Deconstructing large projects in order to learn how to structure them better next time. Tune out now if you don't want to hear about video projects....

I've been working on a project since the beginning of May. Not everyday but the timeline of this video project wound itself through every week and sometimes we were shooting for days at a time. Some of the time spent was involved in chasing weather; which is almost as unpredictable as photographing toddlers with pets... Some of the time was spent scheduling interviews and more time was spent actively re-imagining the project as we worked through it. Here are things that I learned:

First off, I spent way too much time worrying about camera selection and technical stuff. When you shoot a corporate video there is always the need to use existing, archival footage and there is generally the need to grab stock photography of things that can't be conveniently shot in the moment. You won't be shooting an ice storm in Texas in early June.... This basically means that you are cutting together all kinds of visual content and some of it won't be state of the art. At first I felt compelled to either rent a high end, dedicated video camera or to use the Sony A7Rii because it's 4K video (when shot in the "super 35" mode) is really wonderful. Transcoding the 4K video into a 1080p timeline gets you close to what you might get shooting with a 1080p camera that does 10 bit 4:2:2.

I quickly abandoned those options because a lot of the anticipated production was outdoors, in the rain and in heavy duty weather. I knew I wanted a camera with a fixed lens and at least the promise of weather sealing. If I was going to lose a camera to water damage I sure didn't want it to be a pricey rental or my main photography camera and, shooting in the driving rain hardly shows off either the dynamic range of a sensor or the overall detail and resolution of the system.

I ended up using the two Sony RX10 cameras I already owned; the RX10ii and the RX10iii, with the "3" being my "go to" camera. I love the heft and feel of that body. One thing I discovered as I got more and more experienced and comfortable shooting with the "3" was that the 4K footage emerging from the camera was even better than I expected it to be. It is so detailed that I would think twice before using it in a studio situation while interviewing people who expect to be visually flattered. I turned the sharpening down in the profiles I used to create the 4K content.

Speaking of rain, I used some inexpensive Goja rain covers for the cameras and the covers got a real workout. In several instances we were able to shoot in rain coming down at a rate of several inches per hour. With clear filters sealed onto the front ring of the lens with electrical tape, and everything covered, the cameras emerged dry after hours of use. I carried microfiber cleaning cloth under my own poncho to dry the front surfaces of the filters. It all worked well.

The next thing I learned is that the noise from "small" sensors that people seem to worry about is largely non-existent at normal ISOs and, even with higher ISOs, is largely mitigated when you are reducing all of the information from a 4k file into a smaller, 1080p file. It's a four times reduction and the noise, as you can imagine, gets reduced in kind. In fact, we wound up using the RX10iii for almost all of the primary "footage."

When we sat down and compared our content with samples from the Sony A7R2, the FS-5 and the RX10's we found the work done in daylight, and in well lit interview situations, to be so close as to be, in some cases, indistinguishable. Since almost all of our video work was done in lit situations, or in gloomy and inclement weather, the additional dynamic range that may have been a feature of the most expensive and complex cameras was not required. That meant we spent less time trying to become  experts in the ways of S-Log and more time was spent actually working on the project.

Then I re-learned that the director of a project, which contains a number of interviews, must craft questions for people who are not used to being interviewed; questions that lead them efficiently into getting what is needed for the program. We weren't doing any muckraking, investigative journalism, and we were really trying to get the authentic story for our project, but many times questions delivered from the client were far too wide ranging and open. There is the fantasy amongst the general public that film makers will set up the cameras and microphones, get everyone settled and then just let things roll. Many people love the attention of being interviewed and would talk for hours, if we let them. It's important to step in, redirect, and let the "talent" know exactly what you are looking for. They might say something great on their own but chances are it will be buried under a lot of other words, and most of the words will arrive without the pauses that would allow editing cuts. If I had my way every single person in front of the camera would be scripted. I know, I know; that's not the way it works.... but being able to craft the messages, and the time spent delivering them, is crucial to the pace and length of a video.

