Speech Recognition (SR)

* NOTE

Most of the material below is now out of date. It in now Faaaaaar easier to use speech recognition on a phone or use Google Docs speech recognition via the internet on a computer.

Watch this video:

http://youtu.be/XIBWYH1Rekw

Below here is out of date material

Rather than reading all this material you can watch a recording of were it was all presented by clicking Here :Webinar Recording

During the Webinar that was presented on the 8 March a number of question where asked in the Chat that I did not answer. The answers are Here.

See also a list of items that people said they used typing for, Here

This is reference material for the Broadband for Seniors (BFS) Webinar on Speech Recognition that was/will be presented on the 8 March 2013.

Click here for information about using speech recognition in Vista

These notes are not designed to be printed out because they use a lot of hyper-linking but ther is nothing stopping you from printing them.

You can use these notes any way you like as I have released them as copyleft and they are covered by the Creative Commons ShareAlike licence. That basically means to you can copy distribute and virtually use this material any way you want as long as you credit me (Geoff Greig) as the author of it. See: http://creativecommons.org/licenses/by-sa/3.0/

You are free to use any of this material in your BFS sessions. Please do so. That is what they were designed for.

You can contact me at any time at Email Address: geoffreykgreig@gmail.com

See also Speech Recognition Exercises

Before we start. I see a lot of people, I teach, trying to find keys on the keyboard. It can take many hours to learn where they all are. Or many hours can be spent learning to touch type (explained later) I feel it could be quicker and more efficient for those people to learn how to use speech recognition. Hopefully I can prove so during this session.

If you want to review this material after the webinar you can access it at the web page show in the address bar. Or via this link on the BFS Wiki site ie http://bfsv.wikispaces.com/BfS+resources Its right at the bottom of that page. On the completion of this Webinar I will also put a link to the recording of this Webinar at this point on this page.

Webinar Recording

Copy Address bar to Chat

If you don't know how to copy the address I have just put in the chat then write down this:

geoff.greig.net.au

That will take you to my homepage and the voice recognition link is near the bottom.

If you review this material in the future you can go directly to any part of it by clicking on that part in the table of contents to the right.

At the end of this session I am willing to stay in this room for a while to help anyone in using Speech Recognition. Or as I have my own trial Blackboard Collaborative Room, if you email me after the session I am wiling, at prearranged times, to help you with any problem you may have. With this Blackboard Collaborative with your permission it can be made so I can see what you are doing on your screen.

Any questions about what we will be covering?

Teaching method

I believe I am here to help facilitate your learning rather than to teach you anything. My teaching methodology is to show you how to learn computer functions, in this case speech recognition, rather than teach you those functions. In that way, in the future, you may be able to learn yourself without a teacher. Today there is insufficient time to teach you everything for about speech recognition, so my intention is to show you how to learn about speech recognition.

These notes are designed to be used both during this session and after it. However there would be insufficient time for each of you to try out the windows 7 speech recognition during the session. Unless I asked otherwise please just watch me demonstrating speech recognition rather than doing it yourself.

Selling it to you

In a previous life I used to work for Olivetti and at one stage sold typewriters. One of my objectives today is for you to start using speech recognition on an ongoing basis and to promote it to your students

All this material is based on what I no. I dont no everything. And even some of what I do no may be wrong. And I make mistakes. And because of the speed at which technology changes I may make statements that may be correct now but may not be in the future.

How many of you have seen speech recognition used before?

How many of you have used speech recognition?

To be able to do the Internet version of speech recognition, you need the Google Chrome web browser installed on your computer.

Who has Google Chrome installed on ther computer?

You can get Google Chrome from here

NOTE The demonstration below will not work with Internet Explorer or Mozilla Firefox. It only works with Google Chrome. However for the Windows 7 speech recognition (shown later) you do not need a browser at all.

Now within Google Chrome go to the the normal google home page for Australia ie http://www.google.com.au/ At the Bottom right hand corner you will see Google.com. Click on it. You will notice that you will now be at the Google home page in the USA. Notice the microphone at the right of the search bar. Click on it and say “speech recognition” What happened? The words you said should have appeared in the search bar. If they did not try it again. Notice the result of your search was returned after you press Enter or click on the search Icon. If you want to no more about Google Speech recognition you can right click on the microphone icon and select Voice Recognition Options.

