Monday, July 25, 2011

Novacut, An Open Source NLE and More; interview with Jason DeRose

Over the last couple of weeks there's been some buzz about a KickStarter campaign for an Open-Source video editor called Novacut. After reading the page and some of the associated links (see below) I was curious about the project, and managed to talk to Lead Developer & Project Visionary Jason DeRose about this interesting project.

As of today, the project is more than half way to their $25,000 goal.

Michael: What is Novacut?
Jason: It’s two things. We’re building an editor, and we are also building a distribution platform.

What’s your role?
I’m the lead developer, and this is my brainchild too. I’ve had numerous help refining the idea from the other two co-founders, Tara Oldfield and Jeffrey Ballagh.

Where did the idea come from then?
My wife Tara is a photographer and when she got a 5D Mark II we started playing with the video, and started realizing there was going to be an explosion of new content producers because of the cost reduction.

I’m an open source developer and open source has had version control tools for quite a while. It’s basically a way to allow teams of developers from all over the world to work together in a coordinated fashion. If you can take from a huge pool you can find the very best people, people who are the right fit for what you are doing and I thought the same thing would apply to content creation. Especially for steps that are independent like the colorist, it shouldn’t really matter where they are located.

Is that the key difference between this tool and the other NLE’s that are out there?
That’s the real kicker, for sure. Some of the differences are that because of the way we had to build it for the distributed/collaborative workflow, it’s also super easy for us to spread out rendering over a local cluster or the cloud. There’s just a lot of flexibility.

Novacut logo

With the distributed workflow, will you still need to have the files locally?
Yes. It’s not like a website. It doesn’t only exist on the cloud. We can utilize the cloud, but you can work locally also.

Usually with most NLEs you have content and you have a project file. Do you have the same sort of model?
That’s another big differentiator, our Distributed Media Library. One of the things it can do is, if you and I are editing, and you opened a project that you haven’t worked on yet that I shared with you, it can figure out what files you already have locally, and what ones you need. If there was some overlap with a previous project of assets you need, it knows what you’re missing and it can download them from the cloud or from me, or it can tell you to insert a certain removable drive that I mailed to you.

How is the initial connection made?
You’ll have me as a contact, and you’ll say “Share this project with Jason”

And that’s done through your server?
We’ll have our own cloud infrastructure and people can have accounts there, though we also want to make it so that you don’t have to go through us if you don’t want to. We’re going to have to come up with some creative solutions there, but it will be something like sending an instant message or something like that.

What platforms will it run on?
Ubuntu is the primary platform. That’s a good place for us to do our initial development because it’s one that we know and also there’s a lot of infrastructure that helps make it a very productive platform for us.

The Distributed Media Library component will probably be ported over to OS X and possibly the Editor, it just depends on where the demand is at that time. So far there hasn’t been a lot of demand for Windows so it doesn’t seem an especially important target right now.

You’re developing this in HTML 5?
Our user interface development is all HTML 5. We may encounter a few places where it’s not feasible, but so far it has been. It’s not being used in a browser; we’re using WebKit embedded in the application, so it looks and works like an application, but the UI internals are done with WebKit.

I don’t normally think of HTML 5 as something you’d develop a video editor in.
Yes, that is a place where we get the most concerned looks and eye rolls, but it’s interesting what’s going on in that market. Especially with mobile where Webkit has got a ton of development recently. For the kind of free-form “canvasy" sort of user interface that you need for an editor I think it’s actually a better way to build an editor than a conventional toolkit.

Will the basic interface and functionality be similar to existing NLEs, or do you see it as being radically different?
It’ll be some of both. We tried to go in with a really blank slate just because our primary motivation is saving artists time, figuring out where artists spend their time, where’s the software fighting against the artist instead of working with them?

There’s a few places we may do things differently, but we’re only doing that when we find another way is faster.

All in all, we think there’s a lot of great work in Final Cut Pro X and most editors we’ve been talking to like a lot of things about it. People didn’t like being blind-sided by it, and having the old one being pulled from the market.

A UI Demo (not the final interface)

How many developers are there?
I’m the only full-time developer. James Raymond is working on it full-time at the moment as he’s out of college for the summer. We’d like to hire him full-time as soon as possible. He’s leading the UI design.

The other developer is David Jordan, who’d also a college student. He’s right now doing a summer project on optical flow; frame rate interpolation if you’re going to step down from, for example, 30 frames per second to 24 frames per second.

Where are you in the development/design stage?
The Distributed Media Library is almost production ready. One of the reasons we started there is that it was the place where the most work and highest risk was, and we wanted to get that done first.

We spent the last year researching and talking to artist after artist, and really honing our priorities and deciding how we were going to measure success, because saving time is such a key thing. We can’t just assume we’re saving time, we have to actually measure it and know that we are.

We’re just getting into serious UI work, and we are prototyping the UI with throwaway implementations. We are going to do a lot of variations of each bit of the UI and we’re trying to get other people to join in on that collaboration. Then we’ll measure the prototypes and that’s how we’ll decide which version to do the full implementation of.

Right now we’re working on making sure we have insanely fast cutting and getting the media organization right.

You’ve been working on it for over a year?
We started about last July, though serious development started in October.

When do you think you’re going to have something that beta users will be able to play with?
It’s a little hard to estimate that. We think we’re probably two or three months away. It’s going to be really simple, it’s going to be media organization, cutting and rendering.

For the editor we’re building on an open source library called Gstreamer; that’s our multimedia backend. Over time we will have to do some work on it - there are some improvements that need to be made there - but starting off we can do basic editing without touching any multi-media stuff ourselves. It’s not nearly as much work as it may seem.

Since you’re an Open Source project, what about the video formats that require licensing?
There’s a bit of legal gray area, but we’re trying to build a close relationship with Canonicle, the company behind Ubuntu. If you’re an OEM selling a computer you can license H.264 through them so that your users have a licensed version. We might do something like that. There are open source implementations for built-in encoding and decoding, even for H.264. The software exists, the issue is that in the US you need to pay per-unit royalties.

Can you talk a little about the distribution platform, is this something like iTunes or Youtube or something else?
More like iTunes. We’re really focusing on native playback applications on tablets and TVs. Pretty much all of these lost cost productions are crowdfunding, and for them the worst thing is obscurity. Getting it out there to as many people as possible is key, and they obviously rely on volunteer payment. They’re doing alright, but not that many of them are very profitable. Where we are stepping up the game is tightly integrating the experience of watching with having the opportunity to crowd fund.

If you are interested in Novacut, make sure you read the Updates section on the Kickstarter page, there's a lot of interesting additional information there.


No comments: