One of the big stories at IBC was the announcement of Adobe Anywhere, a new collaborative workflow platform for video editors, loggers, and motion graphics professionals that will be introduced in 2013.
The platform will enable Prelude, Premiere Pro, and After Effects users to work together from anywhere by accessing a centralized media server. It promises to erase the boundary between facilities based and remote editing.
As details of the announcement have been covered elsewhere, I'll focus on the benefits of Adobe's approach and how it differs from other cloud editing services. I'll also add my thoughts on who it's going to be targeted at initially, how it will be priced, and the hardware requirements.
Before we begin, we'll need to cover some of the basics.
Adobe Anywhere Explained
As mentioned above, Adobe Anywhere is a collaboration software platform that customers will run on a centralized server connected to shared storage in one of their facilities. This could be a broadcast studio, a production facility, or a corporate data center.
Editors will be able to login to the Adobe Anywhere server from anywhere using forthcoming versions of Premiere Pro, Prelude or After Effects. Once connected, they will be able to work on shared projects using centralized media. This will look exactly the same as working with local media. If editors want to incorporate local footage into a shared project, they can add it to their timeline, and it will be transparently copied to the Adobe Anywhere server while they work. Once the footage finishes uploading, other people on the project will have access to it.
Take a look at the video below to get a better feel for how it works. In the demo video, editors use WiFi connected laptops to cut a sequence using video files located in another city. The 1080p source material they are working with plays back very smoothly and behaves as if it's being played back off local storage (I saw a live preview at NAB and the responsiveness in the video seems like an accurate representation of how it works).
The big benefit for editors is that they will be able to pass projects back and forth instantly without the hassle of rendering, transferring files, or relinking. In fact, it sounds like multiple editors will be able to work on the same project simultaneously and that the server will resolve conflicts between their changes.
The big change behind the scenes is that all of the rendering will be done by the central server and the resulting footage will be streamed back over the network via the "Mercury Streaming Engine".
The Mercury Streaming Engine
The Mercury Streaming Engine takes care of rendering the video and delivering it across different networks. Essentially, the computers running Premiere Pro will just be issuing instructions to the centralized server and displaying the output that is streamed back. Since the centralized server will take care of all the heavy lifting, users will be able to edit on less powerful hardware like a MacBook Air.
One of the reasons that editing performance will closely resemble desktop editing is that the Mercury Streaming Engine will intelligently manage media and bandwidth to keep the network load as low as possible. For example, with a multi-stream edit, the server will do all the rendering and compositing and send back a single stream of video rather than multiple streams.
The Mercury Streaming Engine will also optimize bandwidth use by detecting the size of the player window and sending back a stream that has the exact same dimensions. If your player window is 960 x 540, there is no need to send all the extra pixels of a 1080p stream. As the window dimensions change, the stream size will be adjusted dynamically. This should help keep playback from stuttering on relatively slow Internet connections.
To ensure that detail isn't lost, when a user pauses the video, the Mercury Streaming Engine will fetch the full-frame source material and display it. The user experience should appear pixel perfect.
There is no need to manually create proxy files because the Mercury Streaming Engine works directly with the source material. This is a big benefit compared to other approaches where time is spent creating and working with low-res proxies. With this system there is no compromise in terms of quality.
The Mercury Engine makes heavy use of GPU processing to do all of the real-time rendering and streaming. As discussed in the technology section below, doing all this processing for multiple users at once is going to require some serious computing horsepower.
The Collaboration Hub and Server Hardware
The other major component of the platform is the collaboration hub. This is the centralized database that manages all the project information, folders, permissions, activity feeds and links to project media.
It also manages conflicts between editors and takes care of project versioning, so things can be rolled back if things go wrong. Because of how the database works, multiple people can work on the same project at the same time, while using the same footage (i.e. there is no need to duplicate files and lock bins).
Adobe has said that there will be several supported server configurations at launch and that the Hub can be integrated with other media asset management systems using a REST API.
