The Hobbit

Peter Jackson’s The Hobbit: An Unexpected Journey was one of the most anticipated films of 2012. It broke new technological boundaries and presented many creative challenges to its editor. After working as a television editor, Jabez Olssen started his own odyssey with Jackson in 2000 as an assistant editor and operator on The Lord of the Rings trilogy. After assisting again on King Kong, he next cut Jackson’s Lovely Bones as the first feature film on which he was the sole editor. The director tapped Olssen again for The Hobbit trilogy, where unlike the Rings trilogy, he will be the sole editor on all three films.

Much like the Rings films, all production for the three Hobbit films was shoot in a single eighteen month stretch. Jackson employed as many as 60 RED Digital Cinema EPIC cameras rigged for stereoscopic acquisition at 48fps – double the standard rate of traditional feature photography. Olssen was editing the first film in parallel with the principal photography phase. He had a very tight schedule that only allowed about five months after the production wrapped to lock the cut and get the film ready for release.

To get The Hobbit out on such an aggressive schedule, Olssen leaned hard on a post production infrastructure built around Avid’s technology, including 13 Media Composers (10 with Nitris DX hardware) and an ISIS 7000 with 128TB of storage. Peter Jackson’s production facilities are located in Wellington, New Zealand, where active fibre channel connections tie Stone Street Studio, Weta Digital, Park Road Post Production and the cutting rooms to the Avid ISIS storage. The three films combined, total 2200 hours (1100 x two eyes) of footage, which is the equivalent of 24 million feet of film. In addition, an Apace active backup solution with 72TB of storage was also installed, which could immediately switch over if ISIS failed.

The editorial team – headed up by first assistant editor Dan Best – consisted of eight assistant editors, including three visual effects editors. According to Olssen, “We mimicked a similar pipeline to a film project. Think of the RED camera .r3d media files as a digital negative. Peter’s facility, Park Road Post Production, functioned as the digital lab. They took the RED media from the set and generated one-light, color-corrected dailies for the editors. 24fps 2D DNxHD36 files were created by dropping every second frame from the files of one ‘eye’ of a stereo recording. For example, we used 24fps timecode with the difference between the 48fps frames being a period instead of a colon. Frame A would be 11.22.21.13 and frame B would be 11:22:21:13. This was a very natural solution for editing and a lot like working with single-field media files on interlaced television projects. The DNxHD files were then delivered to the assistant editors, who synced, subclipped and organized clips into the Avid projects. Since we were all on ISIS shared storage, once they were done, I could access the bins and the footage was ready to edit, even if I were on set. For me, working with RED files was no different than a standard film production.”

Olssen continued, “A big change for the team since the Rings movies is that the Avid systems have become more portable. Plus the fibre channel connection to ISIS allows us to run much longer distances. This enabled me to have a mobile cart on the set with a portable Media Composer system connected to the ISIS storage in the main editing building. In addition, we also had a camper van outfitted as a more comfortable mobile editing room with its own Media Composer; we called it the EMC – ‘Editorial Mobile Command’. So, I could cut on set while Peter was shooting, using the cart and, as needed, use the EMC for some quick screening of edits during a break in production. I was also on location around New Zealand for three months and during that time I cut on a laptop with mirrored media on external drives.”

The main editing room was set up with a full-blown Nitris DX system connected to a 103” plasma screen for Jackson. The original plan was to cut in 2D and then periodically consolidate scenes to conform a stereo version for screening in the Media Composer suite. Instead they took a different approach. Olssen explained, “We didn’t have enough storage to have all three films’ worth of footage loaded as stereo media, but Peter was comfortable cutting the film in 2D. This was equally important, since more theaters displayed this version of the film. Every few weeks, Park Road Post Production would conform a 48fps stereo version so we could screen the cut. They used an SGO Mistika system for the DI, because it could handle the frame rate and had very good stereo adjustment tools. Although you often have to tweak the cuts after you see the film in a stereo screening, I found we had to do far less of that than I’d expected. We were cognizant of stereo-related concerns during editing. It also helped that we could judge a cut straight from the Avid on the 103” plasma, instead of relying on a small TV screen.”

The editorial team was working with what amounted to 24fps high-definition proxy files for stereo 48fps RED .r3d camera masters. Edit decision lists were shared with Weta Digital and Park Road Post Production for visual effects, conform and digital intermediate color correction/finishing at a 2K resolution. Based on these EDLs, each unit would retrieve the specific footage needed from the camera masters, which had been archived onto LTO data tape.

The Hobbit trilogy is a heavy visual effects production, which had Olssen tapping into the Media Composer toolkit. Olssen said, “We started with a lot of low resolution, pre-visualization animations as placeholders for the effects shots. As the real effects started coming in, we would replace the pre-vis footage with the correct effects shots. With the Gollum scenes we were lucky enough to have Andy Serkis in the actual live action footage from set, so they were easy to visualize how the scene would look. But other CG characters, like Azog, were captured separately on a Performance Capture stage. That meant we had to layer separately-shot material into a single shot. We were cutting vertically in the timeline, as well as horizontally. In the early stages, many of the scenes were a patchwork of live action and pre-vis, so I used PIP effects to overlay elements to determine the scene timing. Naturally, I had to do a lot of temp green-screen composites. The dwarves are full-size actors and for many of the scenes, we had to scale them down and reposition them in the shot so we could see how the shots were coming together.”