NOTE: If you select the search option at the top of the Google home page, the microphone will appear as part of the search box. If you then exit Google Chrome, the next time you start it, the microphone should appear. Thanks to Dianna Ottaway for that hint

Go to this web site http://ctrlq.org/dictation/. Click on the photo of the microphone that is on the right. Say something when you are asked to do so.

What happened? What you said should have been typed on the screen. This is part of what speech recognition is about.

This was using SR on the internet

In both the above demonstrations the speech recognition was not performed by your computer. It was performed by cloud computing. That is your voice was recorded by your computer and sent over the internet to a farm of Googles computers (many of them) that converted your voice to text and then that text was sent back to your computer. It shows what the power of clouding computing is. Spreading task over multiple computers.

A demonstration of what you can do with windows seven speech recognition

Operating of a devise via speaking. eg. pressing of computer keyboard and or using a mouse via speaking. Typing by talking.

http://en.wikipedia.org/wiki/Speech_recognition

Pressing Keys on a machine so that text is produced

Originally done by a Typewriter. A Typewriter was the profession of a person that used a machine to produce type. Later the machine was given that name. The name of the profession was later changed to a Typist. Their are 2 types of typing. Touch typing, http://en.wikipedia.org/wiki/Touch_typing where the Typist does NOT look at the keys they are pressing and uses 8 fingers and one or two thumbs, and a hunt and peck typing, where the operator looks to find the keys to press then press those keys, usually using only 2 fingers.

Touch typing had to be learnt using a typewriter and took weeks or months to do so. Most likely millions of people where trained in touch typing. Hunt and Peck users where never taught but just picked it up. I recon ther are more people that use hunt and peck than can touch type

A indication of who can can touch type?

For those that can touch type, what is your typing speed?

http://en.wikipedia.org/wiki/Words_per_minute

I am a 2 finger hunt and peck typist and can type at about 15 Words Per Minute (WPM) with 90% accuracy or at 250 Words Per Minute with 1% accuracy. Press a lot of keys very quickly and fluke hitting some correct ones. In a recent test I discovered if I make lots of mistakes with two fingers I can type of about 20 words per minute. With the exactly the same amount of mistakes with voice recognition I can do over 100 words per minute. A link it to the website I used to do the test can be found hear.

Typewriters went from mechanical to electric to electronic and then where taken over by dedicated Word processors that where in turn taken over by Computers.

As this happened because text appeared, via electronics, on a screen rather than paper, tying speed increased such that Touch Typists got Repetitive Strain Injury (RSI) because the human body was never meant to repeat this function so often, so fast.

During the development of typing came shorthand and stenography. A different hand writing method that allowed a trained person to hand write almost as fast as people could talk. But most people could not read shorthand and, at times, one persons shorthand could not be read by another person, even though they where trained in the same method of shorthand. Typically Stenographer/Secretaries where employed to take shorthand and then type it up.

With the introduction of email the prominence of secretaries diminished as professional people where expected to type ther own emails.

An indication of what most of your typing is?

Anyone no why the keys on a keyboard are in the order they are in rather than in alphabetical order?

Answer http://en.wikipedia.org/wiki/QWERTY#History_and_purposes

See also http://en.wikipedia.org/wiki/QWERTY#Alternatives_to_QWERTY

Now world wide ther are billions of people using keyboards layouts that where never designed to be used used as they are, and those people where never trained to use keyboards.

Anyone that can not touch type can benefit from Speech Recognition.

An article on Is Touch-Typing an Obsolete Skill, with many comments

Its a wonder Speech recognition has not become more popular. See comment re Starting to use it in Windows 7

Copy typing. This is where what you type is already written and a person types what is already written.

If copy typing from typed text, ther is no advantage in using speech recognition. Better off cutting and Pasting or using Optical Character Recognition (OCR). If from hand written material ther could be advantages in speech recognition.

Creative typing. Have an idea of what you want to type in your mind and then type it. This is ideally suited to Speech recognition.

Does anyone do copy typing?

