 | ||
| ||
| ||
Media Management Nightmares? Game On!
by Dustin Lau| Creative COW Magazine : Games Issue | ||
Our weekly video game magazine program is imaginatively titled "Game." Broadcast on Asia's top sports network, ESPN STAR Sports, Game showcases gaming news, the latest trends, and previews of upcoming titles. A big draw for viewers is our speed: we review the latest video games within a week of their official release in South East Asia and the Pacific Rim. Our challenge is to turn the fourteen hours of game footage that we capture into a 30-minute episode week after week. We have a very short time frame, and a very small team, comprising two full-time gamers/reviewers, a director and myself, all of whom deal with the media - often upwards of 400GB captured to DVCPRO 50-PAL. You may not produce weekly programs for a major network, or have a team of people working with you, or have such a huge amount of footage for every project. But what follows is an extreme and unique example of how taking some time to plan your workflow can make all the difference between hand-to-mouth, skin of your teeth survival, and doing work that you have time to enjoy. Otherwise, it's all too easy to slip into just doing instead of creating. CHALLENGES WORKING WITH GAME FOOTAGE Why do we start with fourteen hours of footage per episode? Because I can't make my "DPs" (the gamers) shoot less. We review the latest game releases, so we have no concrete idea what events will occur as we play. A standard shotlist is simply impossible. This makes for a great number of video files, some more than 20 minutes long. Of course, the first challenge is capturing the footage. Gaming machines like the Nintendo Wii and Gameboy, Microsoft Xbox 360, the Sony PlayStations (PS3, PS2 and PSP) and others were never meant to be ingest sources. Using them as your primary source of footage makes for anything but the typical post-production experience. None of these fine machines generate timecode or offer machine control for recording. They output in a plethora of resolutions, formats and scanning frequencies. The first obstacle we faced is that an overwhelming majority of games are NTSC, whereas we broadcast in PAL. Fortunately, Final Cut Pro now supports mixed resolutions and formats in the same timeline, so this is not quite as big a problem as it would have been in earlier versions.  Input management with the Blackmagic Decklink Extreme The other stumbling block we encountered was the PSP. It outputs an HD resolution that is recognized by all the capture solutions we tried as an SD resolution, but with progressive frames - a particular challenge to match with interlaced HD footage we can capture from other sources. Additionally, some games simply don't offer much control over output resolution at all. We have found ways to deal with this madness through excruciating trial and error, and a combination of hardware and software too complex to explain clearly in an article twice this long. Suffice it to say that conquering challenges like these is not for the faint of heart. Media in hand, it's now time to start managing it. 
KNOW THE NAME, FIND THE FILE FCP has earned a somewhat bad reputation for media management. You can help by taking a few extra steps for every project you undertake. An easy trap to fall into is limiting yourself to folders as a way to organize media in context. But when you take an individual file out of a directory, you may be lost. Can you tell from a filename like "JI.psd" where this file belongs, or what its contents are? This is why I enforce a very strict file naming system for all of our media files - it is the very first step toward making the media manageable. The system I use is simple, but descriptive: Game-S[season number]E[episode number]-[Game name]-[Type of footage]-[Shot Number]-[description] For example: • Game-S02E09-MGS4-Interview-08-Hideo on engines • Game-S02E14-Halo4-B-roll-13-Master Chief mascot This system results in a file with a human-readable format, that leaves very little doubt as to what the contents of file are. Using the examples above, the first is an interview with Hideo from episode 9 of Game's second season. The second is a clip of b-roll, which focuses on the Master Chief mascot, for the section of episode 14 devoted to Halo 4. I always use leading zeros (09 instead of 9) so that FCP sorts multiple shots properly when they are sorted by filename. The Shot Number (08 in the first example above, 13 in the second) is necessary for the video game footage, as the sources do not generate timecode. I would have no other way to know the order of shots within a game session. Such consistent file naming removes all of the guesswork for me, but also for Final Cut Pro. I can now make very specific searches. 
If I need all the voiceovers recorded for Episode 45 in Season 2, instead of searching for "VO," I can look for "Game-S02E45-VO," and not have to scroll through every other VO for every episode I've ever ingested. And if I want to find all B-roll in Episode 15 except for those featuring the game GTA4 (that is, Grand Theft Auto 4), I could search for it like this: You can use similar approaches for any kind of footage, in any kind of workflow - perhaps multiple commercials for the same client. Each clip name might contain the client name, spot name, source, type of footage, and a very short description. 
You obviously needn't follow this exact recipe. The point is that, no matter what size your operation is, don't waste time wrestling with media management when you can easily prevent these kinds of problems. Consistent names take moments to apply, and can save hours. PROCESS MANAGEMENT Everything we set up is to buy us time. One of the best ways to save time in FCP is working with footage in real time. However, we deal with a large number of files that, depending on the properties of a specific project, cannot be used in real time without being converted first. These include: •CD music sampled at 44.1kHZ being used in a 48kHZ project • AVI files from our gaming PC • H.264, MPEG and WMV video game trailers downloaded from the internet The only way to get the savings of real-time editing for these and many of the other formats we work with is to convert them all to the proper formats before the edit begins. I create a few folders for every project, with a tilde at the front, so that they will always be sorted to the top of the project folder. (See image at bottom left.) Any still images, movies and audio clips ready for real-time playback are placed in the folder named ~TO BE IMPORTED. Other elements are placed in the ~TO BE CONVERTED folder, until they have been processed and moved to the ~CONVERTED folder. 
