Article Writing: Is it Better to Type or use Voice?

Peter B. Giblett By Peter B. Giblett, 8th Oct 2013 | Follow this author | RSS Feed | Short URL http://nut.bz/zbh0miqm/
Posted in Wikinut>Writing>Columns & Opinions

We are a mad lot in the modern world, but if you see a person talking to their computer or smart-phone then chances are you should not call the men in white coats - there is a new generation of speech recognition software out there which may be worth looking at to create your articles.

A new Generation of Speech Recognition

The question of whether it is better to use voice to type is actually an interesting one. Typing is one of the most common input methods to any device, we do it almost naturally, yet we speak much more quickly than we type so it would seem obvious that we use our voice as an input device and there is something magical about would simply appearing on the page. Truth is I've been looking at voice recognition software for over 20 years, I guess it is the Star Trek fan in me that is fascinated with speaking to computers, getting the computer to understand you has always been a challenge especially having an English accent and the fact that most software has been built to understand American voices rather than English ones, but of course I'm always wanting to try this again and again and again and so I shall, until voice recognition works properly. Some of the issues it seems to face is the difference between certain words such as "show" and "shall" and in truth it is difficult to get speech recognition 100% perfect, but with a little editing (usually involving typing) anything is possible, perhaps proving that it is the combination of input devices that works best.

In fact I am writing this article using Google's speech recognition software, it is not perfect but it's better than most things I've tried to date. One of the great challenges of speech recognition is actually speaking clearly, background noise can be so troublesome so you need a quiet area in which to space and in in truth that is a problem because most of the places that we are in tend to be very noisy, such a shopping malls or airports. The other thing I've noticed is that all sentence pauses must be done manually adding a comma is something you need to type, however on when I spoke the word "comma" the software actually inserted a "," but when I said "full stop" it has ignored that, so clearly I have to understand the way it forms sentences and the way it uses those special characters, that is clearly going to be a challenge, but they say everything that's worth doing is worth persevering with, and it is only after persevering that we are going to succeed. With further experimentation saying comma will insert a "," into the text and saying period will insert a "." in the present location, although I have found it much easier to insert special characters myself by typing.

Will it be worth using voice as an input method rather than typing? This is already proven interesting because I have actually even been able to edit this document using voice. The accuracy of the voice transcription is always a concern and I think this might have been why I had given up with voice in the past, but saying that I would like to know what the word "fog" has to do with "past" because that is what it gave me when I said "past", but in truth I'm still very surprised at how good this is and I do think I shall continue using it in the future, despite the challenges.

The Challenge of Background Noise

Some of the dictation that I've done here has been with serious background noise in place in order to try and confuse the machine so that it would not as function as well as it could in perfect conditions and so far it has performed well, that said though typing on a smartphone is not always perfect either, the only problem seems to be the fact that it has underlined a large section of my document, certainly undesirable side effect, but this was put right by some editing on the computer, one reason we cannot give those devices up.

From time to time it seems the speech recognition goes haywire and is unable to interpret some of the words that I say, the challenge is words like "unable" which seems to come out like "in a bowl" or "in a ball" and it will take some time to figure out whether it is my understanding of speech recognition or maybe I speak with slight slur. One thing is certain using a cellphone as an input device rather than a microphone connected alongside your mouth is certainly an improvement, I'm actually using this at about a foot away from my mouth and it seems to work reasonably well.

Reality is the background noise is ever-present whatever we do and we cannot get away from this fact, it is good to know that the current generation to speech recognition is managing to avoid the background noise because right now there's a song playing in the background and the only thing that it is recording is my voice despite this it is still necessary to check and recheck what it has dictated and it is necessary to check on what you said because sometimes you actually say the wrong thing and there is no accounting for human error, whatever the situation.

Other People's Speech

One of the problems about speech recognition software has always been the question of training that you have to actually go through a training cycle with the software in order that the software recognizes your voice with the Google Voice keyboard this is different because and in order to prove this I asked my son what he was doing and this is the unedited response:

"I'm watching Star Trek Deep Space nine"

Not bad, because earlier software would have not have been able to recognise an untrained voice and of course record someone else would therefore have been impossible. I tried something else, have it transcribe voices from a TV program, which after testing several times failed to transcribe anything. I do intend to try this however at a meeting and see if it can record individuals and their speeches to see what the results are like.

Watching your Ps and Qs

One aspect of dictation is the fact that the machine will pick up everything you say including undesirable things such as "um" and "er" as well as any slang that you care to mention, swear words however were replaced in part with "***". Part of the challenge here is that we do not think in complete sentences we tend to make things up as we go along and if we are editing a document we will correct it on the fly. As much as I would like to just talk at the computer and make it go on and on and on it is unlikely that I could make this happen but one thing is for sure talking to a computer is certainly going to be a part of the future. Right now we are probably 80 percent of the way there with voice recognition but there is still a way to go, and in part it is about knowing the right words to use in the right place but I think with a little patience people can make it work.

For a computer program to provide the right words in the right scenarios will take another leap in thinking, there are many words that would never be used together and there would have to some research carried out in this area in order that the correct words are offered up when speaking because the sentence should make sense at the end of the day and some of the word combinations offered can look strange, thank goodness for editing.

Transcribing a Speech

Over the next few weeks I do intend to test the ability to transcribe a speech at my local Toastmasters just to see how good it is at recording somebody else from a distance like in the front row in front of the stage and test how much of the speech is transcribed and how many problems there are in doing so. I fully expect to have to pay a lot of attention to doing this as I don't know how much it will transcribe and what will happen if there is a significant pause in the speakers words or what happens if they go to the other side of the stage during their speech.

