Bringing Vista's Speech Recognition Engine to Linux via Wine
Bringing Vista's Speech Recognition Engine to Linux via Wine
I have RSI (hurty hands) and need software for recognition of continuous speech.
Linux has nothing workable. Dragon was bad even on Windows.
What's needed is a Linux speech engine, so that the linux crew can get making a good GUI.
So I'm trying to bring over the Vista Speech Engine. Here's my (failed) attempt: http://womblezone.blogspot.com/
I'm trying to bring it over with the GUI (Speech Assistant). Idea being if that works, it's proof the engine's ported correctly. Well it's not working.
regsvr32 c:/windows/system32/Speech/SpeechUX/SpeechUX.dll
This fails. with the following error:
wine: Call from 0x4b4775ab to unimplemented function USER32.dll.ChangeWindowMessageFilter, aborting
I'm stymmied. I looked up this function on MSDN and it looks like a whole subsystem that isn't implemented. Tomorrow I will try compiling a 'hello world' style speechapp from the WindowsSDK and see if I can port that. But I think this venture is beyond me.
Please please please lend a hand! This is such valuable work! I want to use Linux _so_ much, but I will cripple myself if I continue typing like this.
Sam
Linux has nothing workable. Dragon was bad even on Windows.
What's needed is a Linux speech engine, so that the linux crew can get making a good GUI.
So I'm trying to bring over the Vista Speech Engine. Here's my (failed) attempt: http://womblezone.blogspot.com/
I'm trying to bring it over with the GUI (Speech Assistant). Idea being if that works, it's proof the engine's ported correctly. Well it's not working.
regsvr32 c:/windows/system32/Speech/SpeechUX/SpeechUX.dll
This fails. with the following error:
wine: Call from 0x4b4775ab to unimplemented function USER32.dll.ChangeWindowMessageFilter, aborting
I'm stymmied. I looked up this function on MSDN and it looks like a whole subsystem that isn't implemented. Tomorrow I will try compiling a 'hello world' style speechapp from the WindowsSDK and see if I can port that. But I think this venture is beyond me.
Please please please lend a hand! This is such valuable work! I want to use Linux _so_ much, but I will cripple myself if I continue typing like this.
Sam
Bringing Vista's Speech Recognition Engine to Linux via Wine
On Mon, Nov 3, 2008 at 10:30 AM, Sam <[email protected]> wrote:
--
-Austin
File a bug, that function is unimplemented.I have RSI (hurty hands) and need software for recognition of continuous speech.
Linux has nothing workable. Dragon was bad even on Windows.
What's needed is a Linux speech engine, so that the linux crew can get making a good GUI.
So I'm trying to bring over the Vista Speech Engine. Here's my (failed) attempt: http://womblezone.blogspot.com/
I'm trying to bring it over with the GUI (Speech Assistant). Idea being if that works, it's proof the engine's ported correctly. Well it's not working.
regsvr32 c:/windows/system32/Speech/SpeechUX/SpeechUX.dll
This fails. with the following error:
wine: Call from 0x4b4775ab to unimplemented function USER32.dll.ChangeWindowMessageFilter, aborting
--
-Austin
Re: Bringing Vista's Speech Recognition Engine to Linux via
This won't work. It relies on lots of unimplemented stuff. And you'll have to have vista to use it. So what's the point?Sam wrote:So I'm trying to bring over the Vista Speech Engine.
Use Dragon it's working on Wine even better then on windoze.
speech recog
I sure hope you can get this to work.
I (with MS ie. PPMS) am having more difficulty using keyboard.
Thx
ellj
I (with MS ie. PPMS) am having more difficulty using keyboard.
Thx
ellj
Bringing Vista's Speech Recognition Engine to Linux via Wine
On Mon, Nov 3, 2008 at 1:57 PM, vitamin <[email protected]> wrote:
since the functions are unimplemented, but that doesn't mean it will
never work.
--
-Austin
If it works on Vista, it should work on Wine. It may take a while,Sam wrote:This won't work. It relies on lots of unimplemented stuff. And you'll have to have vista to use it. So what's the point?So I'm trying to bring over the Vista Speech Engine.
since the functions are unimplemented, but that doesn't mean it will
never work.
--
-Austin
This thread got split. I'll post my answer here also.
