Hallo

ich lasse für meinem schwerbehindeten Vater, der weder Arme noch Beine bewegen kann, als
Steuerung einer Umfeldsteuerung Dragon Naturally Speaking (wurde ihm von der Krankenkasse bewilligt)
laufen.
Da er aber nur mit Brille sehen kann und die beim liegen nicht auf hat (trotzdem aber das Umfeld
steuern können soll) ist es für mich extrem wichtig Dragon nicht zu jeder Zeit jedes Kommando zu
gestatten. Da er dann eh den Bildschirm nicht sieht ist ein falsches Kommando schnell ein Chaos.
Das darf in seiner Lage nicht passieren können.
Daher suche ich einen Weg z.B. immer nur die Kommandos zuzulassen die den sichtbaren Bildschirmbereich
betreffen. Dann könnte ich (ich bin des Programmierens mächtig) im Hintergrund ständig dafür sorgen
das der Vordergrund sich immer nur mit einem Thema beschäftigt (und nicht änderbar ist) und das auch
mit akustischen Ausgaben versehen so das er halt keinen Bildschirm mehr braucht.
Quasi wie ein altes MS-DOS Programm mit moderner Spracheingabe (und -ausgabe).

Die Frage ist lässt sich sowas mit Dragon realisieren oder ist das Programm schon zu sehr auf
nicht (oder wenig) behinderte Menschen zugeschnitten?

gruß
MCP

Könnten Sie bitte Ihr Anliegen etwas einfacher und klarer (und in kurzen Sätzen) beschreiben, damit unser querschnittgelähmter Programmierer und Moderator Lindsay Adam (monkey8) die automatische Google-Übersetzung ins Englische versteht?

Gruß, RW

huhu
Ich möchte mit Dragon Naturally Speaking Pro für meinen schwerbehinderten Vater folgendes erreichen:

Nur das gerade aktive (visible) Programm soll steuerbar sein
(Alles andere [z.B. Windows Desktop] soll unterdrückt werden).

Ich kann Programmieren. Gibt es ggf. einen Weg die Sprachkommandos
von Dragon "abzufangen" (vor der Ausführung) und die Ausführung zu
verhindern, wenn sie meine Bedingungen nicht erfüllen?

{ Gesteuert wird ein kleiner PC der als Umfeldsteuerung verschiedene Dinge
steuert. Dieser muss auch blind - also ohne das man den Bildschirm sieht -
steuerbar sein (ohne das ein Chaos entstehen kann). Also sind alle
Kommandos verboten für die er hingucken können müsste. }

gruß
MCP

englisch:
hi
With Dragon Naturally Speaking Pro i try to do the following for my severely disabled father:
(my father can only use his head - nothing else. No Hands,Arms...)

Just the aktive (visible) program shell be controllable
(Everything else [i.e. Windows Desktop] shell be supressed)

I am able in coding. Is there (maybe) a way to capture voice commands from
dragon and block them if they dont match my filter ?

{ My Father control things like TV,Radio,Camera and so on with a little computer.
This computer have to be controllable only with voice (without display)
and this MAY NOT result in a chaos !!! So it´s very important to block
all commands wich are not able to use in my "no look"-system }


greets
MCP

Hi MCP

Welcome to the forum and I hope it is okay for me to reply to you in English.

It is pretty much possible to do what you want but first a little explanation. Dragon Naturally Speaking Professiona comes with something like 10,000 built-in commands, they are accessible by scope, scope meaning:

Global commands (work all the time)
Application Specific (will only work with one specific application)
Window Specific (will only work with one specific application that has a specific window title)

Now if you only want the commands that are for one particular application to be active then you don't have to worry about disabling any other Application Specific or Window Specific commands and you need only concern yourself with the Global commands.

Unfortunately we are still talking about thousands but they basically breakdown into the following categories:

Natural Language Commands
Active Accessibility Commands
Customer added commands/macros/scripts which can be of six different types and are held in one specific file.
A few other various commands implemented by various technologies.

