Workflow From Home: Episode 12 – Behind the Scenes
We’d like to start by thanking all of you for taking the time to watch these shows, or read these articles, as we’ve explored remote workflows for media and entertainment professionals.
Also, thank you for sharing your success stories, and for all of your enormously valuable feedback and questions.
We’re all completely floored by your encouragement, and gratified that you’ve shown us how you’re learning from the series by taking some of the concepts we’ve discussed and putting them into action for your own workflows.
So far, we’ve investigated and unpacked 11 different segments of production and post-production processes, and have heard from industry experts who’ve successfully deployed custom workflows that have empowered them to keep delivering new content while working with distributed teams—often with virtually the same level of productivity.
But the story we haven’t told yet is how an end-to-end remote production can be created, and since you’ve asked for an analysis on how a distributed team can use all the technology we’ve discussed together so far, we thought we’d give you a proof of concept.
So here, in Episode 12, we’re turning the cameras on ourselves to show you how our team at Frame.io has applied everything we’ve learned together, and pushed our own technology to experiment with a few new workflows that we hadn’t tried before.
This episode serves as a snapshot of some of the ways in which crews, producers, directors, editors, and post facilities can increase efficiencies no matter where they’re working from.
We didn’t fully grasp the totality of the situation, but our last day in our offices was Friday, March 13th. Several weeks prior to this, our CEO, Emery Wells, had anticipated stay-at-home orders and began migrating the entire company to a distributed-teams workflow—with every employee working from home.
On March 18th, Emery released a public statement about our plan to work remotely and to offer the Frame.io platform to anyone who needed it. Hundreds of individuals and companies were able to join Frame.io —completely free of charge—and existing customers were given increases in collaboration seats, as well as additional cloud storage.
Then, our leadership team thought that in addition to sharing our technology, we should also share our knowledge and experience as it relates to remote workflows. So we quickly decided that the best way to do this would be to put together a series of informational programs.
We assembled a production team, began outlining topics to cover, and on March 23rd, a mere 10 days after we all said goodbye to each other in the office, we released the first episode of “Workflow From Home.” Our initial episode definitely had a kind of garage-band feel, but as we continued to produce new episodes, we continued to improve our process and our final product.
There are a lot of contributors to the “Workflow From Home” series across many departments at Frame.io, but our core production team is made up of eight principal contributors working from home in different parts of the country, each representing a different facet of the production.
Every single day, we get together on Zoom for a face-to-face update on the status of the current episode, brainstorm topics and ideas for upcoming episodes, and address details for upcoming releases.
Part of what makes “Workflow From Home” special is that when we started this series, we only had two episodes clearly in mind. We didn’t know how many episodes we were going to make, what topics we would eventually cover, or what order they would go in.
That’s why we relied a lot on you, the front line pros and experts following this series, to help guide us forward based on your feedback and suggestions.
But we also used data to support how we proceeded. I can’t stress enough how valuable data can be. Emily Kalen, based in NY, works on our growth team in data and analytics.
First, she helped us conduct a couple of industry surveys to see what the community was feeling. Then, using a methodical approach to how and where we rolled out episodes, she was able to capture information from viewers who watched the series and interacted with the Frame.io platform. Similar to broadcast ratings’ analytics, those findings helped us identify topics to explore and sequence them from week to week.
Those insights were key to helping us stay focused on what the industry was telling us during a period in which timing was important.
It’s important to note that everyone on our “Workflow From Home” team also has a full plate of work outside of this production, so in order to keep pace with a weekly series, we had to set up a killer workflow and a production schedule in order to make our deadlines.
What we eventually refined into a well-oiled machine goes something like this: I shoot my on-camera portions on Sundays; we have a parallel track of offline editing and animation creation that starts on Monday; during the week we write the blog entry for the current episode and scripts for the next episode; we do the editorial turnover on Fridays; we perform our online, color, and sound pass over the weekends; we QC and make any final tweaks on Mondays; and we release the episodes online on Tuesday.