There's good reason why I'm doing this. It's explained if you follow the link in the original post. I'll summarize again:
My main concern is as a developer not a user.
My objective is a FOSS GUI running on a FOSS SR-engine. Linux needs this. Badly.
I can see a path to accomplishing this:
1) pull over a closed source engine
2) create the FOSS GUI for Linux that implements (1), and sends speech-data back to a central database (such as VoxForge)
3) Once this database is suitably populated, we can crank up Sphinx, and we will have everything FOSS.
I feel strongly the Linux community needs a raw SAPI-compliant Speech-Recognition Engine, so it can build a FOSS GUI. This is the reason I was looking at porting the Vista engine. I don't know whether the Dragon port exposes a raw API. I'll contact the maintainer.
Vitamin says 'what's the point?
1. Many people have copies of Vista. It comes with nearly every new PC. Many linux users will have a Vista DVD. The same is not true for Dragon.
2. The MS engine is SAPI compliant and exposes a complete interface. I don't know about the Dragon interface.
3. The MS engine is very good. I will be surprised if Dragon is as good. Last time I used Dragon it was far behind.
Does anyone have the energy to take this project on? I'll file a bug and keep chugging away. But I am a tortoise. It would be cooler than cool to get some dev-fusion on this!
Sam
There's good reason why I'm doing this. It's explained if you follow the link in the original post. I'll summarize again:
My main concern is as a developer not a user.
My objective is a FOSS GUI running on a FOSS SR-engine. Linux needs this. Badly.
I can see a path to accomplishing this:
1) pull over a closed source engine
2) create the FOSS GUI for Linux that implements (1), and sends speech-data back to a central database (such as VoxForge)
3) Once this database is suitably populated, we can crank up Sphinx, and we will have everything FOSS.
I feel strongly the Linux community needs a raw SAPI-compliant Speech-Recognition Engine, so it can build a FOSS GUI. This is the reason I was looking at porting the Vista engine. I don't know whether the Dragon port exposes a raw API. I'll contact the maintainer.
Vitamin says 'what's the point?
1. Many people have copies of Vista. It comes with nearly every new PC. Many linux users will have a Vista DVD. The same is not true for Dragon.
2. The MS engine is SAPI compliant and exposes a complete interface. I don't know about the Dragon interface.
3. The MS engine is very good. I will be surprised if Dragon is as good. Last time I used Dragon it was far behind.
Does anyone have the energy to take this project on? I'll file a bug and keep chugging away. But I am a tortoise. It would be cooler than cool to get some dev-fusion on this!
Sam
Bringing Vista's Speech Recognition Engine to Linux via Wine
On Tue, Nov 4, 2008 at 12:10 AM, Sam <[email protected]> wrote:
--
-Austin
This is unrelated to Wine. Please take such discussions elsewhere.This thread got split. I'll post my answer here also.
There's good reason why I'm doing this. It's explained if you follow the link in the original post. I'll summarize again:
My main concern is as a developer not a user.
My objective is a FOSS GUI running on a FOSS SR-engine. Linux needs this. Badly.
I can see a path to accomplishing this:
1) pull over a closed source engine
2) create the FOSS GUI for Linux that implements (1), and sends speech-data back to a central database (such as VoxForge)
3) Once this database is suitably populated, we can crank up Sphinx, and we will have everything FOSS.
I feel strongly the Linux community needs a raw SAPI-compliant Speech-Recognition Engine, so it can build a FOSS GUI. This is the reason I was looking at porting the Vista engine. I don't know whether the Dragon port exposes a raw API. I'll contact the maintainer.
Vitamin says 'what's the point?
1. Many people have copies of Vista. It comes with nearly every new PC. Many linux users will have a Vista DVD. The same is not true for Dragon.
2. The MS engine is SAPI compliant and exposes a complete interface. I don't know about the Dragon interface.
3. The MS engine is very good. I will be surprised if Dragon is as good. Last time I used Dragon it was far behind.
Does anyone have the energy to take this project on? I'll file a bug and keep chugging away. But I am a tortoise. It would be cooler than cool to get some dev-fusion on this!
Sam
--
-Austin
This is called stealing.Sam wrote:I can see a path to accomplishing this:
1) pull over a closed source engine
2) create the FOSS GUI for Linux that implements (1), and sends speech-data back to a central database (such as VoxForge)
3) Once this database is suitably populated, we can crank up Sphinx, and we will have everything FOSS.
