Creative COW SIGN IN :: SPONSORS :: ADVERTISING :: ABOUT US :: CONTACT US :: FAQ
Creative COW's LinkedIn GroupCreative COW's Facebook PageCreative COW on TwitterCreative COW's Google+ PageCreative COW on YouTube
APPLE FINAL CUT PRO:HomeFCP ForumFCP XFCPX TechniquesFCP TutorialsFC ServerBasics ForumPodcastFAQ

FCPX + Mountain Lion + Dictation = A lot of fun!

COW Forums : Apple FCPX or Not: The Debate

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Share on Facebook
Mark DobsonFCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 11:45:57 am

I read a recent Philip Hodgetts blog and having installed Mountain Lion on a MacBook Pro decided to try out the dictation function from within FCPX.

And as he points out in his post - it works! (as it will within any application)

http://www.philiphodgetts.com/2012/08/final-cut-pro-x-and-mountain-lion-now...

It makes lot of mistakes but gets the bulk of what you want to say down.

So far I have used it to dictate the words for keyword collections, favourites and titles. In fact anywhere where you would type it will work.

Whether it will actually be a useful function within FCPX, I'm not sure but it's certainly fun to play about with.

As a lot of my work involves mechanical and technical vocabulary and it definitely struggles a bit.

And I just can't get it to swear!

If you say F**k it writes Flock.


Return to posts index
Reply   Like  

Michael PhillipsRe: FCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 12:29:18 pm

It will be interesting to see where speech to text goes over time, as the overall percentage of "correct" has flat lined for the most part over the years - probably due to the fact that these are dictionary based conversions and languages change over time.

But that aside, many companies, broadcasters and such use it as means to jump start the transcription process where they will listen to the dialog via headphones and repeat back into the diction software, then make any corrections from there. But in tests so far, a fast typists still beats this method in both time and accuracy, but that is always a decision of budget for who you're hiring. Things may change in the future.

I have also tried using speech as a means to drive the editing software - not this has been a while, but I was much faster using keyboard and mouse than dictating the commands. Simple things like play and stop were almost a wash in time - but everything else was slower. Speech is a real time process, many things including thought are faster than real time.

Michael

Michael Phillips


Return to posts index
Reply   Like  

Mark DobsonRe: FCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 12:56:19 pm

[Michael Phillips] "But that aside, many companies, broadcasters and such use it as means to jump start the transcription process where they will listen to the dialog via headphones and repeat back into the diction software, then make any corrections from there."

Transcription of interviews is at the heart of the work we do, so a viable application that aids this process would be welcomed.

I'm of the generation that never learnt to type correctly, and whilst I can fingerpick a guitar pretty effectively I'm defiantly a member of the two finger and a thumb school of typing.


We don't do a full transcription anymore but rather summarise a statement, sentence or paragraph. So being able to have this summary transcribed within a favourite could actually save quite a lot of time.


Return to posts index
Reply   Like  


Michael PhillipsRe: FCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 1:15:10 pm

I use the ScriptSync function in Media Composer for anything that is scripted or transcripted as it presents the most unique method by which to edit - so the more complete the transcript, the more accurate and granular the edit becomes when working directly off the tran(script).

In your scenario, the summary seems to work fine and perhaps these tools will find a place in your workflow.

I have been following these technologies for a long while now, and one of the more interesting ones is speaker identification is finally becoming something useable - so not only would a search include "someone that says this.." but also "who says it". Imagine dialog replacement... and in those scenarios, no script or transcript is needed. Tools like PhraseFind for Media Composer and Get for FCP and Premiere bring this type of technology to the end user. I am keeping a close eye on what that will become.

Michael

Michael Phillips


Return to posts index
Reply   Like  

Richard HerdRe: FCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 11:00:46 pm

[Michael Phillips] "probably due to the fact that these are dictionary based conversions"

A former linguist here... it has to do with a thing called allophones. Computers just stink at phonology and guessing the underlying representation.

Here's an example: Hold your tongue between your fingers and say a sentence to a person; then, say it to a computer. The human can get it. The computer can't. It has to do with a human trait of inductive reasoning, guessing the meaning.

OR try this one: tap a pencil and say a sentence. The human can filter out the pencil and concentrate on the language, but the computer can't.

Computers don't do very well with similar contrastive pairs, like s/z p/b. But it's really good with the differences between sibilants and ejaculates and fricatives. So it's almost as if someone needs to invent a new LUI -- language user interface. I tried once, but it ain't going to work because I sounded like an idioted talking to my computer. It was all "b f d t."


Return to posts index
Reply   Like  

Liam HallRe: FCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 8:40:48 pm

[Mark Dobson] "If you say F**k it writes Flock."

I said that quite a lot last time I tried FCPX.

Liam Hall
Director/DoP/Editor
http://www.liamhall.net


Return to posts index
Reply   Like  
+2


Bill DavisRe: FCPX + Mountain Lion + Dictation = A lot of fun!
by on Aug 6, 2012 at 11:52:00 pm

I've noted this before, but perhaps not in this forum...

But back in the early Mac days, there was an early speech processor called Voice Navigator.

I got a demo copy for a software review and I was messing around with it late one night after a party programming commands - and when the UNDO command was active I giggled and programmed in "DAMMIT" as the key phrase.

And forgot all about it.

Imagine my astonishment when I was using it a week later, screwed up something, muttered "DAMMIT" - and the computer instantly invoked UNDO!

One of my "best moments in computing" ever.

"Before speaking out ask yourself whether your words are true, whether they are respectful and whether they are needed in our civil discussions."-Justice O'Connor


Return to posts index
Reply   Like  
+1

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Share on Facebook


FORUMSTUTORIALSFEATURESVIDEOSPODCASTSEVENTSSERVICESNEWSLETTERNEWSBLOGS

Creative COW LinkedIn Group Creative COW Facebook Page Creative COW on Twitter
© 2014 CreativeCOW.net All rights are reserved. - Privacy Policy

[Top]