As with most feature film editors, Jabez Olssen likes to fill out his cut with temporary sound effects and music, so that in-progress screenings feel like a complete film. He continued, “We were lucky to use some of Howard Shore’s music from the Rings films for character themes that tie The Hobbit back into The Lord of the Rings. He wrote some nice ‘Hobbity’ music for those. We couldn’t use too much of it, though, because it was so familiar to us! The sound department at Park Road Post Production uses Avid Pro Tools systems. They also have a Media Composer connected to the same ISIS storage, which enabled the sound editors to screen the cut there. From it, they generated QuickTime files for picture reference and audio files so the sound editors could work locally on their own Pro Tools workstations.”

Audiences are looking forward to the next two films in the series, which means the adventure continues for Jabez Olssen. On such a long term production many editors would be reluctant to update software, but not this time. Olssen concluded, “I actually like to upgrade, because I look forward to the new features. Although, I usually wait a few weeks until everyone knows it’s safe. We ended up on version 6.0 at the end of the first film and are on 6.5 now. Other nonlinear editing software packages are more designed for one-man bands, but Media Composer is really the only software that works for a huge visual effects film. You can’t underestimate how valuable it is to have all of the assistant editors be able to open the same projects and bins. The stability and reliability is the best. It means that we can deliver challenging films like The Hobbit trilogy on a tight post production schedule and know the system won’t let us down.”

Originally written for Avid Technology, Inc.

©2013 Oliver Peters

Autodesk Smoke 2013

Autodesk attracted a lot of attention last year with the revamped version of Smoke for Mac OS X. I had originally been working on a review with the earlier version (Smoke 2012), but held off when I found out Smoke 2013 was just around the corner. Indeed, the more “Mac-like” refresh wowed NAB attendees, but it took until December to come to market. In that time, Autodesk built on the input received from users who tested it during this lengthy public beta period. Now that it’s finally out in the wild, I’ve had a chance to work with the release version, both on my own system, as well as at a client site where Smoke 2013 has been deployed. Both of these are on recent model Mac Pros. Although Smoke 2013 is a very deep application, I would offer that the learning curve for this new version is a mere 25% of what it used to be. That’s a significant improvement.

Getting set up

There are several ways to install and operate Smoke 2013. Most users will install the application in the standalone mode. The software is activated over the internet and works only on that licensed machine. Facility users can also purchase license server software, which allows them to float the Smoke license among several machines. Only one at a time is activated, but any of the machines can run the software, based on the permission assigned by the license server application over the internal LAN.

Smoke 2013 operation is tied to the media storage, so the first thing to do after software installation is to run the Smoke set-up utility. This allocates which drives are accessible to Smoke. You can grab media files from any connected drive, but specific locations must be assigned as library locations for media caches, proxies, render files and so on. These can be internal drives, SAN volumes or externally-connected drives. The key is that when you create or launch a project, it is tied to a specific library location. If that drive is unmounted, any projects associated with it won’t show up and are not accessible (even in an offline mode) to the operator.

You should approach Smoke operation with a media strategy in mind. Smoke 2013 handles more native codecs and file formats – and in a more straightforward fashion – than Smoke 2012. If you are working with ProRes media, for instance, no conversion is necessary to get started in Smoke and files can be rendered as ProResHQ, instead of the previous default of uncompressed DPX files. This means drive performance requirements are less than in the past, but it’s still a good idea to use fast RAID arrays. Even two 7200RPM SATA drives striped as RAID-0 will give you acceptable performance with ProRes media. Naturally, a faster array is even better. Smoke will let you render intermediate proxies for even better performance, but if you want to simply drag in new media from the Mac Finder, then Smoke 2013 now performs on par with other desktop NLEs.

Smoke uses OpenGL and not CUDA or OpenCL acceleration, so performance from ATI or NVIDIA cards is on even footing. If you run a dual monitor system, like my set-up with two 20” Apple Cinema Displays, you can enable dual-screen preview. This will let you mirror the UI or display a selected viewport, which is most often the current clip, but can also be the ConnectFX schematic. You are best off with two 1920×1080 or 1920×1200 screens. The scaling function to reduce the full screen viewer to fit my 1680×1050 resolution introduced artifacts and affected the performance of the card. Smoke 2013 can work with screen resolutions starting at 1440×900, but it’s better to stick to one higher resolution screen like a single 27” or 30” Apple Cinema Display or iMac screen. It’s best to run a broadcast monitor connected to an AJA KONA, IoXT or Blackmagic Design card (in a future version). In that configuration, you can’t use a second computer display to extend the real estate of the Smoke user interface, but could display the UI from another open application, like Adobe Photoshop.

