Speech to text

[John Rofrano]

[Bill Davis] “…theres typically a budget for captioning when required.”

I guess if you work at a big post house there is… but if you are an independent delivering broadcast content to cable stations, captioning is always required and there is no way to do it yourself in FCPX. The show in that I was finishing editor for in that video (Painting & Travel) is shot entirely by Roger and his wife traveling the country just the two of them. Roger has no budget and you can’t delivery anything to PBS without captioning so Vegas Pro and my Caption Assistant plug-in allows him to do it himself. Roger actually helped me develop the tool. I was looking to do something similar for FCPX and was shocked it had no provision for captioning because in the US, it’s required now for everything going to broadcast.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[John Rofrano]

[Lance Bachelder] “Nice video John – the last feature I cut in Vegas was in 10 and I wish I had known about your ap. Hopefully you’ll be able to do something like someday for FCPX.”

Thanks Lance. Yea, thats why I was looking into it. I’m trying to figure out what plug-ins I can write for FCPX.

[Lance Bachelder] “I’m really loving FCPX much more than Vegas but do miss a lot of useful features in Vegas that don’t exist in any other NLE. I only wish it hadn’t become so buggy or I still might be using it regularly.”

It is sad that Vegas got so unstable. I’m getting fairly proficient at FCPX as well now although I have a long way to go. 😉

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Craig Seeman]

There’s MacCaption.
https://www.telestream.net/captioning/overview.htm

[Andy Field]

We use the Text to Speech all the time – if someone speaks clear standard English it does a fairly good job – accented…not as good. The Boris Soundbite is great if you’ve already transcribed something and need to find it fast as it’s a phrase find program – it looks for sound patterns and matches them..but doesn’t actually transcribe anything.

Andy Field
FieldVision Productions
N. Bethesda, Maryland 20852

[John Rofrano]

[Craig Seeman] “There’s MacCaption.”

Wow $1,095! I’m definitely not charging enough for my Caption Assistant plug-in! lol 😀

Thanks for the pointer. When we were looking for post houses to do the captioning they wanted $300 per episode so $1,095 pales in comparison to a 13 episode season at $300 each ($3,900) so I can see how they can charge that much. I’ll check it out.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Craig Seeman]

BTW Telestream MacCaption has specific info for FCPX workflow in their KBase. It mentions both file based and tape based workflows.

https://telestream.force.com/kb/articles/Knowledge_Article/Closed-Captioning-with-Final-Cut-Pro-X/?q=AJA&l=en_US&fs=Search&pn=1

[Michael Phillips]

SoundBite, like PhrseFind does not require a transcript to be done. You may be confusing that with ScriptSync. SounBite and PhraseFind eliminate the need for transcripts as it can do ad-hoc searching for words and phrases directly on the media itself. In many cases, it can save the cost of doing a transcript in the first place.

Adobe’s Speech to text did add a feature that “helped” it by providing it is script to start with, but then again, that sort of defeated the whole purpose of a speech to text engine in the first place.

Michael

[Michael Sanders]

When researching setting up my other half as a transcriber (so she could earn money whilst doing her PhD) I was told the industry has maybe ten years at most. Computer transcription is so good now its scary.

Just try this: find a press conference on TV, hold your iPhone up to the speaker in Notes or whatever in dictation mode.. Even on an scottish accent it did a very good job!

Michael

Michael Sanders
London Based DP/Editor

[Andy Field]

Sound bite doesn’t need a transcript but without one how do you know the phrase or words to ask it to find

[Oliver Peters]

[Andy Field] “Sound bite doesn’t need a transcript but without one how do you know the phrase or words to ask it to find”

SoundBite, like PhraseFind indexes audio based on phonetics. You type in a word – regardless of spelling, as long as the pronunciation would match – and it will find the matches. You can dial in accuracy tolerance. It will display as many “hits” as it can locate within the indexed clips. Let’s say you need to find a certain word in an interview where the word ends the sentence. Type in the word and then review the various search results to see if you find that word ending a sentence. The “frankenbite” scenario.

Or search an interview for every time the speaker says “Apple” or a person’s name or certain phrases. It will display the results in the same manner. The lower the percentage, the wider the search results, but also the more sloppy. The higher the threshold percentage, the fewer hits, but with greater accuracy (at the risk of missing some good ones).

I’ve done a lot of interview-based docs, web videos and marketing pieces and almost never have a transcript. I work with the material in front of me. I’ve found that transcripts can be a false aid, since often times things that looks like they would edit together well have the wrong inflections, so I prefer to work without transcripts. OTOH, transcripts (with a way to match locations in the media) can be a get help, when the producer says, “What about statement XYZ? I seem to recall them saying that.” If they can give you a way to find it based on the transcript, then it’s easy to call up and review.

– Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com