You are wrong. Many people don't have that junk and don't want it to begin with.Sam wrote:1. Many people have copies of Vista. It comes with nearly every new PC. Many linux users will have a Vista DVD. The same is not true for Dragon.
Bringing Vista's Speech Recognition Engine to Linux via Wine
2008/11/4 vitamin <[email protected]>:
people have produced and gathered. Please explain your off-hand
one-liner.
- d.
Sam wrote:
I can see a path to accomplishing this:
1) pull over a closed source engine
2) create the FOSS GUI for Linux that implements (1), and sends speech-data back to a central database (such as VoxForge)
3) Once this database is suitably populated, we can crank up Sphinx, and we will have everything FOSS.
How on earth is that stealing? Microsoft doesn't own the speech dataThis is called stealing.
people have produced and gathered. Please explain your off-hand
one-liner.
- d.
Reply to Austin, Vitamin, David, Elij
Austin: The project I'm proposing is inseparable from Wine. Without using Wine it is not possible. Wine is a vital ingredient. Hence it is relevant.
The nub is this: I would like some community focus on getting Vista's Speech Engine ported via Wine. It is likely going to be too difficult for myself. To do this, I need to justify the project, explain the motivation, that it is something beyond myself, of community benefit (note that already one MS sufferer has replied on the thread). Otherwise who is going to work on it?
Specifically I have been called upon by Vitamin (whom I suspect to be trolling) to justify the project: In order to do so, I need to zoom out.
Also please, as a matter of nettiquette, do not paste my entire post simply to make such a comment. It clogs the thread. Please, let us have everything said once and once only, for clarity's sake!
Vitamin, correct me if I am wrong, but if I own a legitimate copy of Vista, it is not stealing. I specifically stated in my post that a large number of Linux users will posess legitimate copies of Vista.
So why are you saying it is stealing? Pray enlighten me.
David: Probably not a good counter to argue. MS almost certainly do own that data. I'm considering the cases (such as my own) where Linux users posess a legitimate copy of Vista).
Elij: I hope it comes to pass. Such work is in the domain of the Samauri that wield the sword of tech at the cutting edge. Let's hope one of them picks up...
Sam
The nub is this: I would like some community focus on getting Vista's Speech Engine ported via Wine. It is likely going to be too difficult for myself. To do this, I need to justify the project, explain the motivation, that it is something beyond myself, of community benefit (note that already one MS sufferer has replied on the thread). Otherwise who is going to work on it?
Specifically I have been called upon by Vitamin (whom I suspect to be trolling) to justify the project: In order to do so, I need to zoom out.
Also please, as a matter of nettiquette, do not paste my entire post simply to make such a comment. It clogs the thread. Please, let us have everything said once and once only, for clarity's sake!
Vitamin, correct me if I am wrong, but if I own a legitimate copy of Vista, it is not stealing. I specifically stated in my post that a large number of Linux users will posess legitimate copies of Vista.
So why are you saying it is stealing? Pray enlighten me.
Your statement is true. And mine is also true. They don't conflict. So why do you say I am wrong? Please pay me the respect of using your brain before you wear out my fingers and patience with your facile one-liners. If you don't do this, I will cease to pay you the respect of a reply.You are wrong. Many people don't have that junk and don't want it to begin with.Sam wrote:
1. Many people have copies of Vista. It comes with nearly every new PC. Many linux users will have a Vista DVD. The same is not true for Dragon.
David: Probably not a good counter to argue. MS almost certainly do own that data. I'm considering the cases (such as my own) where Linux users posess a legitimate copy of Vista).
Elij: I hope it comes to pass. Such work is in the domain of the Samauri that wield the sword of tech at the cutting edge. Let's hope one of them picks up...
Sam
Bringing Vista's Speech Recognition Engine to Linux via Wine
On Tue, Nov 4, 2008 at 10:58 AM, Sam <[email protected]> wrote:
related to general Wine use or development, as these forums are.
of the original post, for those not on the forum.
--
-Austin
While somewhat related to Wine, it should be a separate project, notAustin: The project I'm proposing is inseparable from Wine. Without using Wine it is not possible. Wine is a vital ingredient. Hence it is relevant.
related to general Wine use or development, as these forums are.
