Through work I have just got hold of MacSpeech Dictate. This is a talk/type application that allows you to talk and your Mac will magically type what you are saying, sort of. You do have to remember to say the words "full stop" if you want a ".", along with other such commands. This can be problematic. For example if you want to say the word "ampersand" rather than have an "&" , or if you want to say the word "comma" rather than have a "," .
I am using this software to create this post and I initially refused to go back and correct mistakes via my keyboard. However, that notion didn't last long because aspects of the last sentence of the last paragraph were completely nonsensical Due to the fact that I couldn't say the word "comma" without the software placing a "," where I wanted the word. Teething problems aside, I can see how this is going to be very useful in helping me manage my RSI.
Now for some swearing; Buck, sheet, bollocks. 1 out of 3! Not so good.
It will be interesting to see whether using this application with effect the way I talk and write. So, if you should bump into me and I keep saying the word comma, and if in future posts you keep reading the word Buck, then you will know that the software and I are starting to merge!
11 comments:
http://dictatedtexts.blogspot.com/
Interesting and puzzling in equal measure.
Does it work if you say something like "quote c o m m a" etc? Just curious, haven't tried it myself.
Not sure if you saw where Erica writes:
This blog records my notes on whatever it is I'm thinking about at the moment, transcribed using MacSpeech Dictate. Quote I read somewhere: "Language doesn't need to make sense as long as its communicating"
Richard: If you try to spell things out it can get it wrong, for example, trying to say 'c' and the application decoding that as 'see' or 'sea' is problematic. However, it's early days and I understand the software tunes into your voice and speech patterns the more you use it.
Steven: I did see that. Obviously a definition of 'communication' and 'making sense' would help here as I'm not sure to 'make sense' is really the right choice of words to explain what is meant. For a communication to be successful (and thus communicate) the receiver makes some sort of 'sense' of it, regardless of a full cognitive understanding of the processes by which the message was communicated. Further to this, it also depends on what the intention of the communicator is as a different message may be received than the one that was intended. It could be argued that this is more likely if the chosen language doesn't 'make sense'. That said, I agree with the gist of what I think Erica is getting at. There are similar debates within graphic design and it's ability to communicate vs aesthetic stylings. David Carson famously says of his cut up distorted type, "Don't confuse legibility with communication" and I whole heartedly agree.
It's all semantics at the end of the day!
indeed, also this:
http://en.wikipedia.org/wiki/Entropy_%28information_theory%29
also some visual/concrete poetry works with this:
http://www.rediscov.com/sacknerarchives/welcome.aspx
I am in the middle of trying to write a new module for our design course so I will look at these later.
are you writing it with MacSpeech Dictate?
I started to but then it crashed after making several mistakes. It became time consuming to keep going back and correcting what I had said and interrupted my thought processes so I've gone back to typing.
I only got the software yesterday so will need a fair bit of practice (preferably on less crucial pieces of writing) before I'm proficient with it.
It is exciting though, it makes the McJunk book I've started writing and doing an MA next year much less daunting RSI wise.
Must crack on.
Cue scene from one of the Star-Treks, where Scotty picks up the mouse and say "Computer..."
Voice recognition has been around for donkeys years though... I remember various packages available being investigated at 'the labs' years back, some even being quite competent (and PC based). Given the progress made in audio analysis in the last 15 years (discounting Antares Auto-tune as progress, thanks to T-Pain), one "wonders" exactly where (as in which domain) the accuracy problem lies.. (ie is it the processing/analysis/correlation/etc or the difficult or inexpressible subtleties of language, or all, or none?)
But regardless, the quote from Erica is spot-on, and a model I find myself falling back to when attempts to practice Heidi-deutsch fail and we have to revert to Swinglish to _communicate_.
Post a Comment