The end result is that I always know, instantly, exactly where I am toward my goal of doing as much as possible in real time in FCP. SAVING TIME ON THE TIMELINE As each episode progresses, some creative decisions may be taken that will be maintained throughout the remainder of the series. It is best to have these sequences saved, so that they can serve as a template. Normally, it would be a pain to make changes to the current episode while also updating a template project for future episodes. However, if the files are named with the system I've described, I can use the previous week's episode as the template project. In the example illustrated below, I use Game- S02E06 to make a template. I just make a duplicate and open the sequence. Then a few quick steps: Hit the Home key, then Command-F in the FCP timeline. Type in "Game-S02E06", and click "Find All." The media that remains the same from week to week remains unselected. But the clips specific to Game-S02E06 will be highlighted. (See image below.) I then execute "Make Offline" (shift-D) for these clips and have a new template to start the next episode. The reason I make the clips offline instead of deleting them is so that they retain motion, distort, crop, filters, etc. I can now simply "Paste Content" to insert new footage, and have it automatically take on the properties I've already defined. This is how proper naming makes a template dramatically more useful. All pertinent clips may be treated in one pass, without having to replace clips one at a time, rebuilding the attributes of each one by hand. TRACK MANAGEMENT Consistent track organization also speeds things up immensely for me. My full frame graphic transitions are always on the topmost track (typically V8 or V9 in my projects), unless they are meant to be within graphic donuts, in which case they are on V5. My main footage/B-roll is always on V1 unless it is cropped/keyed and has a background, in which case it will be on V4. 
This track consistency makes it easier to do tasks throughout a sequence, like adding the Flicker filter to graphic files with fine lines, adding Broadcast Safe only to shot footage which has not been legalized, etc. In general, seeing at a glance where things belong makes them easier to manage. RINSE AND REPEAT No show has ever been lauded for having one great episode and being mediocre the rest of the time. The crux of episodic TV is this: "Great! Now do it again, but better and faster." There are no shortcuts to making a show great. But these are some ways to move through the grunt work more quickly, so you have more time to tweak, and bring it to the next level. I typically work 18-23 hour days in the first few weeks of a new series just to get the templates right. It's a lot of extra effort but it pays off by giving us back the time to make that baby shine! All of these processes and procedures are technical. They are not easy to implement, and the people who need to carry them out won't always like them or follow them. I am fortunate enough to be working with a team that generally conforms to my file naming system, as annoying as it must be for them to do so while also killing zombies, racing cars and generally shooting things up. Taking time to plan, consider and settle your workflow before you begin working increases your ability to function over the long term. You and the rest of your creative team will have more time to think about how to make the show look good. You can make creative choices instead of fighting workflow and media management fires.
Find more great Creative COW Magazine articles by signing up for the complementary Creative COW Magazine. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Dustin Lau
| Games | Add Comment | ||
Comments
Hi Cody, glad to hear you liked the article.
I understand it seems like working long hours before the start of a new series makes it seem like the workflow doesn't work but the truth is turnaround times in Singapore can be very tight. Just to give you an idea, for the first season of the show, we had 3 weeks to build the edit and capture suites from scratch, troubleshoot the conversions, prepare graphics, music beds and sound effects and complete the first episode. Other than the graphics, which were done by the graphics department, the rest of the hardware wrangling and setup were done by myself and one of the gamer/reviewers. Our typical episode for the show involves about 3 days of capturing the gaming footage for our main reviews which can amount to 14 hours of rushes. For 2 of those 3 days, I cut the rest of the entire show, convert the main review footage on the night of the last day and cut the 2 main reviews in the next 2 days. On the other remaining day, I cut another half hour show also related to video games. There are also certain massive video games titles that launch on the 'wrong' day of the week for us. We go to air on Thursday. One week, there was a huge game that launched on the Monday, and this wasn't one of those that could be completed at one sitting. It was an open world game with 50+ hours of gameplay and we had no way of getting an advance copy before Monday. So, one of the gamers played the game overnight on Monday, while the other had to host segments on location at the developer's launch, before returning to take shifts on the game. As he was chatting with video game magazine reviewers at the launch, he asked them when their reviews were due for print. They replied that they would have 3-4 weeks to write their articles. In turn, they inquired when they would see our review on air. They were flabbergasted when he replied, "Thursday." For every big game release of the past 2 years, we have gone to air with a relatively detailed review of the game in the same week of the game's release. On occasion, we have done so with the help of advance copies kindly provided to us by game companies, but for the most part, the proof that the workflow works for me is that we've made it to air with a good show for 40 weeks with a 4 man team and 4 days of the week to do so. | |||
Great article. But it killed it for me when he mentions working 18-23 hour days for a few weeks of each new series. Kind of seems like this entire workflow isn't all that wonderful in the end...?
| |||
