My first test with CC in 10.4.2 and a few questions probably.

Hi everyone.

I had a few days to experiment with the CC (closed caption) feature in FCP, since I figure it is just a matter of time before I get a request to do this for a project. Unfortunately I missed Larry Jordan's webinar on just this subject since I was out of town. I'm thinking about purchasing this for reference. I did find this Ripple Training video on YouTube which gave me a good starting point:



So the first thing I did was download SpeedScribe and give that a shot. There are 15 trial minutes to play around with. Working with a final "picture locked" video is the way to make this work.

In FCP, I muted all tracks except for dialogue, and exported an mp3 sound track. That was pretty quick, as well as uploading to SpeedScribe's (SS)servers. I won't go into how to use SS, but I can tell you it is pretty straight forward.

There were two things that confused me though. Upon import of my audio file, there was a pull-down menu asking what the frame rate of the video was. I know my vieo was 29.97 but there were 2 flavors of that DF (drop frame) or NDF (non drop frame), I didn't know which to chose, so I went with the NDF.

It also asked if I wanted the ITT or CEA-608conversion. I first tried the ITT but I found that ITT maybe more geared to upload to ITunes, so I also did a 608 version which in the end worked well with YouTube.

In SS, you can edit the dialogue if there are transcription errors, etc. prior to exporting to ITT or 608. I then exported my transcript to my desktop. BTW- SS did an amazing job transcribing voices that had very thick accents, which I marveled at!

Back in FCP in the project that you were working with, File>Captions>Import Caption and select newly created caption file. And like magic, captions appear at top of project window. Just amazing!!! Could this actually work??? I did notice that the type itself looked kind of crappy, but this was resolved after export somehow. To me, the timing of the captions coming and going seemed to be acceptable as far as I was concerned.

I went t through each transcription to check for spelling, punctuation and so on. There were a few things that needed changing and that took some time. There were also instances where if there were three sentences, the middle sentence was just one word, so I tried to reformat that, which proved to be a little bug-y.

I think there are more formatting options with a 608 file than ITT. I use the term format loosely, since you can't alter the size of the text. Just color of text, and transparency of the text box background, as well as moving text up, down left and right. I chose to have my text center, though I had to manually move the text on some dialog since it didn't seem to reflect that change when applied.

Without going onto all my errors, I exported a QT master file which embeds the CC in the QT wrapper. At first I opened my new file with QT 7, thinking that had more control and viewing options. But nothing!!! DOH. I then decided to try and open the file in QT 10, and was able to see my CC, using the CC button on the QT controller.

It was now onto trying to see what happens when uploading to YouTube. In Compressor, I exported a short 40 second clip from my QT master, just to save time.

Uploaded that short file to YouTube, but after upload I did not se any CC control on the interface. HHMMM. So I went back into my YT settings for that file, and this is where my memory is flakey, I went to the Closed Caption menue tab for that video, and I think I had to select "Upload CC" from their dialog. To be honest, I can't remember what I did exactly, so use this as just a hint.

Whatever I did, I went back to my video, and there was now a control button to turn the CC on, Auto, and 2 flavor of english. Auto was recommended, but it changed my formatting where all my text appeared flush left. It worked fine when I chose one of the 2 English versions, but I can't recall which one. As mentioned above, the quality of the text in the the YT upload (as well as the QT exports) was excellent and looked better than what I was seeing in the FCP Project.

I'm sorry I probably left out some important stuff, but maybe this will make it a little easier for anyone to have a go at this incredible function.

My question would be: Let's say you want to get a translation of a transcription, say English to Spanish. Since the length and structure of the two languages is so different, I wonder if the translation would work timing wise as well as the native language. Another day.

Jonathan