This week I’m back to work on “Godkiller,” Halo-8’s first Illustrated Film. An Illustrated Film is a highly stylized animated movie that mixes original graphic novel illustrations with motion graphics and dramatic voice performances to create an edgy new style of story telling. Its like Liquid Television meets Ralph Bakshi, allowing us to tackle stories that we wouldn’t be able to do with video.
Halo-8 doesn’t have experience producing animated titles, so I didn’t have a post production work flow in place. I sought advice from a couple of people who work with traditional animation, then tried to adapt that work flow to Illustrated Films. The VO recording was handled by The Engine Room in Hollywood. We recorded the voice actors in separate sessions as “wild lines,” meaning there was no cut picture for them to reference. After the session, I would get the audio files organized by script page, which left me with the chore of compiling all the takes together. Basically, I gather each take of every line of dialogue and stack them on top of each other in the timeline. When all of the takes are compiled, I have what I call the “Take Matrix.” There can be anywhere from 2 to 12 takes of every line. By putting each take on a subsequent audio track, I can solo individual tracks and quickly hear every take of a line. After the Take Matrix is assembled, I go through every take and copy the select take to track one. When I’m done with that process, track one represents a completely compiled performance for that actor using the best read of each line. If I need to make tweaks, I can easily go back into the Take Matrix and audition other reads.
This was especially useful in working with Lance Henriksen’s performance. Lance likes to do long takes and fall into a groove. He’ll do a bunch of these long takes, kind of riffing on the dialogue and making subtle variations along the way. His compiled dialogue performance ended up with a lot of edits because I went deep into mixing and matching his variations, pulling individual phrases of a monologue from several different takes. It makes for a really dynamic voice performance.
"Take Matrix" for Lance Henriksen's performance in Godkiller.
Once you’ve got a Take Matrix for each actor, its very easy to combine all of the actors selected takes into the final dialog track for the film. You just go to track one of the Take Matrix for each actor, copy from the compiled performance on track one, and paste the dialog into the master edit. Its easy to make tweaks as necessary by going back to the Take Matrix and auditioning new lines.
The saying goes, “Good, fast, cheap. Pick two.” Its a clever anachronism, and broadly applicable to a wide range of endeavors. I doubt that it was coined to describe working in TV and movie production, but I can’t tell you how many times I’ve heard it used over the years. Trouble is, it makes less and less sense every day. In fact, In the revised landscape of media 2.0, its straight up nonsense.
Let’s break this saying down into its parts:
“Good” - The definition of “good” is entirely dependent on the expectations of the audience. In the fractured, niche driven mediascape of cable TV and web video, where most video professionals do their work, “good” content is that with an authentic voice and a unique point of view. If the production values are positioned to support authenticity, then the content will have value to the viewer. So “good” really means “good for what it is.” Or more accurately, “good for the market in which the content competes.”
“Fast” - Once you’ve figured out how to satisfy your market place, you need a system for the consistent release of timely content. The zeitgeist moves too quickly for us to spend two years producing a movie, TV series, or webisode, and releasing quality content at regular intervals is the only way to stay visible in the vast ether of niche media. If you miss a couple of release opportunities, the fan base generated by your debut will quickly forget that you exist.
“Cheap” - The price of HD prosumer equipment has fallen sharply over the years, so there’s less excuse for video production to be costly. Decide what kind of production value your marketplace demands and produce accordingly. If you drastically over produce, then you end up with a system that can’t sustain the consistent release of material, and your beautiful production values will disappear into the ether. You’re better off using modest resources to build a following, then expanding the production values as applicable.
BC One is a world wide, annual break dance competition organized and produced by Red Bull. They have a massive investment in grassroots culture and sporting events, and BC One is probably the most entertaining event in the roster. Red Bull has been trying to sell this show in the U.S. for years, and with the recent success of other dance competition shows, MTV finally took the bait. The clip below is the intro for the MTV show.