Now the good news is that we can turn off Active Accessibility commands, Natural Language Commands and we can delete absolutely everything in the customer added commands file (this commands file also contains about 25 inbuilt Dragon commands which can be deleted).

Now that should pretty much take care of most of the commands and it just leaves you needing to add all your own personal commands for the specific application you need to use.

If at any stage you find the odd stray command that is still active then effectively what you can do is nullify it to stop it causing any chaos. You do this by simply adding an application-specific command for your application with the same name as the command you want to nullify, with the exception that the command you add does absolutely nothing.

Lastly with the Professional version you can add what are called advanced scripts (a kind of Visual Basic), text and graphics commands (boilerplate text and images), macro recorder commands (where Dragon records a sequence of actions and repeats it when you run the script) and lastly step-by-step commands where Dragon just mimics a set of steps that you perform including pressing particular keys and clicking the mouse in particular places. Personally I would forget about macro recorder and step-by-step commands as they can be very slow, stick with advanced scripting and text and graphics (for the obvious signature type templates etc) commands.

It is also possible to add custom voice commands using free third-party add-on scripting engines with the Premium editions of naturally speaking, like Vocola/Dragonfly etc. Fortunately we have expertise in the forum that covers such commands as well. These free add-on scripting engines can be very powerful allowing you to create very complicated commands or very simple commands.

If you want to take it to another level then the ideal solution is a standalone application that contains commands only relevant to the application that you want to add commands for. That way you start with a blank sheet application with no commands and just add the commands that you need. However to do this with Dragon you would need to use the Dragon SDK which is prohibitively expensive.

The other option is to produce a standalone application using the Microsoft SAPI 5 SDK, which is effectively the same speech engine as used by Windows Speech Recognition (comes free with Windows Vista and Windows 7). Not quite as accurate and feature rich as Dragon but certainly capable of achieving your task. The advantage of the SAPI 5 SDK is that it is free and it can be used with any .net language (managed code) or C++ native code. There are lots of examples on MSDN etc setting up simple SAPI 5 applications containing only voice commands.

Lindsay
www.pcbyvoice.com

hi

Thank you very much for this comprehensive answer. Now i will check the 2 options.

The SAPI-coding looks like a good way (if the quality is well enough) to make my
simple, but not destroyable, command-system to reach the "no-look" commands goal.
I will try to follow this new path and learn to code it.
The remaining is the active (visible) working with the computer (i.e. writing emails,
surfing,...) could remain on Dragon. Maybe i just have the need to control the Dragon
start/stop by my program and add a user reset if something is going wrong too much.

thanks again for fast helping
MCP

There is a simple managed API available below, this will work with any .NET language (Visual Basic, C#, Delphi, JavaScript etc) and will allow you to add custom voice commands to your application.

http://msdn.microsoft.com/en-us/library/...ecognition.aspx

If you need something more in-depth (which it does not sound like you do) then you can use the SpeechLib namespace with Com interop (if you are using.net), The managed API above is basically a subset of SpeechLib.

Just one thing to point out here is that we do not freely support any speech SDK coding on this forum, although we are happy to help with basic questions. Quite frankly it is too time-consuming and it's a never ending line of questions, particularly for someone starting out in programming. However, the good news is that there are numerous Microsoft resources on the web of for the SAPI 5 SDK. Also, because you are an experienced programmer I doubt you will have too many issues.

If you are using C++ then start here:

http://msdn.microsoft.com/en-us/library/...ecognition.aspx

Best wishes,
Lindsay.
www.pcbyvoice.com

Robert the Bruce,

it is just a little unfortunate that both links that your provide are pointing to the same source.

Rüdiger

np. I will have to google now for good samples. And yes my prefered
language is C++. Target System is Windows XP (=>SAPI 5.1 not higher).

thx for the link.

greets MCP

Zitat von R.Wilke

---

Robert the Bruce,

it is just a little unfortunate that both links that your provide are pointing to the same source.

Rüdiger

---

Just testing

http://msdn.microsoft.com/en-us/library/...3(v=vs.85).aspx

Lindsay
www.pcbyvoice.com