Something I didn't know about the Acer — speech recognition
Posted: Wed Feb 08, 2012 7:04 pm
I'm entering this article without typing. The Acer has a built in speech recognition function as part of Windows. I have not had to touch the keyboard at all in writing this article. I am really impressed with this program, although it does have some problems. I notice it gets better as I use it, for example when I first started typing (or rather speaking), it would sometimes misunderstand a word I said. For example when I would say Acer, sometimes it would get it wrong. Often it gets things right, but sometimes it misses a little.
I really like this feature. I am amazed. When you type something in, you mix commands and text together and the computer makes its best guess at what you mean. Sometimes you issue a command and sometimes you just enter text, and based on what you said, it tries to figure out when you're issuing a command and when you're not.
You may not believe this but I've actually not had to type one single character on the keyboard in writing this entire article. I have not had to touch the keyboard at all!
This is probably going to be frightening (for everyone else) because now I am not limited by how fast I can type which is about 50 words a minute, I can now type as fast as I can speak which is about three times that fast. Maybe faster.
I was considering buying the Dragon program — I can't remember the exact name of the program it may have been something like Dragon Naturally Speaking or something like that — but I lucked out and this came with the computer. So I saved about 100 bucks.
It doesn't require any special equipment. I was able to use my cheap $6.00 plug in headset (as opposed to a USB headset). If you've seen any of my YouTube videos you've probably seen me wearing the headset.
This is actually fun even with the mistakes it makes. I'm actually learning how this stuff works. When I want to go to the end of a line I have to say "press end" and I have to be careful that I enunciate both words, or I will end up with an ampersand, instead of it going to the end of the line.
What is amazing is I fucked up while I was <s>writing</s> speaking this. On the screen was a pop-up window for McAfee antivirus, blocking some of the screen real estate, and I wanted to get rid of the pop-up window, which was irritating - to say the least (as "irritating" practically applies to everything in an antivirus program) - and one of the features of this program is a function called "Mousegrid" (and I had to actually do that separately because that word is a command word and when I say if it will activate the function) which allows you to click anywhere on the screen by selecting a region. As it turned out, I wasn't paying attention, and of the two buttons "restart now" and "restart later" I just blindly went ahead and selected "restart now" instead of "restart later".
Fortunately Firefox's restart session function allowed it to restore everything exactly where I left it, including all of this article. When I first made the mistake, I was both laughing and cursing at myself because I fucked up so badly. But I was saved because Firefox does have the restore session function, and if recovered everything, including all of this drivel.
I'm really absolutely astonished by what this program can do. This is really fun!
I have even been able to make corrections to <s>my</s> this message by telling the program to click on "Edit," then tell it to scroll to the bottom of the window, then speak wherever I want put in. There is just one small problem, if you say something that appears to be a command that may activate it. For example if you say the word edit in a sentence, and you just say it by itself, it can activate the edit menu. The most important thing to do is not to panic, to just remember to press the escape key, which will then dismiss the menu.
Again I'm using words in the way I normally use them instead of the way the program activates them, because I don't really press the escape key, I just say the word escape and it presses the escape key for me. Really slick.
Because if you panic, you can make a horrible mistake like <s>typing</s> saying "undo" which then erases everything you said, which of course is not what you want!
One thing which I found which was amazing, was when I would issue the command "show numbers" in order to find the "edit" button, it would tell me how not to activate the edit menu in Firefox, by telling me to use the command "edit post" would then click the edit button in the PHPBB software — and when I spelled out the word PHPBB, the voice recognition software actually got it right — instead of activating the edit menu in Firefox. Again, this is really slick! It even recognizes the mdash "—" symbol, except that it will not allow me to put a space before the symbol, if I say space and then I say the word for the symbol it still pops the space out. It may be that it's a standard not to put a space before that symbol, but I prefer a space before and after the mdash.
The only thing that is kind of irritating, is when you have to issue a set of commands multiple times, like telling the program to scroll down, and then it sees the scroll bar on the page, and it starts scrolling on that one. Now you have to scroll all the way through that before you can scroll the page down to get to the submit button. The can be a little bit irritating, when you just wanted to get to the bottom of the screen.
One nice feature on this is when I am looking at a message I can simply <s>type</s> — there I go again, I don't type I say — say the words "press page down" or sometimes just saying the words "page down" will work or you can say "scroll down" by itself and it will scroll down the page as if you read clicked on the scroll bar.
Again I've not had to touch the keyboard at all while typing this message. Sometimes I get so frustrated with how it did things that I grabbed the mouse to click on a certain area which was faster than telling it how to do it but I have not touched the keyboard at all to do this message.