This text will be replaced
A couple different editors worked on this intro, specifically the “I am a b-boy” section, and the Rokka on-camera intro. I cut the rest. I was really pleased with the graphic open where the dancer is in front of the stenciled walls. The walls were actually shot clean, and the stencils were added in post using After Effects. Our designer, Steven Jones, found those classic French images and treated them by hand to look like painted on stencils. The footage was shot on the Phantom high speed camera. The Phantom is an amazing toy, but it can create some political baggage if its not used properly. The Phantom cost far more to rent than a standard HD camera. Its about three times as much, depending on the configuration. When you are trying to justify a rental that is three times the standard rate, you might aid your cause by showing amazing demo clips like the infamous shark catching the seal video and you say “We’re getting the camera that shot this.” Everyone is blown away by that shot, but it sets a ridiculous expectation for delivery. In our case, the phantom footage was beautiful, and the slow motion helped the aesthetic, but the effects were very subtle. Inevitably, the Executive Producer asked why we needed to triple the camera rental for such a subtle effect. In other words, “Where’s our shark catching a seal?” Its not that our footage wasn’t beautiful, but we had an editorial and graphic concept in place that dictated the structure of the piece and it didn’t allow time for long, dramatic slow motion effects. If you promise the shark catching the seal, you better deliver the proper equivalent:
Editors have been complaining for years about Apple’s ambiguous support for professional products. I feel their pain, because its seems crazy when Apple calls something a “pro” model but removes commonly used access ports, or fails to innovate over several successive software releases.
I was an early adopter of Final Cut Pro. Back in 2000 I used version 2.0 and a Digital Voodoo card to edit a promo for Fox Sports. That was a horrible experience, but I finished the spot and stuck with the software because it was affordable and flexible, two traits not shared by Avid back then. A few years later, Apple released the first Ipod, then the Itunes store. I realized then that Apple didn’t care about my needs as a Pro Apps user, and our relationship was not made to last.
Apple is a hardware company. They only produce software to generate demand for hardware. They have no interest in creating innovative post production solutions. The Wikipedia site for Final Cut Pro states very plainly that Apple bought Final Cut Pro from Macromedia as a “defensive move.” Macromedia’s beta version of FCP was built for Mac and Windows, and Avid and Adobe were moving toward the Window’s market, purportedly because of Apple’s notorious policies of secrecy and exclusion. Apple thought it was worth buying FCP to keep desktop video users on Apple hardware. It was an effective strategy for maintaining a market foothold, but its not a plan for innovation in video production.
My prediction: Final Cut Pro has three years to live. Maybe four.
Apple’s domination of the gadget market takes the focus off of desktop computer users, so they no longer need video production professionals to bolster the Mac market share. I’ve been squawking for weeks about the imminent rise of the “Consumer-Producer” and the dissemination of professional video production into common knowledge. Apple will use this growing trend to combine Final Cut Pro with IMovie and bundle it with all new Mac computers. The Apple “reality distortion field” will frame it as a democratization, movie making set free, or whatever. This consumerization will allow Apple to cease development on professional features, and the Pro Apps department will die quietly. Software companies with a real interest in post-production innovation will fill the professional void and we’ll be better off.
I’ve been making an argument that the coming demand for video content created by the increasing viability of web delivered video will not be filled by television and film professionals, but by novice consumers with access to affordable HD cameras and post-production freeware. Today I heard some news that adds a new chapter to the story. Wikimedia is adding user generated video to wikipedia.org. In the same way that the pages of the site are user edited by the public at large, wikipedia will facilitate the addition of user generated “documentaries” to each entry that can be challenged and edited collectively. The video will be managed by Kaltura, a company offering content management solutions and a remarkable web based editing tool. You can take it for a test drive here. Its rudimentary at this point, but this is the future of post production.
The first generation of You Tube videos were horrible because it was like America’s Funniest Home Videos multiplied exponentially. Consumers didn’t know how to produce videos for You Tube publishing, so they uploaded what they had. But it didn’t take long for people to start producing for You Tube. They figured out what tools they needed, they learned how to use them, and now thousands of amateur producers create video content for self publishing on You Tube and other user generated video services. The technology available to them is improving, their skills are rising, and the demand for Internet delivered content is booming. With some coordination, and a little effective direction, these advanced Consumer-Producers will be ready to create TV shows, documentaries, news packages, EPKs, etc., for a fraction of what that work costs now.
As I stated previously, the business of producing television will be drastically undercut by the proliferation of free editing technology, and the dissemination of video production techniques into the realm of common knowledge. Much like what has happened to the business of photography, the bottom will drop out of video production budgets, and the middle class of working producers and editors will be marginalized. This will create a two tiered production system:
1) The elite Hollywood studio system that produces ultra expensive, high profile movies and TV.
2) The aggregate underclass of freelancers.
Ironically, the lowest paid members of the underclass, amateur video enthusiasts, are already hard at work for the elite studio system, and they’re doing it for free. Judging by the amount of “fan trailers” on the Internet, I don’t know why any film studio needs to pay an editor to cut trailers for highly anticipated movies. Imagine if Warner Brothers uploaded all of the footage for the new Harry Potter trailer and allowed consumer-producers to create their own trailer edits. Within days they would have hundreds of trailers to choose from, and probably some worth cleaning up and using for theatrical. The studio could eliminate the exorbitant expense of producing trailers internally while exciting a core fan base of movie geeks and generating terrific buzz.