The editing experience

The reaction to Apple Final Cut Pro X kicked up interest in Smoke. Users who wanted a 64-bit, track-based application that didn’t stray too far from FCP 7’s operational style, felt that Smoke 2013 might be the hypothetical “FCP 8”. Autodesk indeed sports an editing workspace that is closely aligned with the look and feel of Final Cut Pro “legacy”, as well as Adobe Premiere Pro. It even defaults to FCP 7 keyboard shortcut commands. If you can edit on Final Cut (before FCP X) or Premiere Pro, then you can be productive on Smoke with little relearning.

The user interface is divided into three panes – a browser, a viewing area and a workspace. Across the bottom are four tabbed interface pages or modes – MediaHub, Conform, Timeline and Tools. MediaHub is where you search drive locations for files. It is analogous to Adobe’s Media Browser within Premiere Pro. Locate files and drag or import them into the editing browser window. Conform lets you reconcile imported media with edit lists and is also a place to relink media files. Timeline is the standard editing workspace and lastly, Tools holds clip tools and utilities, such as deinterlacing, pulldown, etc. Each pane changes the information displayed, based on the context of that mode. In the Timeline mode, you see viewers and a timeline, but in the MediaHub mode each pane shows completely different information.

Editors will spend most of their day in the Timeline mode. This interface page is organized into the standard editing view with player windows at the top and a track-based timeline at the bottom. Smoke always loads at least two timelines – the edited sequence and the selected source clip. Effects can be applied to the source clip, as well as to clips on the timeline. The viewer pane can display clips on a single, toggled viewer (like FCP X) or traditional source/record windows (like FCP 7). There’s also a thumbnail and a triptych view. The latter is helpful during color correction, if you want to display previous/current/next frames for shot matching. The browser displays all imported source clips for a project. It can be placed on the left, on the right or hidden entirely. Within it, clips can be organized into folders. You may have more than one sequence in a project, but only one project can be open at a time. As you select a clip, it immediately loads into the viewer and timeline window. No double-clicking required.

Smoke is a good, fast editor when it comes to making edits and adjusting clips on the timeline. There are some nice touches overlooked on other NLEs. For example, it uses track-based audio editing with keyframable real-time mixing. There are a set of audio filters that can be applied and the output has a built-in limiter. Formatting for deliverables is built into the export presets, so exporting a 1080p/23.976 sequence as 720p is as simple as picking a preset. The edit commands include the standard insert, overwrite and replace functions, but also some newer ones, like append and prepend. Ripple and snapping are simple on/off toggles.

While editing is solid, I would still categorize Smoke 2013 as a finishing tool. You could edit a long-form project from scratch in Smoke, but you certainly wouldn’t want to. It lacks the control needed for narrative long-form, like detailed custom bin columns, a trim tool, multi-camera editing and more. On the other hand, a scripted short-form project, like a TV commercial – especially one requiring Smoke’s visual effects tools – could be edited exclusively within Smoke.

The better approach is to do your rough cuts in another desktop NLE and then send it to Smoke for the remainder. You can import various edit list formats – EDL, XML, FCPXML and AAF. Cut on Final Cut Pro 7/X, Premiere Pro or Media Composer and export an edit decision list in one of these formats for the sequence. Then import and link files in Smoke and you are ready to go. In my testing, XMLs from both FCP 7 and FCP X worked really well, but AAFs from Media Composer were problematic.  Typically Smoke had difficulty in relinking media files when it was an Avid project, most likely due to issues in the AAF.

Come for the effects

The visual effects tools are the big reason most editors would use Smoke 2013 over another NLE. There are four ways to apply effects. The first and easiest is the effects “ribbon” that flies out between the viewers and the timeline. It contains eight standard effects groups – Timewarp, Resize, Text, Color Correction, Spark, Blend, Wipe and Axis. (Spark is the API for third party filters. GenArts Sapphire is the first effects package for Smoke 2013.) The “ribbon” effects are always applied in the same order and some are multiple purpose tools. For instance, the Resize effect is automatically applied for format correction, such as a ProRes4444 clip in a ProResHQ timeline. When these effects are added to a clip on the timeline, a reduced set of parameters appears in a fly-out panel at the top of the timeline. You can immediately apply and adjust effects in the timeline without the need to step deeper until you’ve mastered the simpler methods.

The last effect, Axis, is a “super tool”. It’s the 2.5/3D DVE effect, but you can enter its effects editor and do a whole lot more. Axis lets you add text, lighting, 3D cameras, plus adjust surface properties and surface deformations. Once you enter any of the effects editors, the mode changes and you are in a new user interface specific to the context of that effect. The controls flow left to right and change options according to the selections made. For instance, picking “object” within the Axis effect editor gives you controls to adjust the scale, position and rotation of the clip. Pick “lights” and the control parameters change to those appropriate for lighting.