The forums are linked to the mailing list. Quoting keeps the contextThe nub is this: I would like some community focus on getting Vista's Speech Engine ported via Wine. It is likely going to be too difficult for myself. To do this, I need to justify the project, explain the motivation, that it is something beyond myself, of community benefit (note that already one MS sufferer has replied on the thread). Otherwise who is going to work on it?
Specifically I have been called upon by Vitamin (whom I suspect to be trolling) to justify the project: In order to do so, I need to zoom out.
Also please, as a matter of nettiquette, do not paste my entire post simply to make such a comment. It clogs the thread. Please, let us have everything said once and once only, for clarity's sake!
of the original post, for those not on the forum.
--
-Austin
Re: Bringing Vista's Speech Recognition Engine to Linux via
Who told you they don't own it?David Gerard wrote:How on earth is that stealing? Microsoft doesn't own the speech data
people have produced and gathered. Please explain your off-hand
one-liner.
If it came as part of Vista, they own it not you not me not anyone else who has Vista.
Re: Reply to Austin, Vitamin, David, Elij
Wine is made to replace windows not depend on it.Sam wrote:Vitamin, correct me if I am wrong, but if I own a legitimate copy of Vista, it is not stealing. I specifically stated in my post that a large number of Linux users will posess legitimate copies of Vista.
You are proposing to add something into Wine that can not be legally used by most people. Also read the fine print on your "vista dvd". Just having it doesn't mean you can use it (or parts of it) on any PC you own.
Re: Final response to Vitamin
You won't achieve much with ad hominem attacks against Wine developers. Vitamin has valid points. Please try to be civil.Sam wrote:What is your mission Vitamin? Are you plain stupid? Bored? Are you trying to pick an argument?
Wine is NOT meant to replace Windows. Wine is designed to allow applications designed for Windows to work in a Linux environment. This includes applications bundled with Windows, such as MineSweeper. Oh, and speech assistant.Wine is made to replace windows not depend on it.
It seems you can't string together more than a couple of words without making a nonsense. Yet I notice you have over a thousand posts to your name. Have you considered STFU?You are proposing to add something into Wine that can not be legally used by most people.
For a start, I am not proposing adding anything into Wine. I am proposing people with the know-how fix problems in Wine that currently prevent a certain Windows application from running in Wine.
Now look in the AppDB. Pretty much EVERYTHING in that DB can only be legally used by someone who posesses a legitimate copy of the application. So (to put it in a way you may understand) a WoW user can only use WoW in Wine if they have a legit copy of WoW. My proposition requires the user to have a legit copy of Vista. More people own a legitimate copy of Vista than pretty much any app in the AppDB. You're an idiot.
Why are you bringing in the idea of multiple PCs? Who said anything about more than one system?Also read the fine print on your "vista dvd". Just having it doesn't mean you can use it (or parts of it) on any PC you own.
What is your intention Vitamin? Why are you here? Why are you writing 100% garbage? How old are you? Please go and watch the Simpsons or something, and let the grown ups get on with the work. You are obstructing, dirtying and obfuscating this thread (and probably countless others).
These questions are rhetorical, for I am disengaging at this point - else I will be crossing the line between exposing a troll and feeding one.
Sam
If you want to file a bug report, it would probably be in your interest to find some 3rd party software, ie. that does not ship with windows that relies on sapi 5.3.
But also, as has been suggested, try out some of the existing solutions, like dns if your goal is just to have something working rather than developing free software for compatibility with windows.
Bringing Vista's Speech Recognition Engine to Linux via Wine
Hi Sam,
I'm sorry that I cannot participate by anything valuable than some words,
I'm total noob in windows, so I cannot judge anything based on this technology,
while i can classify myself as an average *IX expert . However, I think
even the words can be of a litlle help here.
to bring us users the best product possible. However, he doesn't like to
explain by many words things, which can be summarized by one simple sentence.
And his wording may seem a bit offending to somebody, maybe because he is not
a native English speaker, but also because he has not so much time to search
for every word he writes .
Let me try to explain, what he is trying to explain to you or the others:
1) Windows OS, including Vista, is always a complex package of various
components, layers, parts etc., which cannot be cleanly separated. Microsoft
has been sued in the past for this "deep integration", especially that MSIE
cannot be separated from the OS and sold separately. It means, that any windows
application is highly depending on various other parts of the whole OS.
