Production Update #2: Transcribe, Transcode

July 25, 2016

Aaaaand we’re back with another (much more informative) production update. Here’s what we’ve been up to:

I, Michael, (this is Michael speaking.) broke my foot. It’s a long story, and it’s been an even longer recovery, but in that time, a new life has come over me, and about this project in particular. I’m very excited as it has led me to embark on a new journey, which I’ve Twitter-deemed #MichaelBecomesAnEditor.

I’ve never been one much for editing, that is, until I started to realize I was doing it all wrong. The more technique I’ve learned, and the more streamlined my workflow, the more efficient I’ve become, and the more I’ve just honestly enjoyed it. Andrew’s pretty much a natural at the stuff, but we’re learning together. After all, the jump from four-minute sorority videos to feature films is not a small one. Getting into the meat of this update, I want to share three very important facts I’ve learned in the earliest stages of my journey to becoming a feature film editor:

1. There’s always a shortcut you don’t know.

There are a whole lot of shortcuts in Premiere. A lot of them. Just when you think you’ve got it all figure out, you’re watching some tutorial, and the guy casually rattles off some insanely simple way to perform an otherwise time-consuming task, lickety-split. I guess the lesson worth learning here is to keep digging, keep watching videos, reading tutorials, and looking through the shortcut editor. Any time you find yourself doing something menial that takes more than two clicks over and over and over and over, see if there’s a shortcut, or if you can create one. Again, the most efficient workflow is the one that’s the most enjoyable, and vice versa.

2. TransCRIBE your footage.

Look; it’s unavoidable. Trust me, I tried. I fought it. I found out that Adobe Prelude cut their auto-transcribe function and said, “No way.”

Then I tried to sit my cute, naive little self down and edit a documentary, so of course I’m now a firm believer in transcribing footage. It doesn’t have to be hard. It’s really pretty straightforward: you need to know what people are saying, what file they say it in, and at one point in the file. It’s very time-consuming, but until you know all the material you have to work with, how can you ever start to create? Editing a doc without transcribing interviews is like building a skyscraper and just hoping you have all the brick you’ll need.

Monday night #views. #transcribing #videoediting #postproduction #documentary

A photo posted by Thrift Store Symphony (@thriftstoresymphony) on

I’ve found a pretty simple solution to the problem: using Soundflower to route my computer’s output to the input, Google Doc’s Speech-to-Text function, and a few hours every night. It’s not perfect, but it does a huge bulk of the typing for me, and then I go back and clean up/format. I’m doing about ten minutes of interviews in ~25 minutes – not a bad ratio! I may do a tutorial on the specifics pretty soon, since there’s not a lot of info on transcribing footage out there.

3. TransCODE your footage.

Okay, question for the ages: why waste the time transcoding your footage? Transcoding is taking a media file using one codec, and changing it to use another (basically. Simple enough explanation for now). So, what’s the point? It can’t make your footage look better, and it won’t make them smaller without a lot of compression, right?

Right! In fact, it’ll probably make your files one to two times BIGGER. But that doesn’t matter in terms of performance. You transcode for a more efficient workflow. Any NLE that doesn’t have a native format (here’s looking at you, Premiere Pro) can benefit from an intra-frame codec. This is a kind of codec that contains the entire information for every frame, as opposed to inter-frame codecs, which only store what changes from frame to frame.

Why does this matter? Well, you computer, when playing an inter-frame codec like H.264, has to take into account what the new frame says is changing, then go all the way back to the last frame that didn’t change to find out the parts that didn’t change, and then composite the whole thing together. I know it sounds complicated, but basically, it’s the difference between your computer just showing a bunch of images, or performing a bunch of complex calculations. Two hints as to which one it can do faster.

We’ve chosen to transcode our footage to ProRes 422 for that very reason; speed. Our editing bay is well-equipped, but we’ll never see the real power of it until we’re using good, editable files. We’ve also chosen, in hopes of a theatrical release, to edit the film at 2K (2048×1080) resolution, and in the transcoding process, we’re upscaling to 2K.

So THAT was a lot of yadda yadda, but hey, one for the nerds, amiright? Thanks, as always, for keeping up with us. Progress is being made, and we’re so so SO excited to show you all what we’re cooking up. Stay tuned for the next update, and as always, #SupportTheSymphony.

Help Make TSS Happen!

Latest Tweets

Sometimes progress is quiet and looks like a three ring binder full of sticky notes. #SupportTheSymphony

Weekly Inspiration

A weekly playlist of inspiring tracks, and generally what we're listening to. If you're a Spotifier, make sure your subscribe to the playlist!