A pre-YouTube music related example of high quality user generated content was the Squarepusher music video contest that Warp Records held years ago. They posted dozens of entries on their website, and they were amazing. Sadly, Warp didn’t give those videos a permanent home, or migrate them to a YouTube channel. The surviving videos are scattered around the internet in a way that doesn’t help Squarepusher or Warp Records.
Another example of user generated editing comes from Nine Inch Nails. Trent allowed a small crew to shoot three concerts with HDV cameras, then the raw footage was seeded as a torrent. There are now hundreds of these edits floating around the web, and the best have been collected on a NIN fan site dedicated to producing a free DVD of the best cuts. Trent gives up the revenue of another concert film in exchange for a deeper relationship with his most devoted fans.
For strong brands that have a good faith relationship with enthusiastic fans, user generated marketing is a smart way to stay ahead of the overwhelming demand for web content.
I found an article about the remote color grading process used by Company 3 for Transformers 2. They made the very ambitious leap of sending a 2k stream from Santa Monica to Japan where the director could watch the high res output and give real time direction. Sadly for us, this article doesn’t say anything about the technology used to transfer the data, or the process involved. What information they did give makes it sound like an overcomplicated wank for frame fucker clients like Mr. Bay. Don’t get me wrong, I like to frame fuck too, but this seems like a lot of leg work for a grading session. Did it have to be a realtime session? He couldn’t screen the files after the session and make notes? They could have compressed the graded shots into Apple Pro Res HQ and delivered the files over FTP. With a laptop computer and an HD monitor/projector he could have screened the files in his hotel room. That seems easier too me, and probably about 1/10th the cost, but I guess that’s why I don’t make $200 million movies. God bless the Studios.
For the rest of us, the dream of untethered post production is slowly coming true thanks to affordable editing gear and accessible internet technology. Setting up a post-production office for a TV show or movie creates a huge amount of overhead, driving up the cost of production. There are three reasons for establishing that kind of overhead, and they’re all becoming less relevant every day.
1) Access to technology
The technology problem is basically solved. Most editors have a functioning edit rig at home built out of the same off the shelf gear that you would find in a post production department. Ultimately, I think all of the current editing solutions will be replaced by web based video editing apps that will function without specific hardware. Imagine if Final Cut Pro or Media Composer ran in a browser from a centralized server, and you could access your projects from any computer with an internet connection; Cloud Editing.
2) Access to media data
Shared media storage is a necessity for most workflows, so that makes it tough for editors to be off site. But the ever expanding internet bandwidth will eventually meet up with the ever shrinking compression of HD video, and bingo: Cloud video storage. Imagine Google docs for HD files.
3) Supervision
Supervision of the editors is a psychological hurdle, but one that is easily overcome. The emerging tools of social media could be easily modified for closed network professional use in production and post. Imagine a Facebook type web app for the production staff, but integrating video conferencing, realtime feeds from remote edit bays, media management, approval and delivery systems, and other custom tools to accommodate the production. The staff does need to be supervised, but they don’t need to be down the hall from you. They can be across the country as long as there is a workflow that provides communication and accountability.
I found this technically thorough explanation of why high end HD cameras shoot a superior image even though prosumer HD cameras often look quite good. If you ever find yourself pushing for a better camera rig for an upcoming production, this guy has the math to back you up.
Now let me tell you why that kind of math doesn’t matter for 90% percent of producers:
Off the shelf, prosumer HD cameras are suitable for at least 90% of today’s content production demand. The bulk of TV and web video consists of reality shows, how-to, documentary, news, magazine shows, EPKs, etc. With the shrinking budgets of cable TV production, and the hyper syndication and accelerated life cycle of web video, real value is found not in glossy production, but rather in the consistent output of serialized content that meets the demands of a niche market. In other words, content that is good for what it is. That doesn’t mean producing for the lowest common denominator. It means producing authentic content that will resonate with the target audience. The aesthetic production values should support the authentic voice of the content, and nothing more.
If you produce TV content, don’t automatically go with the $2000 per day HD camera package out of habit. Analyze the market where you compete, understand the niche you are trying to own, and produce accordingly. You might find that you can use more affordable, more accessible equipment and still satisfy the market demands. Use the money saved to produce more added value content for your show, or a pilot presentation for another show, or take a trip to Barbados.
If you’re an independent producer of low budget movies or web content, never wait for technology and never let it bog you down. Use the tools that you have to dominate your niche with the consistent release of authentic stories told from an original point of view.
