. Has anybody had any luck with VirtualBox, other Windows emulators, or any of the Mac emulators?
However, I still would love to see someone wipe that smug "we have the market cornered so we can walk all over you" smile off of Nuance's face.
I am definitely interested in helping to develop speech to text programs for Linux and, although I am not a programmer (I took a few classes and, although my grades tell me that I have the brains, I really don't have the patience), I might be able to help, both because I do use Dragon regularly, and because I am, as my user name indicates, an anthrogeek, albeit a cultural anthropology geek rather than a linguistics geek. Being very into studying all four fields of anthropology, I do know a few things about language that might be useful. I would love to help if I can.
I also got an interesting suggestion from someone on this speech recognition community: there is apparently a Linux program called AutoHotkey, and there are a couple of windows programs (PhraseExpress and MacroExpress). They also mentioned something about programmable keyboards. They are too less resource intensive and, while they might not help me (anthropologists use words in weird ways, and not always the same ones), I'm guessing that they would be really useful for programmers-- no more typos!-- and they might be a way to make Linux development a viable option for somebody who has trouble typing, which means more people for the speech to text projects-- which in turn means that something that is more useful to my discipline is likely to happen more quickly.
Are you familiar with this?
What is the status of the Sphinx project, is it in a place to assist building speech-to-text software for linux?
What about via voice and wine?
I used to use ViaVoice on Windows (a long time ago).
It was designed to work with Windows, which means it integrated with programs such as MS Word. Clearly, this is pointless on a Linux system (unless you use MS Word, in which case why are you using Linux?).
For Linux, it would be better to have something that integrates with Linux rather than with Windows.
Having said that, I can't test ViaVoice on Wine because I've lost my ViaVoice disk.
well for better or worse I ordered this
http://voxin.oralux.net/index.php
For $6 it's worth trying but how they claim you get a via voice license with that seems.......strange.
I and my friends have been working on developing a Speech Assistant software called "VEDICS: Voice Enabled Desktop Interaction and Control System". The software is nearly complete and the user can access almost any element visible in the system. It can even access words that are not there in the English dictionary.
The software works perfectly with Ubuntu 9.10 and Ubuntu 10.04.
We are currently in the testing phase and we still have to create an installer.
It will be about a month before we release the software.
We'll soon release screencasts of it to show a demo.
Stay tuned!!!! :p