In the early days of the transition from analog film to digital still photography, the common wisdom was that you had to shoot digital images flat to get a proper exposure. This was a widespread misconception that I ended up debunking with many photographers when I showed them that not only could their film look be duplicated, but that it could be done in real time, shooting tethered using custom presets. The only real secret to capturing a great digital file is to properly light and expose the scene (shocking, I know). The rest of the image making chain comes down to workflow choices.
Now we are in the early days of the digital filmmaking transition and everyone looks to film/DI workflows that have been established for multi-million dollar projects using very expensive cameras and involving many steps and a multitude of specialists. This workflow compartmentalizes the jobs of Cinematographer, Camera Operator, Camera Assistant, DIT, Colorist, Editor, Lab etc.. The person exposing the film will not necessarily be involved in every step of this process. The workflow that makes sense in this world is to capture a scene with a wide latitude that has the most overhead to make corrections in post. Final color looks are rarely “baked-in” on the set.
Then comes inexpensive video capable DSLRs and the world of micro production and filmmaking is opened up to a much wider audience. The barrier to entry for creating high quality cinema style motion images is lowered and many photographers find themselves already owning the gear that makes digital filmmaking possible.
When we began the color correction process we quickly realized a new grading process was necessary as the old rules did not apply…We soon understood that with this fragile color space you had to move the knobs very delicately. – Shane Hurlbut
The tricky part has been figuring out the photography/cinematography (PhotoCine) workflow. Photographers search out information on the web in places like DVXUser and other traditional video forums and start piecing together a process that has trickled down from film/TV where the budgets are high and the work load is spread out. Then this information spreads to photographer centric filmmaking forums and there is suddenly a common wisdom that has taken root (remember the common wisdom/misconceptions for digital stills?).
My problem with this workflow from the beginning has been the fact that files coming out of HDSLR cameras are heavily compressed. The advice color experts are giving seems to be counter intuitive. Would you tell a newbie still photographer to shoot their JPGs flat and correct in post? No, you would tell them that since the look is baked in due to lossy compression, they should treat JPGs like they would transparencies – getting as close to a final image as possible. Tom and I have had this discussion many times and end up feeling like outsiders in the HDSLR workflow world. I have spoken with color and post people that I trust who tell me not to worry about the compression and reassure me that shooting flat and correcting in post is the way to go.
I treated the 5D like I was exposing reversal film stock, you had to get it close to what your final product would be. Any extreme manipulation in color was difficult – Shane Hurlbut
I am sure that if you are shooting a feature film using a dozen cameras, shoot-flat workflow could be advantageous, but if you are a still photographer using one or two cameras for a motion ad campaign, you might want to think differently. You would probably want to shoot as close to the final look as possible to minimize the amount of post work and limit problems in the image files by not having to make major color/contrast adjustments later. As a matter of fact, the H.264 format is a finished format designed for uploading to the internet so if you needed to, you could deliver quick dailies straight out of the camera without going through any post process. If you shoot flat and correct later, then the dailies will have to go through transcoding, color correcting, then output encoding. Skipping steps to deliver quick uncorrected clips to the client is a recipe for disaster.
Recently a producer friend was working on a feature for a big studio using a RED One camera for the first time. The show was being shot out of town and the dailies were being sent back to hollywood uncorrected due to time constraints. Keep in mind that the RED files are close to being in a Raw format and the workflow is usually to shoot flat and process the look in post. The studio freaked out and wanted heads to roll because they thought the dailies looked terrible. I explained what I thought was happening and that information was sent up the chain to help ease tensions and start a dialog with the DP and DIT. Of course the RED footage looked amazing once it was color graded and everyone was happy. The funny part was that another of their out of town projects was being shot on Canon 7Ds and the dailies were coming in great because the look was at least partially baked in during production. The point here isn’t that Canon is better than the RED. It’s that you can’t rely on the kindness of strangers to understand workflow shortcomings due to time constraints.
Neil Smith at HdiRAWworks has a saying that deliverables determine the prep. What that means is that once you know your final output (film out, digital projection, Blu-Ray, DVD, web, ipod etc..) you can then determine what camera and workflow best maximizes your needs and budget. I know that those of you reading this are shooting amazing footage that will someday be shown to the Academy on an Imax screen converted to 3D, so money is no object and making every shoot into a big production isn’t a problem. But for those of us doing it ourselves as one-man-bands on small projects tagged onto photo shoots, there might be a better workflow that revolves around making decisions during the shooting process and baking in the look to maximize image quality and minimize time spent in post.
Happy in-camera color correcting – Shane Hurlbut
You can probably tell from the quotes that Shane Hurlbut’s article about color correcting is what inspired me to write this rather long article that goes against the conventional wisdom of the day. Shane is not only a visionary but he is an experienced DP who works closely with HDSLR files in post production and constantly pushes the envelope of digital filmmaking. He has the crew members and the resources at his disposal to “go with the flow” and rely on the traditional shoot-flat post process if it worked for him. I think his approach is a hybrid workflow that treats HDSLR footage more like a JPG or transparency than a color negative film. Read his article to find out the tools and the process he uses to get the most out of a heavily compressed file format before shooting.
Not being a post production expert, I have mostly kept these opinions to myself, but since someone like Shane is thinking along the same lines, the digital image expert in me feels like this conversation needs to be explored more and that conventional wisdom should be challenged every now and then.
Addendum: You can read my response to the discussion this generated here.