To me this is a period of experimentation and with experimentation there are good things that will happen and bad things that will happen, things will go well and things that will go badly, in this case the purpose of experimenting is just to see what will happen is it good is it bad? I don't know.

One thing is very very clear you will have to do much editing to your transcribed document before it can be accepted as a final copy, bit the same is true with any piece of writing.

Recent Thoughts

The following articles, have each been published by Peter Giblett and may be of some interest to you:

Thinking about making your own personal contribution? You should do something about it now! Your viewpoint is important! The first thing you could do is ADD a comment providing your own thoughts and feedback, which is always appreciated. Sometimes however your thoughts transcend the limits of a simple comment, then you need to make your own contribution of 300 words or more; perhaps penning your own article is more appropriate, well you can make your own views felt.

Join Wikinut today and have some fun telling the world about your views and how the world needs to change.

Tags

Article, Article Writing, Articles, Background Noise, Dictate, Dictation, Manners, Other Peoples Speech, Speech Recognition, Typing, Typing Skills

Meet the author

author avatar Peter B. Giblett
Author of "Is your Business Ready? For the Social Media Revolution"

Social media consultant, with C-Level background.

Share this page

moderator Mark Gordon Brown moderated this page.
If you have any complaints about this content, please let us know

Comments

author avatar Mark Gordon Brown
8th Oct 2013 (#)

I wonder about using these and if my cat happens to meow at the time, would it have meowing throughout my article? It would sure be easier to write my thoughts as they flowed though.

Reply to this comment

author avatar Peter B. Giblett
10th Oct 2013 (#)

When I talked about background noise I meant heavy noise with the TV blasting and the Dyson going, unnaturally loud and it still worked.

Reply to this comment

author avatar MarilynDavisatTIERS
9th Oct 2013 (#)

Good Evening, Peter. I used Dragon Speak years ago and after much trial and error, it learned, or probably more accurately, I learned the commands and was pleased with the experience. Thanks for an undate on what's out there now. ~Marilyn

Reply to this comment

author avatar Peter B. Giblett
10th Oct 2013 (#)

I have used several generations of Dragon without success and soon gave up, never being that pleased with the results.

Reply to this comment

author avatar Phyl Campbell
9th Oct 2013 (#)

Ha Ha Mark!

Peter -- this is great! I need to work more with software like this, as I certainly talk faster than I type. I also need to record my books since people have asked me for something they "can just listen to in the car." Hopefully technology can make this easier for me! Great article.

Reply to this comment

author avatar Peter B. Giblett
10th Oct 2013 (#)

Phyl, we talk somewhere between 160 and 200 words per minute but only the best typists can get to speeds of 80 to 100 words per minute.

Reply to this comment

author avatar Connie McKinney
9th Oct 2013 (#)

I think I would have the same problem Mark would. My cat would meow at the most inopportunte times. Seriously, interesting thoughts, Peter. I can see pros and cons to using a voice recognition system over plain old-fashioned typing.

Reply to this comment

author avatar Ptrikha
9th Oct 2013 (#)

Is "Google's speech recognition software" a paid software? Overall, it would be interesting to use speech recognition software once in a while, and see what the results are.

Reply to this comment

author avatar Peter B. Giblett
10th Oct 2013 (#)

Google voice keyboard is actually a default part of the Android operating system.

Reply to this comment

author avatar Ptrikha
18th Oct 2013 (#)

ok

Reply to this comment

author avatar Mariah
9th Oct 2013 (#)

Advancement in technology is amazing, but in agreement with Mark, background noise in this case could be a frustrating drawback.
I would have to shut the dog away for certain.

Reply to this comment

author avatar M G Singh
9th Oct 2013 (#)

Never tried speech software, but looks good.

Reply to this comment

author avatar MMD
9th Oct 2013 (#)

I would prefer to write rather then speech.

Reply to this comment

author avatar Denise Salmon
9th Oct 2013 (#)

you shared some very interesting facts, great artcle

Reply to this comment

author avatar Ellen
9th Oct 2013 (#)

Interesting thoughts on the subject. I could not do it. It takes me awhile to get my thoughts together, but for those who feel they can do the voice all the more power to them. Good article.

Reply to this comment

author avatar Fern Mc Costigan
10th Oct 2013 (#)

Nice and interesting post!

Reply to this comment

author avatar Randhir Bechoo
13th Oct 2013 (#)

Interesting article.

Reply to this comment

author avatar William Power
23rd Oct 2013 (#)

Interesting article. I guess, as someone familiar with editing, I know I`m going to have to proof-read the dickens out of a "voice-text" work because of homonyms and other errors. Like those that claim they still love a good book in their hands when they read (rather than an eBook), I love to write with my keyboard. That being said, the future often drags us along--like it or not. WP

Reply to this comment

author avatar Peter B. Giblett
24th Oct 2013 (#)

William, that is an interesting thought, but even with typing has to be proof-read to ensure good quality - something I wish some Internet writers would do more of.

Reply to this comment

author avatar Nadine May
7th Jan 2014 (#)

Gosh how interesting that you use "voice-text" I never thought of doing that for articles! I have used Dragon Speaking for when I write dialog in my novels when I have been on my own in our office, which hardly ever, but I will keep it in mind..

Reply to this comment

Add a comment
Username
Can't login?
Password