Episodes average about 12-15 minutes each, with more than 60 percent of the screen time filled with custom animations and graphics. What’s important here is that because we were in different time zones and adding this series to our daily workload, we really needed to leverage an optimized cloud-based workflow to keep us on track.
Writing the Script
When it comes to the best writing tool for speed and collaboration, we use the application Quip to act as our “show bible.”
Lisa McNamara, who’s based in Oregon, starts writing in a shared Quip doc that 23 total people at Frame.io can access. There, we can not only all collaborate and make notes, we can also tag each other.
What we like about Quip is that it automatically builds an outline which makes it easy to sort, search, and jump between episodes.
Graphics, stills, and other references can be inserted directly into the Quip doc so everyone can get a sense of the direction of each episode, and of the series as a whole.
By keeping the entire series in one document, everyone is able to collaborate—we can adjust the order of episodes, search for key concepts, and keep tabs on progress without having to manage independent or local documents.
Shooting the episode
Shooting is a subject a lot of people were asking about a couple of months back—especially those who were creating their own shows or conducting interviews from home. So we did a deep dive into tips and tricks for recording yourself (or a remote interviewee) and dedicated all of Episode 6 to examining this workflow.
You can watch the full episode to get the specifics on all the techniques we outline, but briefly, we shoot my segments of the shows in 8K on a RED Helium at 9:1 compression and a ProRes sidecar in Log3G10.
One of the things we’ve been working on at Frame.io is getting editable proxy files directly to the cloud, so I’ve actually been testing a prototype of that workflow for this series, and it means that my shots are instantly and automatically delivered to the editor—an industry first.
We’re using gamma encoded files so our editor immediately receives 1080p, 23.98 proxy files in the Frame.io cloud for the offline, that will eventually be matched back to the original 8K RAW files for the online.
Vince Masciale is a freelance editor and director based in Los Angeles, who came in to help us by dedicating his time to editing this series. We decided to use our own hybrid-cloud workflow and chose Final Cut Pro X as the NLE because it’s able to do a few things no other NLE can.
I’ve been pushing the benefits of camera-generated proxy workflows since 2015 with the Panavision DXL and DXL2 cameras, and FCP X is optimized to edit natively with H.264 and H.265 files—essentially, my camera video output is directly connected to Vince’s Final Cut computer.
The Frame.io extension lives within FCP X so every time a new asset is generated, Vince gets a notification and can instantly preview and pull that asset into his project Library for the most seamless hybrid workflow.
What we’ve demonstrated is that as a shot is being captured, it’s also being encoded and sent to Frame.io, where the Final Cut extension automatically delivers it to Vince.
Because FCP X is exceptional with intra-frame and GOP codecs, the H.264 files we’ve captured look, sound, and act like ProRes or DNx files, which means this workflow allows us to shoot and stream directly into a timeline.
If you watch the video of this episode, you’ll see how we’ve connected a camera, automatically, in real time, to the Frame.io cloud where it went immediately into the cutting room, and was ready for editing.
For all the questions and concerns this workflow may invite, once you start collaborating this way, you won’t go back. It’s why cloud-camera proxy streams will be how everyone will work in a few years.
As soon as Vince gets the footage, he starts building a “radio edit” in FCP and identifies through title cards the name and description of an animation. The FCP Frame.io Integration allows Vince to export and upload edits directly into Frame.io in a single process.
Once the file is online, we use a powerful form of organization through Range-based tagging. Vince navigates to each element in the episode, and Tags them using frame-accurate Range-based comments along with the type of element using hash-tags—such as animation or graphics. Vince also tags a person so they get a unique notification for their assignment.
What’s significant about this is that we don’t need to keep a separate spreadsheet for the external elements, because Frame.io organizes everything just as a spreadsheet would—only it has more functionality and flexibility.
By working this way, we can see how many elements are in this episode as evidenced by total comments; we can sort by the type of element that needs to be created; we can export a table as a database in a CSV; or we can share a PDF with notes, timecodes, and thumbnails.
This turns Frame.io into a database with a sound and video reference track for every contributor to know exactly what needs to be done without sending a single email.