What Makes Adobe Anywhere Different From Other Cloud Video Editing Concepts
I wrote about cloud video editing back in March. What's exciting about Adobe's approach is that they have found a way around many of the issues or compromises that I talked about in my post. Avid has also introduced Interplay Sphere, which has many similarities to Adobe's product.
A big benefit of Adobe's approach is that it uses a standard version of Premiere Pro, so there is no sacrifice in terms of what editors can do, and there is no learning curve.
Media is centralized on site, versus a remote server. In the approaches I was looking at, the underlying assumption was that media would reside in a facility where the user wouldn't easily have direct access to the servers for video ingestion (like Amazon Web Services). The implication was that all video would have to be uploaded over the Internet, and this would create a huge bottleneck due to file sizes and connection speeds. In Adobe's approach, video can be ingested directly at the facility where the server is stored. The only uploads required are when remote editors want to incorporate footage from their own machines. Even this is handled sensibly because the user can work with the local footage while the upload takes place in the background.
There is no need to create proxies. I suggested that working with proxies was one way around the upload bottleneck and challenges of playing back full-resolution footage over the Internet. There are two big issues with proxies. First they take time and computing power to create. Second, they are by definition lower quality, and this creates issues linking to the source footage at the end of a project. In the Adobe Anywhere solution, there is no compromise. You have snappy playback when you need it, and access to the full resolution footage whenever you pause playback.
No recurring storage costs. While I suspect that the monthly recurring cost of using cloud storage like Amazon S3 may be less than the total cost of ownership when using enterprise grade shared storage used by broadcasters, switching to cloud storage for video editing would be a huge paradigm shift. With Adobe Anywhere, there is no need to make this switch in order to take advantage of the benefits of collaboration.
Benefits of Adobe Anywhere
Saves time and increases efficiency. Video production is an inherently collaborative effort. The bottleneck in producing video is no longer software and processing speeds. It's working with others. Time is spent shuttling media back and forth, encoding proxies, conforming video, linking assets, etc. These are all activities that generate zero value. By providing a way for people to collaborate across all phases of a project, Adobe stands to save its customers a significant amount of time. Imagine having someone immediately pick up exactly where you left off on a project without having to do a thing.
Reduces costs. One of the biggest potential cost savings will come from being able to edit on less powerful hardware. There are significant savings if editing can be shifted from Mac Pros to cheap laptops. Other IT savings could be realized by centralizing operations and eliminating duplicate storage.
Other cost savings will come in the form of reduced travel and shipping costs. For example, if a production is located in California and some of the post work is being done by an editor in Toronto, then media is either going to be shipped back and forth, or the editor is going to travel to California to work with other post production staff. Either way, this can get costly.
Project teams can be created without regard for geography. For some, the fear is that this will precipitate the offshoring of jobs overseas to people who will work for pennies on the dollar. For others, this will open up the potential to work with their dream team of highly skilled people without having to worry about the logistics of shipping media and managing travel schedules.
Virtually no learning curve. By using the applications editors know and love, the training costs and organizational change required to adopt Adobe Anywhere will be kept to a minimum.
Simultaneous access to projects. The ability for people to work on the same project at the same time, while managing conflicts, and having the ability to roll-back to previous versions is a big leap forward in terms of collaborative editing.
Security. The centralized database will allow administrators to control what each user can do in a fine grained manner. Files can be locked, people can be removed from projects immediately, and menus can be customized to prevent users from doing things like exporting video. This could be handy when you are passing a project off to a contractor and you want to make sure that they don't export it and post it on YouTube.
Thoughts on the Big Outstanding Questions
While Adobe Anywhere looks very promising, there are more questions than answers at this point. While much more will be revealed between now and the 2013 release date (NAB seems as good of a guess as any), that's not going to stop me from speculating. Maybe Adobe has learned from the Apple machine that rampant speculation only serves to build interest in future products :).