The third way to build an effect is to select ConnectFX. This brings you into Smoke’s world of node-based composting, where you are presented with a flow chart schematic, a viewer and a set of filter tools. An effect like Color Correction may be applied directly to the timeline as a single filter or as a filter within a ConnectFX build. It’s entirely up to the comfort level of the editor and how many additional effects will be applied to that clip for the final look.

One of available tools within the set is Action, which is a separate compositing method. It forms the fourth way to build effects. You can composite multiple media clips in an Action node, such as a title over a background. Once you step into an Action node, you are presented with its own schematic. Instead of a flowchart, the Action schematic shows parent-child links between layers of the composite, such as a light that is attached to a media clip. Action is where you would make adjustments in 3D camera space. Some tools, like the 3D lens flare effect are only available in Action.

Smoke detractors make a big deal out of the need to render everything. While this is true, I found that a single effect applied from the FX “ribbon” menu to a clip on the timeline will play in real time. If you’ve applied more than one effect to a clip, then usually the last one in this string will be displayed live during playback. When rendering is required, the processing speed is pretty quick. If you export a sequence with unrendered effects, then all effects are first processed (rendered) before the finished, flattened master file is exported.

Conclusion

Smoke 2013 is likely to be one of the deepest, but powerful, editing applications you will ever encounter. It’s deceptively simple to start, but takes a concentrated effort to master the inner workings of its integrated, node-based compositors. Nevertheless, you can start to be productive without having to tackle those until you are ready. In an editing world that’s gravitating towards an ever-growing number of canned, one-button preset effects, Smoke 2013 unabashedly gives you the building blocks needed for that last 5% of finesse, not available from a preset effect. You can even build your own complex presets to be applied on future projects. That takes time and talent to master. Fortunately Autodesk has gone the extra mile with good tutorials available on their Area community site and the Smoke Learning Channel on YouTube.

Smoke is ideal as a finishing tool in a multi-suite facility, the main system in a creative media shop or the go-to system for broadcast promotion production. It is designed to fulfill the “hero” role and is targeted squarely at the Adobe suite of tools. The sales pitch is to stay within Smoke’s integrated environment rather than bounce among several applications. While Smoke 2013 largely meets that objective, it still gets down to personal preferences – compositing in nodes versus a track-based tool like After Effects.

Installation is easier than it was, but I’d still like to see Autodesk improve on the activation process – especially for those interested in using more than one machine. Smoke uses a Unix-style file structure, so project files (other than media index and render files) are hidden from the user. This makes it difficult to move projects from one computer to the next. Smoke 2013 lives up to the commitment made at NAB 2012, but now that it’s a released product, Autodesk has a chance to hone the tool to be more in line with the needs of the target user.

Originally written for Digital Video magazine

©2013 Oliver Peters

Phil Spector

Phil Spector became famous as a music industry icon. The legendary producer, who originated the “wall of sound” production technique of densely-layered arrangements, worked with a wide range of acts, including the Ronettes, the Righteous Brothers and the Beatles. Unfortunately, fame can also have its infamous side. Spector abruptly came back into public notice through the circumstances of the 2003 death of actress Lana Clarkson and his subsequent criminal trials, culminating in a 2009 conviction for second-degree murder.

The story of his first murder trial and the relationship between Spector (Al Pacino) and defense attorney Linda Kenney Baden (Helen Mirren) form the basis for the new film by HBO Films. Phil Spector, which is executive produced by Barry Levinson (Rain Man), was directed by celebrated screenwriter/director David Mamet (The Unit, The Shield, Hannibal, Wag the Dog). Rather than treat it as a biopic or news story, Mamet chose to take a fictionalized approach that chronicles Spector’s legal troubles as a fall from grace.

One key member of the production team was editor Barbara Tulliver (Too Big to Fail, Lady in the Water, Signs), who has previously collaborated with Mamet. She started as a film editor working on commercials in New York, but quickly transitioned into features. According to Tulliver, “I assisted on David’s first two films and then cut my first feature as an editor with him, so we have established a relationship. I also cut Too Big to Fail for HBO and brought a lot of the same editorial crew for this one, so it was like a big family.”

As with most television schedules, Phil Spector was shot and completed in a time frame and with a budget more akin to a well-funded independent feature, rather than a typical studio film. Tulliver explained, “Our schedule to complete this film was between that of a standard TV project and a feature. If a studio film has six weeks to complete a mix, a film like this would have three. The steps are the same, but the schedule is shrunk. I was cutting during the thirty-day production phase, so I had a cut ready for David a week after he wrapped. HBO likes to see things early, so David had his initial cut done after five weeks, instead of the typical ten-week time frame. Like any studio, HBO will give us notes, but they are very respectful of the filmmakers, which is why they can attract the caliber of talent that they do for these films. At that point we went into a bit of a hold, because David wanted some additional photography and that took awhile until HBO approved it.”