My two moments of either hubris, or misplaced confidence came at the same location. I was shooting interviews in warehouse with so-so lighting. I brought my own light, via a couple of LED fixtures, but I still found myself working at ISO 800. No, the problem was not electronic noise, it was an issue with focus. I had the camera on a stable tripod, I composed the scene carefully, I lit the faces to match the ambient light but with about 3/4 of a stop more pop. I set the camera to use the face detection AF and watched it engage; the box turned green. Then I stepped to the side of the camera to conduct the interview. Big mistake. The camera was interested in the subject for about a minute and then started to lose interest and focus on stuff in the background. By the time I noticed there was an issue we were almost done with the interview. We got sharp "footage" at the front and the back of the interview but the stuff in between was dicey. The camera requires a fairly high light level to work its face detection magic with any reliability.

The only thing that saved the footage was the content from the RX10ii that Ben was using as a "B" camera, from a different angle. We were able to use the front end and back end footage from the "A" camera, and the audio from the "A" camera, but depended on Ben's footage for everything in the middle. That, and lots of cutaways to still shots that reinforced that particular interview. For the rest of the project I switched to manual focusing. But if you really need to see exact, fine focus with the RX10iii you'll need to exit the video mode and magnify in the still mode. The magnification in the video mode is only to 5.8X while the still mode allows more than double that. Once you get the lens in focus it's brilliant. Getting there can be another story...

The second bit of bad luck was nothing but hubris on my part. We were trying to do our interview with one person during a driving rain storm. Real pounding, torrential rain, and it was banging into the  uninsulated metal roof like a full percussion section of an orchestra that only does punk. I pinned a lavaliere microphone to the subject and decided to go for it anyway. After hours and hours and hours in all sorts of audio restoration programs I finally had to throw in the towel. You can't separate bad, loud noise from a voice once they are all mixed together into an audio stew. I should have moved the whole production to a different location; even though time and logistics were not on our side.

We talked in an earlier post about making a packing list so you don't pull a dumb stunt like I did. Came out during another rain storm expecting dismal lighting and as soon as I got on location the sun peaked out from behind a cloud just to remind me that I'd left the variable neutral density filter at home.

So, lots of travel miles were logged. We saw a lot of rain. We shot a lot of video. What did we do right?

We worked to a strong initial script. There was some back and forth to refine it but a good script provides a backbone for a project. You might not follow everything in your script verbatim; plans change when you see the location (and the talent), but it gets you thinking in terms of the overall message, and of the continuity required to pull it off.

When you get to a location to shoot video don't let your "video blinders" allow you to ignore the still photography potential of a place. If you have a subject being interviewed on video you might find that you can shoot a ton of different angles as photographs, and you might have more control over the image than you would in video. We inserted a couple of images that were taken with a 14mm lens on full frame and they work very, very well. Having lots of stills means you have a library of possibilities, in addition to the primary footage, when you head indoors to edit. They can add a lot more spice to your project.

Work hard to select your music bed before you start editing. A good music track will inform the timing and rhythm of your edit and make it all work more fluidly. In my recent experience, if you've taken care to consider your (approval) audience and have worked hard to find music that works you might never even hear from the committee about musical choices. There is an ocean of stock music out in the web. Make sure you buy it and use it legally.

Keep track of hours and expenses as you go along. We tend to bill by the project and almost never by the line item, but we keep track of details for those rare moments when we have to justify something. Good accounting also informs you about ways to do projects more profitably in the future. If you grossly under estimated your mileage on the current job (for example) knowing where you went wrong means you probably won't make the same mistake when you estimate the next one...

Speaking of billing. Be sure to be clear about what is included in each step. Adding days requires additional fees. Adding new sections adds fees. After the approval of the rough cut changes of images, music and content requires additional fees. Any major edits after the approval of the "final" review edit require additional editing fees and administration fees. These are routine in order to compensate us for our time. The agreement on additional fees also focuses clients on getting to internal consensus in the first place, rather than on the other side of the editing process.

Back to what we did right... We stayed flexible for the client's benefit. If we needed a good shot of hydro-electric power generation and we couldn't find the right one in a stock house we drove out and found just the right one and grabbed good content with our cameras. If we missed a great shot because of weather we sucked it up and re-scheduled for the next opportunity. Larger production houses, with staffers and freelancers (large crews in general), don't have the luxury of going into overtime without killing the budget. We'll do it if we can get a better shot and a better outcome.