One thing where regular editing is much easier than speaking is making corrections. It's a lot more work to simply change a word when you have to issue the whole string of commands, such as if you say the word "examine" and it thinks you meant "salmon," then you have to issue a bunch of words to make the correction in the form of "correct examine" then when the window pops up, you say the correct in word and it should find it or it may have the word. For occasional corrections it's OK but for lots of corrections it's not gonna work (it's too much work) as opposed to just making the correction.
I really like this feature. I am amazed. When you type something in, you mix commands and text together and the computer makes its best guess at what you mean. Sometimes you issue a command and sometimes you just enter text, and based on what you said, it tries to figure out when you're issuing a command and when you're not.
You may not believe this but I've actually not had to type one single character on the keyboard in writing this entire article. I have not had to touch the keyboard at all!
This is probably going to be frightening (for everyone else) because now I am not limited by how fast I can type which is about 50 words a minute, I can now type as fast as I can speak which is about three times that fast. Maybe faster.
I was considering buying the Dragon program — I can't remember the exact name of the program it may have been something like Dragon Naturally Speaking or something like that — but I lucked out and this came with the computer. So I saved about 100 bucks.
It doesn't require any special equipment. I was able to use my cheap $6.00 plug in headset (as opposed to a USB headset). If you've seen any of my YouTube videos you've probably seen me wearing the headset.
This is actually fun even with the mistakes it makes. I'm actually learning how this stuff works. When I want to go to the end of a line I have to say "press end" and I have to be careful that I enunciate both words, or I will end up with an ampersand, instead of it going to the end of the line.
What is amazing is I fucked up while I was <s>writing</s> speaking this. On the screen was a pop-up window for McAfee antivirus, blocking some of the screen real estate, and I wanted to get rid of the pop-up window, which was irritating - to say the least (as "irritating" practically applies to everything in an antivirus program) - and one of the features of this program is a function called "Mousegrid" (and I had to actually do that separately because that word is a command word and when I say if it will activate the function) which allows you to click anywhere on the screen by selecting a region. As it turned out, I wasn't paying attention, and of the two buttons "restart now" and "restart later" I just blindly went ahead and selected "restart now" instead of "restart later".
Fortunately Firefox's restart session function allowed it to restore everything exactly where I left it, including all of this article. When I first made the mistake, I was both laughing and cursing at myself because I fucked up so badly. But I was saved because Firefox does have the restore session function, and if recovered everything, including all of this drivel.
I'm really absolutely astonished by what this program can do. This is really fun!
I have even been able to make corrections to <s>my</s> this message by telling the program to click on "Edit," then tell it to scroll to the bottom of the window, then speak wherever I want put in. There is just one small problem, if you say something that appears to be a command that may activate it. For example if you say the word edit in a sentence, and you just say it by itself, it can activate the edit menu. The most important thing to do is not to panic, to just remember to press the escape key, which will then dismiss the menu.
Again I'm using words in the way I normally use them instead of the way the program activates them, because I don't really press the escape key, I just say the word escape and it presses the escape key for me. Really slick.
Because if you panic, you can make a horrible mistake like <s>typing</s> saying "undo" which then erases everything you said, which of course is not what you want!
One thing which I found which was amazing, was when I would issue the command "show numbers" in order to find the "edit" button, it would tell me how not to activate the edit menu in Firefox, by telling me to use the command "edit post" would then click the edit button in the PHPBB software — and when I spelled out the word PHPBB, the voice recognition software actually got it right — instead of activating the edit menu in Firefox. Again, this is really slick! It even recognizes the mdash "—" symbol, except that it will not allow me to put a space before the symbol, if I say space and then I say the word for the symbol it still pops the space out. It may be that it's a standard not to put a space before that symbol, but I prefer a space before and after the mdash.
The only thing that is kind of irritating, is when you have to issue a set of commands multiple times, like telling the program to scroll down, and then it sees the scroll bar on the page, and it starts scrolling on that one. Now you have to scroll all the way through that before you can scroll the page down to get to the submit button. The can be a little bit irritating, when you just wanted to get to the bottom of the screen.
One nice feature on this is when I am looking at a message I can simply <s>type</s> — there I go again, I don't type I say — say the words "press page down" or sometimes just saying the words "page down" will work or you can say "scroll down" by itself and it will scroll down the page as if you read clicked on the scroll bar.
Again I've not had to touch the keyboard at all while typing this message. Sometimes I get so frustrated with how it did things that I grabbed the mouse to click on a certain area which was faster than telling it how to do it but I have not touched the keyboard at all to do this message.
One thing where regular editing is much easier than speaking is making corrections. It's a lot more work to simply change a word when you have to issue the whole string of commands, such as if you say the word "examine" and it thinks you meant "salmon," then you have to issue a bunch of words to make the correction in the form of "correct examine" then when the window pops up, you say the correct in word and it should find it or it may have the word. For occasional corrections it's OK but for lots of corrections it's not gonna work (it's too much work) as opposed to just making the correction.