The fastest way to input creative written text

It can be faster to learn to use speech recognition than to learn how to touch type

You can concentrate on what you are writing and not which keys to press and how to spell

It is very good for getting ideas down if you think faster than you can type

In some cases it can be faster than using a mouse

It allow people with disabilities to use a computer where they may not otherwise be able to

It has to be learnt by the user (but so does typing)

It is not 100% accurate (but either is typing)

It can be frustrating to use (but all computer use can be)

You may have to put on a head set

You may not be in the best situation to be able to use it. Like in a noisy place or you may annoy other people when using it.

To control what otherwise would require physical movement. eg Dialing a number on a phone or responding to voice prompts

Dictation

Interacting with a device.

In the Windows XP era was only supplied as Part of Microsoft Office

Was supplied as part of Windows Vista

Was supplied as part of Windows 7

Is supplied as part of Windows 8

Click here for information about using speech recognition in Vista

It is not easily found in “All Programs” because it is many levels down . You can get to it this way via: Start, All Programs, Accessories, Ease of Access, Windows Speech Recognition. It is easier to find in control panel . You can make a shortcut to it that will appear on the desktop See Creating a short cut on the desktop. It can also be set to start when your computer starts.

To start Windows Speech Recognition. Start, Control Panel, Ease of Access, Speech Recognition, Start Speech Recognition (That is why you may never have seen it)

A dialog should appear on the screen that shows a Microphone icon, and to the right of that should be to word “Off” So although speech recognition is started it is not operating. To get it to operated click in the microphone and the “off” message should change to “Listening”

SR can also be started via the speech reference card. See about it later.

See starting speech recognition exercises

Click the X at the top right of the speech recognition Dialog. Or right click anywhere on the speech recognition Dialog and then click Exit

A computer with Windows 7 installed

A working Microphone. Speech Recognition has an option to set up a microphone which will indicate if the microphone is working. Although any computer microphone will work it is best to use a headset microphone.

Preferably a quiet room

Optionally working speakers, as speech recognition gives audible feedback sounds.

NOTE It is not necessary with the Windows 7 version of speech recognition to train it before using it. It trains its self as you use it. Although you can train it, I have not found it to be of any great benefit.

The best way to lean to use Windows 7 speech recognition is use the “Take Speech Tutorial”

This can be access via Start, Control Panel, Ease of Access, Speech Recognition, Take Speech Tutorial.

Or if you have the Speech Recognition dialog open, right click on it and then click on Take Speech Tutorial

It is worth while going through the speech tutorial multiple times. The fist time just to see all the things that speech recognition can do the second time pick up the finer points of how to do it and possible a third time to note what you have to say to perform certain functions. Although ther is a reference card.

As I will be talking to my computer during this tutorial I will not be able to respond to you during this time. So please wait until I finish the first sections of the tutorial before asking questions. It would take quite some time for me to allow you to read each screen of the tutorial and as we dont have that much time I will move through the screen quite quickly. I will say “Next” and the SR will move to the next page. I will not use the keyboard or mouse at all.

A demonstration of using the first few sections of the Speech Tutorial

Any Questions about the Speech Tutorial

See Going through the speech recognition tutorial exercises

The second best way to learn speech recognition is to use the speech reference card. This can be accessed via Start, Control Panel, Ease of Access, Speech Recognition, Open Speech Reference Card.

Then either print the parts of the reference card out that you want to try or have its window open at the same time that you try each function

Anyone not no how to have multiple windows open on the screen at the same time, re size those windows and switch between then?

A short demonstration of some of the functions on the speech reference card

Any questions

Anything specific anyone want to see demonstrated?

See using this speech recognition reference card exercises

Most everything that you can do with the mouse and keyboard with the exception of what is in the next section

Easily drag the mouse

Have someone other than me speak in this Webinar and have ther voice converted to text

Have a recorded sound file converted to text. You could have it play to a microphone

Prior to Windows 7 if you wanted to use a non Microsoft program with the dictation part of speech recognition you had to create the text in a Microsoft program like Notepad or Wordpad and copy and past the text in to the non windows program like Google Chrome or Open Office Write. However with the Windows 7 version of speech recognition a dictation scratchpad will appear so that all the dictation and editing functions of speech recognition are done in the scratchpad and the resulting text can then be inserted into the non windows program.