One place where we ended up practicing overkill was in our audio (after the disaster with the rain noise). While I felt that the cameras could be counted on to provide clean audio straight in I had also read lots and lots of articles on the web leading any reasonable reader to believe that the audio pre-amplifiers in any but the most expensive production cameras ($20,00+) are absolute crap.

With that in mind we ran everything into a digital audio recorder with very clean pre-amplifiers, XLR connectors, phantom power and all the right tools. But, on the last day of shooting, I decided to take the signal from my Sennheiser wireless lav mics into a little, passive Beachtek D2A (to make sure the balanced and unbalanced connections were compatible) and directly into the camera. Surprise, there was no readily discernible difference in sound quality between that method and the more complex, dual sound set up we'd been using. Seems that Sony is actually pretty good at making audio circuits in most of their cameras. Who could have known?

Here is some random stuff that I jotted down in my notebook during our shoots:

The sandwiches at the Redbud Cafe in Blanco, Texas are very good. So is the poblano potato soup.

When shooting in storms add to your kit --- extra pants, an extra pair of socks, waterproof shoes, a second set of shoes to drive home in if you are too tired to scrap mud off the first shoes. Bring a couple of towels. You'll want to dry off once you get into your car. Bring a plastic seat cover, your leather seats will adore you.

Always bring a good flashlight. You will drop something vital in the dark.

Video always looks sharpest when the color balance is right on the money. Make a custom white balance whenever you can.

My cheap, Ikan shoulder rig worked very, very well in conjunction with the camera system's image stabilization. Using a shoulder rig can serve as a tripod substitute (for a short period of time).

Log your footage at the end of every shooting day. Don't make it harder to find or take the risk of overwriting a memory card.

Always bring two tripods. You never know when you'll want to shoot simultaneous "B" roll on a day with no scheduled assistant.

If you shoot with Sony cameras you can never have too many batteries in your case.

If you shoot with multiple Sony cameras be thankful that all the ones you probably want to shoot with share the same batteries.

Bring water with you.

Fill up the tank whenever you hit the halfway point on the gas gauge. When things happen quickly one hundred miles away you may not have time to stop and fill up en route. When the power is out the pumps don't work.

The Sugar Shack Bakery in Wimberly, Texas makes great pastries.

Thyme and Dough, a cafe and bakery in Dripping Springs, Texas, has great coffee. Really great coffee.

Ask interviewees to take a little pause after each few sentences. You'll need somewhere you can cut.

Always put your stuff back in the same place. Batteries always go in one pouch. That pouch always goes in the right side pocket of the camera bag, etc. It means you can find stuff quickly, even in the dark.

If you are recording audio (dialog, interviews, narration) always wear a good pair of "over the ear" headphones. Buds won't hack it. Better than nothing but never as good as the big, goofy ones.

Build your edits carefully because about half the time you'll get a change request that blows them right apart. Make sure changing an edit doesn't take out most of your timeline.

Be sure to work with an editor who is smarter and better than you. (Not hard in my case...).

Figure out how to make all of this fun.

Realize that video is much harder and more time consuming than photography. Rush back to embrace your happy career as a photographer!!!!

Added note (June 14th): Final cut now officially approved at all levels, including, but not limited to, the CEO. Celebration at the VSL headquarters begins immediately....


George said...

Fascinating - I'll probably never shoot video, but the process, and the repeated reminders to pay attention to every dang thing and to bring redundant, never-use-it gear are priceless. Thanks!

Michael Matthews said...

Seems to me that Sony has been innovating superior means of portable audio recording for decades. If what they put into today's consumer video cameras is anything like the quality of PCM audio (the linear pulse code modulation track recorded to moving tape in earlier times) the limiting factor is not going to be the recording circuitry. It's what you're able to bring to it in the way of a clear, clean signal properly matched in level, impedance and other characteristics. The attention you've paid to that via the Beachtek matchbox and the more complex double-system approach of feeding through a standalone recorder has paid off.

Kepano said...

Nailed it on the issue of interview questions. That alone can make or break a corporate video. Life is so much easier in post when you have good sound bytes and lots of b-roll.

Stephen Greszczyszyn said...

Thanks for this, I really enjoy these behind the scenes workflow posts, videos, etc. Best way to learn outside of going out there and doing it yourself.