The production itself was handled like a film shoot using ARRI Alexa cameras in a single-camera style. An on-set DIT generated the dailies used for the edit. Although you wouldn’t consider this a visual effects film, it still had its share of shots. Tulliver said, “There were a lot of comps that are meat-and-potatoes effects these days. For instance, the film was shot in New York, so in scenes when Spector arrives at the courthouse in Los Angeles, the visual effects department had to build up all of the exteriors to look like LA. There are a number of TV and computer screens, which were all completed in post. Plus a certain amount of frame clean-ups, like removing unwanted elements from a shot.”

Mamet wrote a very lean screenplay, so the length of the cut didn’t present any creative challenges for Tulliver. She continued, “David’s scripts are beautifully crafted, so there was no need to re-arrange scenes. We might have deleted one scene. David makes decisions quickly and doesn’t overshoot. Like any director, he is open to changes in performance; but, the actors have such respect for his script, that there isn’t a lot of embellishment that might pose editing challenges in another film. Naturally with a cast like this, the performances were all good. The main challenge we had, was to find ways to integrate Spector’s songs into the story. How to use the music to open up scenes in the film and add montages. This meant all of the songs had to be cleared. We were largely successful, except with John Lennon’s Imagine, where Yoko Ono had the final say. Although she was open to our using the song, ultimately she and David couldn’t agree to how it would be integrated creatively into the film.”

Phil Spector was cut digitally on an Avid Media Composer. Like many feature editors, Barbara Tulliver started her career cutting film. She said, “I’m one of the last editors to embrace digital editing. I went into it kicking and screaming, but so did the directors I was working with at the time. When I finally moved over to Avid, they were pretty well established as the dominant nonlinear edit system for films. I do miss some things about editing on film, though. There’s a tactile sense of the film that’s almost romantic. Because it takes longer to make changes, film editing is more reflective. You talk about it more and often in the course of these discussions, you discover better solutions than if you simply tried a lot of variations. In the film days, you talked about the dramatic and emotional impact of these options. This is still the case, but one has to be more vigilant about making that happen – as opposed to just re-cutting a scene twenty different ways, because it is easy and fast – and then not know what you are looking at anymore.”

“Today, I cut the same way I did when I was cutting film. I like to lay out my cut as a road map. I’ll build it rough to get a sense of the whole scene, rather than finesse each single cut as I go. After I’ve built the scene that way, I’ll go back and tweak and trim to fine-tune the cut. Digital editing for me is not all about the bells-and-whistles. I don’t use some of the Avid features, like multi-camera editing or Script Sync. While these are great features, some are labor-intensive to prepare. When you have a minimal crew without a lot of assistants, I prefer to work in a more straightforward fashion.”

Tulliver concluded with this thought, “Although I may be nostalgic about the days of film editing, it would be a complete nightmare to go back to that. In fact, several years ago one director was interested in trying it, so I investigated what it would take. It’s hard to find the gear anymore and when you do, it hasn’t been properly maintained, because no one has been using it. Not to mention finding mag stripe and other materials that you would need. The list of people and labs that actually know how to handle a complete film project is getting smaller each year, so going back would just about be impossible. While film might not be dead as a production medium, it has passed that point in post.”

Originally written for Digital Video magazine.

©2013 Oliver Peters

Final Cut Pro X versus Premiere Pro CS6

The struggle within most shops that invested in Apple’s Final Cut Pro is whether to stay put a while longer, adopt Final Cut Pro X or cut the cord and move on. For many this means shifting to the Adobe Production Premium bundle – part of Creative Suite. Most of the editors and facilities in my sphere are doing just that. I’m one of only two local editors that I know of, who is actually using FCP X professionally. The rest are in the process of shifting to Premiere Pro, while maintaining some continued use of FCP “legacy”. This post is not intended as a “shoot out” or to say one is better than the other. Both are good tools and much of the choice gets down to personal preference. Instead, my goal is to lay out some random considerations in making the move.

Cross-platform / performance

Premiere Pro runs on Mac OS X and Windows workstations and laptops, while Final Cut Pro X is a Mac-only product. The biggest consideration is that by having the tool available to Windows, you open your access to the fastest machines and GPU cards. Premiere on a PC can tap into the faster NVIDIA CUDA-enabled cards, which is not an option for either Premiere or FCP X on the Mac. Although Premiere runs with both CUDA and non-CUDA cards on Macs, the selection is limited.

Adobe’s standalone software must be purchased with either a Mac or Windows license and switching platforms requires cross-grading the license. Unlike Avid, you cannot simply go from a PC workstation at a facility to a MacBook Pro at home with a simple de-activation/re-activation process. The exception is the Creative Cloud subscription, with permits access to both Mac and Windows licenses on up to two machines, as long as they aren’t used at the same time.

Naturally, if you opt for Final Cut Pro X, you have software that has been tweaked for the most current Apple hardware. We can argue the merits of CUDA, OpenGL and OpenCL acceleration, but it’s pretty clear that FCP X running on a decked-out iMac outperforms the application on a Mac Pro tower.