The fxguide interview with Adobe Senior Product Manager Michael Coleman, points out that the initial target market for the service is going to be large broadcasters. Further to this point, the Adobe Anywhere micro site points to the broadcast solutions group for more information. This is the professional services group that helps organizations like the BBC and NBC solve big problems like streaming coverage of the Olympics across multiple platforms.
I think Adobe is positioning itself to win the hearts, minds, and professional services revenue of big broadcast and cable television networks. With a true collaborative workflow tool, and a wide array of tech partners, it can support the entire video life cycle for these organizations: Content creation with Adobe Anywhere; distribution and content protection with Adobe Media Server and Access DRM; and Monetization with Adobe Pass and Digital Marketing Suite.
If Adobe is successful in this market, this could be the move that drives a stake through Avid's heart. Only 12% of Avid's revenue comes from NLE sales, the rest comes from "services and big iron hardware that drives big media production facilities." Any momentum that Adobe gains in this marketplace is going to hurt Avid's already precarious financial situation.
Based on the target market outlined above, I expect that his is going to be a relatively pricey solution that isn't exactly plug-and-play easy (see my thoughts on hardware below).
The demo had a couple of hints that running Adobe Anywhere will take some serious hardware at the centralized location.
First, the demo system was attached to an Omneon Media Grid. It seems reasonable that the demo system would be based on one of Adobe's blessed configurations, and it's certainly not cheap hardware. I'm not an expert in Omneon's products, but they seem like they are directed at large broadcasters and media companies rather than smaller post facilities.
Second, Adobe Anywhere doesn't eliminate the need to render video. It just shifts where rendering is done from a local machine to a centralized one. It stands to reason, that if this centralized machine is going to simultaneously do the rendering for a bunch of notebooks out in the field, then it's going to need some serious processing and GPU horsepower. If 5 people are working on a project simultaneously, it may not necessarily need to be as powerful as 5 stand-alone workstations (not everyone is going to be loading the system by playing video at once), but it's still going to need the muscle to deal with their peak viewing / editing load.
All this is to say that I'm eager to see what the supported configurations are. The hardware required to make this work will really determine who can adopt the platform. Hardware requirements will also figure into the ROI of the service. If the savings from moving from big iron workstations are completely offset by increased spending on powerful servers and shared storage, then the investment in Adobe Anywhere will have to come from harder to quantify (but super important) things like time savings, efficiency, and the ability to coordinate global a workflow (i.e. outsourcing).
Who Adobe Anywhere Could Work For
Although I think Adobe's initial target market is going to be big broadcasters, I think there are several other workflows where the product makes sense.
News and sports production. For things like news and sports, a tool that helps reporters, videographers, and editors work together simultaneously will help get breaking stories to air quicker. Not surprisingly, this is the market that Avid is going after with Interplay Sphere.
Education. Although Adobe hasn't said it, I think another early market for this product is going to be educational institutions. This type of tool that could work very well in classrooms where lots of students are working with the same footage. This would eliminate all kinds of duplicate storage, and could reduce the cost of outfitting media labs with large numbers of high-end workstations. Students could bring their own laptops and edit from there.
Television shot in one location with editing in another. This scenario has peaked the interest of Shane Ross, who is dealing with a production company in Virginia, production in Virginia and post being a mix of Los Angeles and Virginia. Depending on the cost of the system, this could be a good fit. I've spoken with a number of companies in Toronto with a similar type of workflow.
Hopefully this technology will trickle down from big broadcasters to smaller production companies over time.
Over the next 5-years, I think Adobe Anywhere and products of its ilk will transform the post production workflow. It will change the hardware we buy, the people we work with, and where we work. It will also help satisfy the real need that people have to improve how they collaborate on productions. Budgets keep shrinking, time lines are getting tighter, and a better workflow can help editors carve out a little more time to work on telling great stories. Here's hoping that the software makes its way beyond the big broadcast market soon!
Let us know what you think that Adobe Anywhere and its potential to change your workflow in the comments below.