2) Some parts of the OS are encumbered with various patents, which are giving
Microsoft a guarantee that nobody can reuse these parts/interfaces/methods
without a special permission from them. And wine cannot ask for these.
3) If you read your Vista EULA, you will definitely find a clause saying that
you cannot use any part of it out of the original OS. This is related not only
to programs, but as well as to other, otherwise freely reusable components,
like fonts, icons, wallpapers etc. For example even I am violating this,
because I have installed a new family of Vista True Type fonts, which can be
freely downloaded from the net, onto my Linux system. So, by running anything
coming from Windows out of the windows OS, you are violating your Vista EULA.
4) The whole Wine project is under the focus of Microsoft and another big
software companies, they are very closely watching every progress and
verifying, that anything cannot be used against Wine. For example, there is a
very strict policy that anybody who have seen anything from windows source
code, cannot contribute to the wine sources anymore. And because Microsoft
now opens its sources to more and more people (even in our university the
students are working on it), it's more and more hard to find a "non-tainted"
developers for the Wine project. So, it's really NECESSARY for the Wine
project to stay absolutely clean, because otherwise it would be very dangerous
for it.
Please, think as many times as you like, about these things, and try to
re-evaluate your wishes, it's REALLY better to stay out of Microsoft stuff
for any Open Source project!
With regards,
Pavel Troller
I'm sorry that I cannot participate by anything valuable than some words,
I'm total noob in windows, so I cannot judge anything based on this technology,
while i can classify myself as an average *IX expert . However, I think
even the words can be of a litlle help here.
Vitaliy is neither of this. He is a very good wine developer, which works hardWhat is your mission Vitamin? Are you plain stupid? Bored? Are you trying to pick an argument?
to bring us users the best product possible. However, he doesn't like to
explain by many words things, which can be summarized by one simple sentence.
And his wording may seem a bit offending to somebody, maybe because he is not
a native English speaker, but also because he has not so much time to search
for every word he writes .
Let me try to explain, what he is trying to explain to you or the others:
1) Windows OS, including Vista, is always a complex package of various
components, layers, parts etc., which cannot be cleanly separated. Microsoft
has been sued in the past for this "deep integration", especially that MSIE
cannot be separated from the OS and sold separately. It means, that any windows
application is highly depending on various other parts of the whole OS.
2) Some parts of the OS are encumbered with various patents, which are giving
Microsoft a guarantee that nobody can reuse these parts/interfaces/methods
without a special permission from them. And wine cannot ask for these.
3) If you read your Vista EULA, you will definitely find a clause saying that
you cannot use any part of it out of the original OS. This is related not only
to programs, but as well as to other, otherwise freely reusable components,
like fonts, icons, wallpapers etc. For example even I am violating this,
because I have installed a new family of Vista True Type fonts, which can be
freely downloaded from the net, onto my Linux system. So, by running anything
coming from Windows out of the windows OS, you are violating your Vista EULA.
4) The whole Wine project is under the focus of Microsoft and another big
software companies, they are very closely watching every progress and
verifying, that anything cannot be used against Wine. For example, there is a
very strict policy that anybody who have seen anything from windows source
code, cannot contribute to the wine sources anymore. And because Microsoft
now opens its sources to more and more people (even in our university the
students are working on it), it's more and more hard to find a "non-tainted"
developers for the Wine project. So, it's really NECESSARY for the Wine
project to stay absolutely clean, because otherwise it would be very dangerous
for it.
Please, think as many times as you like, about these things, and try to
re-evaluate your wishes, it's REALLY better to stay out of Microsoft stuff
for any Open Source project!
With regards,
Pavel Troller
Bringing Vista's Speech Recognition Engine to Linux via Wine
Sam wrote:
It is Micro$oft, not anyone of us. Take it easy, think of Stallman, how
he build the FSF? No one helped him at that time.
I'm so sorry about your situation. There is one you can pour your fury.And what response do I get? Everywhere it is the same. 'We don't do this. This is not our department. Take it somewhere else. It is not going to work.' Or, to paraphrase, fuck off please.
It is Micro$oft, not anyone of us. Take it easy, think of Stallman, how
he build the FSF? No one helped him at that time.
Re: Bringing Vista's Speech Recognition Engine to Linux via
This topic is closed. Don't bother posting here.