Suite versus “suite”

Premiere Pro is generally purchased as part of the Production Premium or Master Collection software bundles – or as part of a Creative Cloud subscription. Final Cut Pro X is only available as standalone software through the Mac App Store. The beauty of the Adobe software is its integration, with direct links between Premiere Pro and After Effects, Prelude, Audition or SpeedGrade. These aren’t all fully developed yet, but it’s a key reason some editors prefer Premiere Pro.

On the other hand, there’s a large ecosystem growing up around Final Cut Pro X that constitutes much of the same. It’s not an official “suite” of software and interoperability is limited to translations of FCPXML. For similar dollars, you get similar capabilities – only with the added ability to pick and choose what’s right for your workflow.

Plug-ins

The plug-in architecture for Premiere Pro has historically been weak. Many of the third-party After Effects plug-ins show up and work within Premiere Pro, but some don’t. If you edit in Premiere, you are best off doing your effects in After Effects. Lately, developers have been tweaking their filters to make them work – or work better – inside Premiere.

To compare, Final Cut has no plug-in architecture. Instead third-party plug-ins use FxPlug through Motion and then show up inside FCP X as a Motion template, rather than a traditional plug-in. This allows developers to not only create updated plug-ins for Motion and FCP X, but also add new and unique effects and transitions built strictly as Motion projects. These in turn are published to FCP X as effects. Since this latter approach requires less programming skill, the market for low-cost (and even free) FCP X plug-ins has exploded. Not to mention, there are effects and transitions for FCP X that simply don’t exist – or can’t easily be re-created - for any other NLE.

Organizational tools

All NLEs are giant databases tracking information. Final Cut Pro X takes this to a new level and uses ratings, keywords and smart collections as a means for fast and automatic organization of your media. Plus a considerable amount of camera and textual metadata is tracked in the background. This doesn’t mean that Premiere Pro doesn’t track a lot of data, as well. Open the metadata display window and you find plenty of fields that are assignable to each clip. Bins can be filtered by a search field, which will reduce the amount of clips displayed according to the search criteria being typed in.

User interface configurations

Final Cut Pro X’s interface design is based on panels and windows that can be opened and closed as needed. It is arranged well for single and dual-screen layouts, though you have very few options to move any of these windows around and create custom screen layouts. Premiere Pro uses a system of dockable tabs common across several of Adobe’s applications, including After Effects and Photoshop. Re-arrange these as you see fit and save custom workspace layouts.

Tracks versus trackless

Premiere Pro uses the “traditional” track-based timeline structure, where audio and video is separated into tracks and clips are positioned on the timeline based on a reference to absolute time. Final Cut Pro X’s timeline does not use tracks, but instead lays out clips according to storylines and connected clips. These are linked to each other in a parent-child relationship. This allows groups of clips to be moved, by simply moving the clip on the storyline to which the others are attached. There is no vertical hierarchy to audio and video content as tracks. Although video is displayed to the viewer from the top down, audio and video connected clips can be linked above or below the central primary storyline.

Project and clip management

Premiere Pro creates a single, self-contained data file for every edit project. This file contains the links to all media on your hard drives and the edited sequences created from these. Final Cut Pro X divides its structure into Events (source media) and Projects (edited sequences). These correspond to separate folders on your hard drive as well as divisions within the FCP X interface. Event folders can contain either actual source media content – or alias files pointing to other locations on your hard drives for that source media.

Media management

At the end of a production, many editors like to organize the final edited sequence and the clips used within it into a single “consolidated” project. This means the source clips have been trimmed to only the portions used, plus a few seconds of “handles” on the ends of the trimmed clips. Premiere Pro allows you to do this via its Project Manager tool. FCP X currently does not allow any clip trimming. You can copy a Project (edited sequence) with its used clips to a new Event, but then it requires a second step to organize the media. That step copies the media itself for all used clips into the new Event that was created.

Multiple editor interaction

Right now, neither tool is very good for collaborative editing. Final Cut Pro X works best to have all Event and Project folders at the root level of drives and only one editor can access those at a time. There is an “add SAN location” feature for shared storage environments, but it doesn’t appear to work with all SANs. The best method is to have media on a SAN, but keep the Event and Project files local to each system, with the media linked to these. If one or more editors is working on the same production, then each can have local, “mirrored” versions of the Event folders. To exchange edited sequences, simply copy and transfer the Project files that you’d like to share.

In the case of Premiere Pro, the current workflow is similar to that of FCP 7. It will likely change after NAB, where Adobe is expected to show Adobe Anywhere as a real product and its entry into collaborative editing. Currently, if multiple editors work on the same Premiere-based production, media can be on a SAN, but the project files should be on local drives. Unfortunately, you cannot open multiple project files at once. When you import another editor’s sequence into your project, it annoyingly imports all the associated master clips, even through they may already exist within your project. These cannot be removed, otherwise clips in your imported sequence will go offline.

[EDIT - My Premiere Pro import issues were challenged by a reader, so I went back and did some testing. It appears that if two editors create two unique projects, but using the same media (like in a shared storage facility), then duplicate master clips are created upon import. However, if the second project is created using a "save as" command, then sequences imported from it back into the original project do not create duplicate master clips.]

