Wedding Speeches - Transcribing

[James Hollingsworth]

Given the current vogue to include snippets of wedding speeches in short wedding films, does anyone know of any software that will transcribe the speeches so that we can print them off and read through them rather than painfully having to listen through them again. I know there is plenty of software that will give a voice to your words, but is there any that does it the other way around?

[Chris Harding]

There are plenty of packages you can get that do voice to text James ...the popular ones like Dragon Speaking work but often need some "training" to recognise the way you pronounce words. I can imagine it would struggle with some of the accents in the UK ... However it's worth a try! You would have to go thru all the text afterwards and correct mis-spelt words ... Imagine the brides father is North of England and the groom's father is Scottish!! That would be a tough one!!

I'm sure you can get a trial version of something??

[James Hollingsworth]

Thanks Chris. Will take a look at that. I don't actually like including snippets from the speeches but I do it when asked. Personally, I think the words date the images over time, after all, no-one likes the sound of their own voice, wherever they are from.

[Gary Huff]

I've heard Boris Soundbite does a pretty good job, though I'm not sure if it does full-on transcription.

[Vince Pachiano]

I've only had to do this once, but I relied on the most powerful company on the earth.
I used my Android phone and let Google listen to the video, and then emailed myself the Text.
Of course I still had to copy/paste from my desktop email app into my NLE timeline, but it got it 90% of the way there.

[Darren Levine]

if you edit in premiere, there is a speech analysis tool which is worth a go.

[Darren Levine]

nevermind apparently, adobe removed that feature.

[Gary Huff]

Originally Posted by Darren Levine

nevermind apparently, adobe removed that feature.

It's still in CC 2014, but it was removed from CC 2015 and onward because of how poorly it performed.

[Daniel James]

Youtube has a closed captioning tool.. upload the clip to youtube, allow it to do its stuff, and then download the text, and delete the clip.

As with other text to speech technology its very variable in its output.

[James Hollingsworth]

None of these methods seem very time saving, is there a gap in the market? I am thinking that the Dragon software that Chris spoke of is the best option but it isn't cheap. Those of you that do put speeches highlights within the short film, what is your method of working? I always edit the full version first including the ceremony and speeches in their entirety and then do the short film last. Normally without speeches but if I do include some bits, I have to wade through the speeches looking for bits that might accompany the images. Speeches are extremely personal and we are employed for our skills with a camera and edit suite. I don't offer a second draft facility as it would just make it too expensive so important to get it right first time. Easy for us to pick the best pictures but not so easy with the words. Perhaps you ask the b and g what bits they prefer? If so, do you send them the audio file?

[Steve Burkett]

I've recently starting added Speeches to Highlight Videos and in the first case, I asked the Bride and Groom to select the bits they wanted to use. This doesn't work with all my clients as some dither enough over music choices. In such cases, as I edit the Speeches, I simply put a marker against something I think would work. A written transcript would be nice but I feel too time consuming. I tend to cut down Speeches for the shorter video anyway, removing the gift giving, the endless thank you's to every single person who contributed to the Wedding, saving those for the full length uncut. So I have less to plough through when it comes to the Highlights.

[Rob Cantwell]

I tried to transcribe a 12 minute audio before and it took ages, any voice recognition stuff i tried failed pretty miserably, maybe its our accents i'm not sure! I ended up going like a court stenographer except going back over it again and again until it was correct.
I wouldn't offer this at all - way too time consuming for me.

[James Hollingsworth]

I agree, it is way to time consuming, whoever thought it was a good idea to start including bits from speeches must either have a tried and tested method or be charging a fortune. I too add markers as I go through but this isn't easy to do as I am also editing for camera angle so am focused on that. I only include a few bits which makes them harder to pin down. Will try to post a link to the one I am doing right now. Also agree that back and forth with the client on either music or speeches is a bad idea. Once again, fine if you are charging enough but in my opinion, £2000 plus is just not viable in the current economic climate for a single camera person unobtrusive shoot.

[James Hollingsworth]

Here is a link to the final film I did that included snippets of speeches, mainly at around the half way mark. All shot with Sony A7s and 28-135 and some on Sony PXW-X70:

Justin & Georgina's Wedding Day on Vimeo

[Andy Wilkinson]

Originally Posted by Darren Levine

nevermind apparently, adobe removed that feature.

Hi all, thought I'd revive an old thread rather than start another one. Probably got a job coming in (from California!) that will need some (technical) narratives from 3-4 scientists transcribing to text so the client can cherry-pick the bits they want in the final video...so I thought I'd see if I could do it using software.

So, is there a good, hopefully relatively cheap, software way of doing this that people are now using in 2018 with relative success?

Bit of background history below.

Fired up Adobe CC and loaded a test video, then created an audio only file to analyse...then after a lot of head scratching as to why there was no "Analyse" button in the metadata workspace I resorted to web searching...bit of digging around and I've just discovered that the feature was abandoned - and this old thread! (Transcribing was indeed removed by Adobe CC at the 2015 version according to an official Adobe help page - I'm currently using 2017 in my "locked down" Sierra OS system new Mac Pro).

No worries I thought. I still have CS6 on my old/classic Mac Pro - maybe that had it (it did). So I fired that up and did some speech analysis on that WAV file to see how it worked...and it's got about 1 word in 3 correctly transcribed into text, with gibberish words for the bulk of it :-(. Going to take a long time correcting it if that's a typical success rate...might be quicker to do it manually listening myself line by line!

But, since we're near the end of 2018 now I'm hoping that speech recognition must have come a long way since Adobe abandoned this feature 4 years ago (think Siri, Alexa etc.). So, over to you. Anyone using software with good success for this task?

Any help much appreciated.