Its good to know the math of why high end HD cameras are better than consumer cameras. There’s a time and place for pristine images and high end production value. But when technology stifles authenticity and consistency, then the production becomes a wank that won’t serve the needs of the market. No matter how much it costs, content produced as a wank is quickly buried in the landfill of anonymous and ineffective media.
I do quite a bit of my own sound work, partly because I enjoy it, but mostly because many productions don’t have the time or money for a mixer. On projects where I do work with a professional sound mixer, I spend a lot of time in the audio bay asking questions and picking up techniques. Below are four tips for making a decent sound mix without a formal mixing session. I used these techniques to mix the trailer for Godkiller, Halo-8’s new Illustrated Film. I used Final Cut Pro and Soundtrack Pro, but these techniques can be applied to any software package.
1) Reference Professional Mixes. Find movies and TV shows that have a similar sound aesthetic to to your project and use those mixes as a reference point. Listening to reference mixes on the monitors you plan to mix with is invaluable, but there’s more to see and hear if you’re willing to dig into those reference files. Its surprisingly easy to tear apart the sound from a movie and analyze the basic elements. In my case, I was mixing a trailer, so I went to the Apple trailer site where they offer high quality downloads of hundreds of trailers. If you drag those quicktimes into the Soundtrack Pro timeline, you can see the six individual surround channels as individual waveforms, and see each channels level on the output meter. I downloaded the latest Harry Potter trailer and found that they set the average dialogue levels between -12 and -6db. I used that as a reference for the dialogue in my Godkiller trailer.
Its nearly as easy to make a detailed analysis of any feature film. With a couple of Mac OS freeware apps, you can rip the audio off of a DVD and individually monitor each channel of the surround mix. I used Mac The Ripper to rip the AC3 surround file off of Kill Bill. Then a program called mAC3dec split the encoded AC3 into six mono AIFF files that can be played in Final Cut or Soundtrack Pro. That’s a fascinating, detailed way to see how the elements of a surround mix are spread out, and how the average mix levels are set for dialogue. Another useful tool is a program called Audioleak which analyzes audio files and measures average and peak audio levels for the entire program. You can run the same test on your mix to see how your average and peak levels compare to a studio movie.
2) Stay Organized.
Keeping track of individual audio elements is only possible if your tracks are methodically organized. Many editors still try to cram everything into 8 audio tracks. This is a bad habit left over from the days when the Avid could only monitor 8 channels of audio. Most Avid’s can monitor 16 tracks now, and Final Cut Pro has no such limitations. It took me a while to integrate so many audio tracks into my workflow, but I now work with 16-30 tracks of audio in Final Cut Pro, depending on the project. This allows very quick access to individual elements, and easy soloing of certain groups. If you use OMF or AAF to transfer your sound into an audio mixing application, working with a more detailed level of organization will save you the hassle of having to re sort every audio clip by hand in the audio app.
3) Start Mixing Early.
I find that the final mix is much easier if I had been paying attention to the audio levels as I was editing the dialogue and picture. Once you’ve referenced some other movies or TV shows, choose an appropriate average level for the dialogue and try to maintain that level as you edit. Sometimes I get lazy about this, and I always regret it later.
4) Mix Around The Dialogue.
When you’re ready for a detailed mix pass, mute everything except the dialogue tracks. Go through the entire show and adjust the dialogue levels so they are even throughout the timeline. How easy this is depends on how well your dialogue was recorded. Use your ears and the audio meters to make sure that everything stays within the average level range you have established. You can get pretty micro about this process, including using the keyframes and mix automation to adjust individual words, boost syllables that might trail off, and kill short peaks. Once the dialogue is set at an appropriate level, add the sound effects channels to the mix and go through the timeline again, mixing sound effects in underneath the dialogue. Do that process again for the music tracks, keeping an eye on the peak audio level, and you’ll have a solid mix right out of the edit. From there, you can send the sequence to Soundtrack Pro or any other sound app using OMF export, and all of your mix work will migrate. If your tracks are well organized, it won’t take long to add some plug ins and dial in a tight mix.
Early this this year, Halo-8 made an ambitious leap into the new medium of Illustrated Films. The first production is “Godkiller: Walk Among Us”, featuring the voices of Lance Henriksen, Daniel Harris, Bill Moseley, Nicki Clyne, Davey Havok, Justin Pierre, and more. I’ve been working on the Godkiller post-production on and off since February, and the first release is slated for September. You can read more about our release plans at Bloody Disgusting, and more about the movie at Godkiller.tv.
I want to say thanks to the people who helped bring GK to life. Its a small, talented team of enthusiastic professionals, and we couldn’t have done this without them. I’ll be writing more about the production process in later posts. If you’re curious, stay tuned.