List interchange

Final Cut Pro X only interchanges data with external applications using the FCPXML data format. This is different than other versions of XML, which means you have to use translation to get from FCP X to FCP 7, for example. Premiere Pro supports XML, EDL, OMF and AAF (limited).

Tape handling (or not)

Neither application is great for videotape-based workflows. Premiere Pro has capture-from-tape and output modules, but it’s not as solid as FCP 7 and definitely not as good as Avid Media Composer. On the other hand, FCP X’s is non-existent. There is limited support for Firewire-enabled videotape decks, like HDV, but you really end up using the capture/output utilities of the third-party hardware cards (AJA, Blackmagic Design, Matrox, MOTU).

The Cold Mountain moment

Feature film editors’ use of specific NLEs does not amount to a large market segment for either of these manufacturers (nor Avid, for that matter, either). But the association with a Hollywood blockbuster fuels aspirational marketing in other sectors. It wasn’t until Walter Murch cut Cold Mountain – along with the Coen Brothers’ use of FCP – that Final Cut started to get noticed by a large portion of the professional editing community as a viable tool. Neither Apple nor Adobe have had that yet with Premiere Pro or FCP X. There has not been a major feature film cut with either.

Adobe is a bit closer in that many films have touched on Premiere Pro as a conduit to get into After Effects or handling some conforming tasks. Naturally, Adobe is more than happy to let the lines be blurred through omission between these roles and doing an actual creative edit of a film. That is likely to change this year. This is purely a guess, but I wouldn’t be surprised to see a studio film being cut on FCP X or Premiere Pro at some point during 2013.

Pricing and openness

For now, when you compare cost, FCP X is the better deal. Even when you add more applications and utilities to fill in the gaps, the cost is lower than Premiere Pro. Add to that the fact that Mac App Store purchases can run on multiple machines under a single Apple ID, thus bringing the cost per machine even lower. Adobe is in the process of changing its sales and licensing structure through the Creative Cloud mechanism, but the limit is typically two machines. The beauty here is that you have access to the whole host of Adobe creative applications for video, photography, web and design as part of a monthly subscription model.

Over the course of the first year on one machine, cost is probably going to be similar for most users. If you use multiple machines, own the software for several years or use more than the average number of applications, then the scale will tip in favor of one company or the other. Both the Mac App Store and Creative Cloud models allow for more rapid updating of software than in previous years.

Openness and company response may also be a factor in your purchasing decisions. Apple is secretive about new product development. They do listen to customer feedback, but they don’t show it publicly. Adobe has tried to be very proactive in their outreach to the professional creative community. In the end, the net result may be the same in how it translates into new features that you can use.

Interesting tools

Both NLEs offer tools beyond just media organization and editing. For instance, both include stabilization, but Premiere Pro includes extra touches to fix rolling shutter artifacts. FCP X includes optical flow processing for high-quality variable speed effects. Final Cut features a number of non-destructive “automatics” for image and sound analysis on ingest and shape/shot recognition. Premiere Pro offers speech-to-text analysis. A lot of these tools fall into the “nice to have, but I never use it” category for me. Still, if these are worthwhile for you, then take a closer look.

Batch exports

Neither tool offers good batch export tools like I’m used to with FCP 7, however, each offers queued exporting functions of edited sequences. In Premiere Pro, if I want to export multiple sequences, or the same sequence as multiple deliverables, then all I need to do is set up a queue in Adobe Media Encoder. With Final Cut Pro X, I can use the Share menu to export straight from the timeline or send Projects to Compressor. Unfortunately neither one lets me export a QuickTime reference file that can be used in other encoders. You first have to export a self-contained master file if you intend to use it with other software.

Responsiveness

This is a big one for me and a good place to end this list. In its current form, FCP X feels a tad buggy to me. Response is generally better on an iMac. I’m mainly on Mac Pros, so playback often just doesn’t look smooth. It’s definitely not dropping frames, but looks like the graphics card (usually an ATI 5870) simply isn’t refreshing as well as it should. The interface also tends to feel “sticky” as I’m going between windows. It “forgets” where it is during skimming when switching between a clip in the Event browser and the timeline. Then it takes a bit of clicking around in the interface to get it to “wake up”.

There seems to be some type of RAM leak issue,too. The longer I work on it in a day with large Events (bins), the more sluggish it becomes. This requires me to close and relaunch the application to get peppier performance.

I don’t see any of this with Premiere Pro. I do miss the skimming features of FCP X (no, hover scrub is NOT the same), but otherwise, the Premiere Pro user interface interaction seems to be better for now. I’d say for me, this is an annoyance and not a deal-breaker, but it definitely needs to be addressed by Apple.