This technique was entirely new to me—I’d never worked this way previously—and it became the core reason we could expedite so many simultaneous assignments without tripping up.
What’s more, it wasn’t until we wrote the script for this workflow analysis that we realized we never sent a single email in making this series.
Imagine a multi-department series that is constantly up to date with specific assignments, assets, review and revisions, all managed without email. We even onboarded several freelancers throughout the series to help with the workload—and they were instantly able to navigate this fluid process and become immediately productive.
Because different people learn in different ways, a key part of each episode is creating the animations that serve as visual diagrams for the sometimes complicated concepts we’ve addressed.
The first step in this process is that our NY-based animator, Abbie Bacilla, reads the script in the Quip doc so she can start thinking through what might best be served by animation in each episode.
What’s great about doing this in Frame.io is that we start by creating a Team for the “Workflow From Home” series. Then we create individual Projects for each episode.
Within every Project, we set up the same folder structure to keep things organized and compartmentalized. We have folders that contain the Dailies, Music, Elements, Edits, OCF, and Masters. Within the Elements folder we have sub-folders, including a directory for Abbie’s Animations.
From there, Abbie creates sketches and storyboards, and she and Vince will go back and forth as she works in After Effects, which also integrates directly with Frame.io.
Next, she’ll create a rough animatic at 1080 H.264, and give that to Vince to drop into the cut. As the animations become more defined and refined, she’ll drop them into the folder and Vince will get a notification that there’s a new version to drop into the cut.
Conversely, if anyone leaves a comment on an animation, Abbie will get a notification. It’s all completely seamless and iterative, and there’s never any confusion about what the latest version of an animation is and what’s in the cut.
When all animation notes have been addressed, Abbie renders the final animations at 4K ProRes for the online.
Feedback is always an essential part of the editorial process, and it can sometimes be the hardest part—especially for the offline editor.
For one thing, it’s difficult when you’ve spent a lot of time on a cut, only to get a ton of notes to address from several different people. The notes can be anywhere from confusing to vague to contradictory, and trying to sort through them and address them in a way that makes everyone happy can be challenging.
As we explained in Episode 11, it’s why Frame.io was created in the first place.
It was specifically designed to allow anyone who needed to give feedback on a cut the ability to do it directly on the video, frame accurately, which eliminates any ambiguity in communicating exactly what you want the editor or colorist, or animator, or VFX artist to do in terms of changes or adjustments.
Some people have challenged us that it’s harder to leave feedback on a video than it is to just talk about what’s working or not working for them. In Episode 10 we heard what the colorists at Light Iron thought about synchronous versus asynchronous review, and how commenting worked efficiently for them.
When it came to commenting in this series, through the first 11 episodes, our small team left close to 1500 comments. That may sound like a recipe for drowning an editor in notes, but Vince, who was initially concerned about this process, later shared that this fluid approach of centralized asynchronous, spontaneous commenting actually ended up making him more productive.
First, Vince didn’t have the lag time of waiting for notes to be written, organized, and then sent by individual parties. Second, the Frame.io extension in FCP has a feature called Presence, which we first showed in Episode 11. It allows the editor to monitor the status of external reviews of the edit, in real time, which opens up a whole new way of communicating.
For example, when our NY-based producer, Andrew Gomez, was in the cut and leaving comments, Vince could see them as Andrew was writing them.
What Vince discovered is that having the notes right there in the edit as we were leaving them made them easier to understand and, what’s more significant, because everyone could see each other’s notes, we could arrive at a consensus and solution in less time that everyone was happy with—and then move on.
It’s basically a new kind of workflow, and one that all of us on this series have said will change the way we work from here on out. And it builds upon what the Light Iron colorists said in Episode 10 and Zoic’s Andrew Orloff said in Episode 8.
What we’re describing here is not exactly synchronous collaboration, but it’s not exactly asynchronous either—because we’re able to capture a synchronous moment in an asynchronous format. Final Cut’s Frame.io Presence extension is essentially inventing a new kind of workflow that is semi-synchronous.