A demonstration of how this works

The disadvantage of this method is that it will not allow you to edit text outside of the dictation scratchpad dialog

Voice recognition (another name for speech recognition) is usually built in to most smart phones and pads/tablets. In the case of Android tablets and phones you use it by touching the microphone icon on the virtual keyboard (it could be not showing because it is switched of in settings) and a prompt will ask you to speak. After speaking the text will appear within the app that you where using. Android being a Google product the conversion of the voice to text is done via the Google servers on the internet (cloud) so is very fast and reasonably accurate but requires an internet connection.

The latest Apple Ipads have SR for dictating text via a microphone on the virtual keypad similar to Android. But not all have Siri

Apple Siri is a way to interact with a some Apple Iphones and IPads. You say what you want it to do or what information you want and it answers you, sometimes.

Ther is similar functionality on Android devices.

1. - Has improved over the years
2. - Is better on my phone than my computer
3. - Is more useful on a phone or pad than on a computer (Because it is difficult to type on a virtual keyboard)
4. - Is much faster than typing, on any device
5. - Better to correct its mistakes by saying “correct that”, so it learns
6. - It is better to dictate short sentences or part of sentences than longer ones, that way they are easier to correct if necessary
7. - If you have trouble in it understanding a word try saying “Spell it”. That way it learns
8. - Sometimes it is easier to say “ delete paragraph ” and dictate that paragraph again
9. - If it constantly makes the same mistake use the dictionary editor.
10. - It picks up words easier within a sentence than the words spoken just by itself.
11. - If the word you say is a command, e.g. Right click, say the word “literal” before the command eg Literal Right Click or literal iteral
12. - Like most writing it can be easier to get it roughly written and then do multiple edits.
13. - In some instances it can be quicker to use a combination of speech recognition and the keyboard
14. - I find it difficult to remember the commands and to find them in the speech reference card
15. - Watch the microphone level to make sure it is working correctly, it should roughly be in the middle
16. - Sometimes nothing seems to work at all, switch off voice recognition and switch it on again
17. - I'm sure not all the commands are in the speech reference card
18. - It is not possible to see or modify the words used to operate a command (Can use additional words with Speech Macros)
19. - Tends to make you write differently
20. - May not spell the way you spell.

A very good series of 12 Video tutorials on speech recognition http://www.youtube.com/course?list=EC0C6EE1970D9B1447

Mainly on the use of the tutorial http://www.youtube.com/watch?v=gpHB_jW_TEc

Mainly on Manipulating the screen http://www.youtube.com/watch?v=NFKHAjHrc5A

Comparison of dragon naturally speaking with windows 7 SR https://www.youtube.com/watch?v=VJ0i1dHZH48

These two below do not show how to use speak recognition, but rather show you how NOT to use it

A good example of how not to use speech recognition https://www.youtube.com/watch?v=Z3KU9nLEueo (This guy obviously did NOT do the tutorial)

Some other things that can be done with SR https://www.youtube.com/watch?v=qLWh1bQuNy4 (This guy should have had his radio in the background switched off)

A link to an online typing speed tester: http://www.typeonline.co.uk/typingspeed.php You can use this to compared you're speech recognition typing speed with your keyboarding typing speed.

Many people, including myself in the past, dismiss speech recognition because they say it is easier to just type it. See its Easier Learning speech recognition is like learning any new function, initially it will be faster to revert to the old away by typing, however after a very short period time you will find it quicker to use speech recognition.

I have discovered that I now find it frustrating to go back to typing, because it is soooooooo sloooooooow compared to speech recognition.

I strongly suggest you spend the hour or so going through the speech recognition tutorial and reference card and if necessary the exercises I have throughout these notes. If you also want further training watch the videos in the links I had earlier .

If you have any difficulty learning speech recognition you can e-mail me and I am willing to help anybody out at a prearranged time.

My Email Address is : geoffreykgreig@gmail.com

Are you going to try speech recognition?

Copy address bar to chat again

So as I said at the beginning I'm willing to stay in this room for a while to help anybody out with speech recognition