©2013 Oliver Peters

PluralEyes 3

The concept of synchronizing clips by sound seems so obvious in retrospect, but when Bruce Sharpe showed his first version of PluralEyes at a small NAB booth, it struck many as nothing short of magic. The first version was designed to sync multiple consumer and prosumer video cameras by aligning their sound tracks in the absence of recorded timecode. With the unanticipated popularity of the HDSLR cameras, like the Canon EOS 5D Mark II in late 2009, PluralEyes gained a big boost. It became the easiest way to sync 5D clips with double-system audio recorded using low-cost devices, such as the Zoom H4n handheld digital audio recorder. PluralEyes expanded from a plug-in for Final Cut Pro to add the standalone DualEyes, used to sync double-system sound projects. In a very short time period, PluralEyes went from an unknown to a brand name synonymous with a product or process, much like Coke or Kleenex.

Now that Sharpe’s Singular Software products are part of the Red Giant Software family, PluralEyes is available as the new and improved, standalone PluralEyes 3 (currently in version 3.1). It encompasses all of the features of both the original PluralEyes and of DualEyes. This means that PluralEyes 3 supports two basic processes: a) synchronizing camera files with external audio, and b) synchronizing multiple cameras to each other or to a common sound track. This is all done by comparing the audio tracks against each other without the use of timecode, clapsticks or other common reference points.

PluralEyes 3 analyzes and matches audio waveform shapes to accomplish this, so without belaboring the obvious, all camera files have to include an audio track recorded in the same general environment. Since PluralEyes uses very good audio analysis tools and audio normalization to aid the process, the camera audio does not have to be pristine. The most common scenario is a high-quality audio recording as a separate digital audio file and camera audio that was recorded solely with the onboard mic. Naturally the cleaner this onboard recording is, the more likely that synchronization will be successful.

The new features of PluralEyes 3 include a brand new user interface, faster synchronization, NLE round-tripping support (Apple Final Cut Pro, Final Cut Pro X, Avid Media Composer and Adobe Premiere Pro) and direct exporting of new, synchronized media files. To synchronize double-system projects, simply drag your camera files into the interface’s camera section and the audio tracks into the audio section. PluralEyes 3 lets you create multiple bins as tabs across the top of the interface for use in organizing your files. For instance, you might want a separate bin for each camera or shoot date or location.

As you add the camera and audio clips to these sections, they will be lined up in ascending order within the lower timeline window. Once the timeline is filled, click “synchronize” and watch PluralEyes 3 do its magic. If the audio recording is low, you can opt to level the audio (normalization) during this process. That will make it easier for successful matching, but it’s an extra step, so the total synchronizing process will take a little longer. Part of PluralEyes 3’s new interface is a 2-up view, which makes it possible to see how the audio tracks align. This view will aid you in adjusting sync if needed.

When synchronization is complete, PluralEyes 3 offers several export options. If you are sending these files to Premiere Pro, Final Cut Pro or Final Cut Pro X, simply export the appropriate XML version. You can choose to replace the camera audio tracks with the audio file’s track as part of this step. Then import that XML into the NLE you selected. When I ran this test with FCP X, the export options let me send two new Events (synchronized clips plus synchronized clips with replaced audio) and a new sequence (Project) representing the PluralEyes timeline. This timeline had both sets of audio channels turned on, so you’ll have to mute the camera tracks first if you intend to use this timeline.

A new feature is the ability to export new media files. For instance, if you want new clips where the high-quality audio has replaced the camera’s reference track, PluralEyes 3 will export these and write new media files. The advantage is that this approach is independent of your NLE choice, making the self-contained, synchronized files easy to migrate between systems.

PluralEyes 3 can also sync multiple cameras for a multi-camera edit session. First, start in the NLE by building a timeline with the clips for each camera placed on a separate video track. Video 1 = camera 1, video 2 = camera 2 and so on. Multiple broken clips from the same camera angle should be placed back-to-back on the same track. In the case of FCP X, group multiple clips from the same camera into a single secondary storyline, before proceeding to the next camera. Once you are done, export an XML file for that sequence. For Avid Media Composer projects, export an AAF file with the media linked and not embedded.

The XML or AAF file is then imported into PluralEyes 3. You’ll end up with a timeline that is populated with the different camera angles corresponding to your NLE sequence. Next, click “synchronize” and watch as PluralEyes realigns the camera clips by referencing the sound tracks against each other. The 2-up view is handy to compare two cameras (as well as their audio tracks) against each other, in case you have any question regarding their synchronization. Once this process is done, export a new XML or AAF from PluralEyes. Import that file into the NLE and you will have a timeline with camera clips rearranged in sync. This would represent what editors typically call a “sync map”. In the case of FCP X, the PluralEyes 3 export settings offer the option of exporting new events, as well as multicam clips. These can be used in FCP X’s standard multicam editing workflow. Open the FCP X angle viewer for access to editing between camera angles.

Red Giant’s PluralEyes 3 is a major advance over the original concept. It’s no longer tied to a single NLE, but is useful both in standalone and NLE-specific workflows. As editors deal with an ever-increasing, diverse spectrum of media sources, a tool like PluralEyes is an essential part of the kit. It was a no-brainer on day one, but even more so in this new and improved version.

Originally written for DV magazine / Creative Planet Network

©2013 Oliver Peters