In other words, semi-sync has the benefits of documented and timecode- specific notes within a cut, but delivered in a more realtime, synchronous experience. Vince was able to manage and address the comments more easily—and in some cases, instantly—without having to track anything manually.
Semi-sync is something entirely new that cloud-based workflows just enabled us to create.
Online and the Frame.io Transfer App
Another key reason for choosing FCP as our NLE was because it’s so easy to go from offline to final conform. This meant we didn’t even have to wait for a locked cut to begin our online.
The FCP database allows me to relink the H.264 edit files back to the original Helium files, which we take into Resolve for a color correction pass on the live action interviews. Then, that round-trips back into FCP where it’s all re-conformed in 4K ProRes HQ 3840 x 2160. I use the new Frame.io Transfer app, uploading an XML from the edit and downloading all the elements I’ll need from the Frame.io cloud for the final online to my G-Speed Shuttle.
This allows me to work in a hybrid cloud workflow, limiting only the relevant media to live on my drive.
After that, I use the Logic engine within FCP to do audio sweetening and the final mix.
Since many of the interviews of our guest industry experts are shot on their iPhones in their living rooms or home offices, the audio can be compromised. Because FCP has best-in-class audio mixing features, this important aspect brought another significant benefit to the offline and online process within FCP.
The next step in our process is to have “Eagle Eye” Ben Bailey in New Jersey perform the final QC. And here’s the thing: no matter how careful we are or how many of us have gone through the cut, Ben always finds something. I highly recommend getting an Eagle Eye of your own. Since we didn’t go through a normal QC house for this series delivery, Ben served this important role.
Once the video is finished, we prepare to launch it into the wild.
Because different parts of our audience consume content in different ways, we wanted to support individual preferences. Lisa and Ben create a written version for people who like to read (like this one), which is posted on our blog along with guides and tutorials like how to set up a remote Resolve grading session in Episode 9. Other people may like to binge so we created our series micro-site. And some people like to share videos within their network so we published each episode on YouTube, as well.
Every week, just prior to launch, Megan Stanford in Phoenix, Arizona, puts together a team launch document to get us all organized. She makes sure that we know what all the deliverables are, where they’ll appear, what time they’ll launch, builds and distributes hyperlinks, and manages distributed responsibilities. She also makes sure that everyone in the company is aware of the launches and where they can find the content so they can help spread the word.
We also produce short promos for each episode for our various social channels, which brings us back to the value of gathering data. Because we have solid analytics, we’re able to strategize our distribution and make sure that the right audience is discovering the right content. Then, based on each week’s performance, we can more flexibly strategize what topics we should cover next, which inspires the birth of a new episode topic, and the cycle repeats.
As we touched on earlier, we were actually floored when we started to see the numbers coming in around how many of you were watching the series, sharing it, and most importantly, really using this information. So far, over the 11 episodes of this series, we’ve had over half a million visits to the frame.io/remotework landing page.
But here’s what I find most gratifying about those numbers. Sometimes I’m just working away, and I’ll get a text from someone, or a comment on my Instagram messenger or a Tweet from someone who just needed to tell me that they watched an episode and tried something out that they hadn’t tried before and it helped them solve a problem they’d been having!
And that’s the kind of stuff that just makes all of this worthwhile. We all kind of pulled this series out of nowhere, and to have it come together in a way that was meaningful for our industry is truly rewarding.
To say that we learned a lot in the process of making this series would be a major understatement. It’s funny, because when you’re creating a product like Frame.io, you think you know all the ways that it’s valuable. But until you really put it through the paces of an actual use case, you don’t always know how much you’ve gotten right, or what areas need additional development to make it more powerful.
So, while we started out thinking that we’d share our knowledge with you, the truth is that by making this series and discussing your workflows, we actually know so much more now than we did when we started. We’ve learned far more about remote workflows than I can summarize here.
How our entire community will work as we go forward will significantly change as a result of this experience. So we hope you’ll come back for our next, and final, episode of this series, where we’ll take a look at what we’ve all discovered in our journey together, and use everything we’ve learned to make some predictions and conclusions about what the future of cloud-based workflows will really be like.