9.25.2017

An evening at the theatre, watching the tech rehearsal for "Singing in the Rain" and learning just what the GH5 is capable of...


Last night I had the opportunity to document the technical rehearsal for "Singing in the Rain" at Zach Theatre. I really like using new cameras in the theater environment as I bring them into the fold in order to understand how they work in what I think is a challenging situation. The lighting on stage sometimes changes minute by minute, and those changes include shifts in color as well as intensity. In addition to lighting changes the actors are constantly moving and I have no control or ability to really anticipate changing expressions and gestures. It's a situation wherein you have to capture stuff you like, in the moment, and then edit toward the best stuff in post. 

In order to make it even more rigorous a test I sometimes deny myself the "crutch" of using the raw file format so I can get a sense of how well the cameras make Jpeg files. It's a bit idiosyncratic but then I never held myself out as a paragon of logic or consistency....

Our rehearsal started at 7pm. I got to the theater a half hour earlier so I could say "hello" to the many people on the technical staff that I've worked with for years. I also wanted to sit quietly and set up my two cameras at, potentially, the optimum settings.

I used two Panasonic GH5 camera bodies and two Olympus lenses. It seems counterintuitive but I think the two Olympus Pro lenses that I ended up buying are a perfect match for the GH5s. When I owned the GH4 cameras I chose (among other ancillary lenses) the two f2.8 zooms from that system; the 12-35mm and the 35-100mm. They were both good as well as smaller and lighter than my current  choices. I think the Olympus Pro 12-100mm is a better choice for all around photography than the 12-35 by dint of the much greater coverage on the long end. I also think it is a sharper and contrastier lens system than either of the Panasonic lenses I owned. Being able to cover most of the focal lengths I use in day to day practice is a time saver and means that most of the time I am wearing only one camera over my shoulder rather than two.

I bought the Olympus 40-150mm f2.8 Pro mostly to use as an adjunct to the 12-100mm in theater and corporate event photography where I might need both the one stop of extra speed and the extra 50mm of reach. While I found myself using the 12-100mm a lot last night I am happiest with the images I shot with the longer zoom. It's not that the files are "better" or sharper it's that they seem a touch cleaner and more detailed. I'd be perfectly happy with the shorter lens but there were a few shots, taken at the longer focal lengths, that just made me smile.

So, that was the gear inventory. The "kit." Two bodies and two lenses. Both set up nearly identically. 
I selected the finest Jpeg setting at the largest file size of 20 megapixels. I chose the standard profile for each camera and left all the parameter settings at their presets. zero, zero, zero. 

I set the camera with the 40-150mm on it to ISO 800 and the camera with the 12-100mm on it to ISO 1600 to compensate for the one stop slower maximum aperture. After consulting with the lighting designer I settled on 4200K as the color setting for the general illumination. The follow spot is cooler and some of the side spots are warmer. There's nothing much you can do about a wash of light that's homogeneously red, blue or magenta. With Herculean efforts you might be able to render a neutral color but it would be a fool's errand to try. After all, they are called accent lights for a reason. 

I tried not to use the shutter under 1/125th or above a 1/500th. If I had enough light to reach for 1/500th the logical thing to do would be to turn down the ISO. 

I shot both of the lenses wide open for the entire evening. On a wide stage shot I was hanging out near the 12-20mm range and figured that depth of field would cover any small disparity between camera and subject distance. At the longer focal lengths I was trying to grab tighter, one person shots and would depend on focusing accuracy of the system for best results. 

In earlier tests I found that the screens were brighter than the calibration on my studio monitor or my Atomos Ninja Flame monitor and knew that setting the brightness one or two notches below zero would help me push up the exposures in a good way. I also enable the histogram and put it in the bottom right hand corner of the screen, consulting it often.

While I had trepidation that the GH5 would stumble in lower light situations I found that anything I shot at ISO 800 or less was as perfect as anything I would expect to see out of any previous camera I had used to shoot this kind of work. At ISO 1600 I started to see (when viewing at 100%) the wavering hand of noise reduction impinge on overall image quality. It manifested itself in overly plastic skin tones and some harsher sharpening of bigger detail. After seeing this I tested setting the camera profile differently. I decreased sharpness by one notch under the default and brought down the noise reduction as well. With noise reduction I experimented with one and two notches under the default and I found that one notch under was enough to make the files perfectly acceptable to me while not requiring me to do anything in post production while a two notch correction required a bit of noise reduction fine tuning in post. 

I tried several different focusing options with the two cameras. I started with my usual: A custom configured AF pattern consisting of four smaller squares in the center of the frame, driven by S-AF. This method nailed every single shot I pointed the camera and lens at, and without any hesitation. It was much quicker and more certain than my Sony A7Rii or A7ii. 

Emboldened I thought I'd try out the methodology that seems to be the preference of many other photographers: To choose all the focusing squares across the frame (wide) and to set the focus to C-AF. Amazingly the camera locked on to the closest thing in the frame and worked predictably. I didn't notice any time lag at all and anytime the camera made an (infrequent) mistake it was because I started out pointing it at the wrong object. Following one actor as they walked across the stage was remarkably consistent and well locked in. 

Since these cameras are both highly competent video cameras I strayed from still imaging from time to time to shoot video snippets that I'll use in the video I am editing of conversations with the director of this play and the choreographer. I thought it would be really nice to have good b-roll of the lead actor actually tap dancing in the rain. When the stage hands turned on the rain device and we got a downpour on stage I switched to 4K video at 30 fps and followed the action. Again, the camera and the 40-150mm were able to follow the dancer (or his feet ) as he tapped his way across the stage under a convincing shower of rain. It's really beautiful footage and you'll get to see it sometime soon. 

We photographers are often given to hyperbole. Smaller format shooters would have us believe that their particular choice of camera is so special that it can transform the laws of physics and deliver performance equal to full frame (35mm) sensors across all performance parameters. Conversely full frame shooters would have you believe that anything smaller than their (35) frame size destines a user to end up with files that are nothing but mush and crap. The truth lies somewhere else. The scale isn't a linear one with full frame to 100%, APS-C at 75% and micro four thirds at 50%. The band of technical behavior and results is much tighter than that. All the cameras are in the 90%s. And there are always many compromises in each direction. As science and industry get better at solving imaging problems the ability to apply computational processing is aggressively flattening the field and reducing the differences we used to see between formats. 

I'm sure I will not be able to blow up these images to sizes as large as I would if using full res files from an A7rii but I'm equally sure that I can get close; and the reality is that either camera will make prints that are large enough so that the optimum viewing distance veils any substantive differences in overall quality. Fast lenses on small formats can equal the look of slower lenses on bigger formats. If f1.4 equals f2.8 (m4:3 versus FF) and we have f1.4 available then we gain two stops of shutter speed or two steps of ISO. Stuff works out. 

After carefully examining about 1200 photographs and six minutes of 4K video I'm happy to say that I feel comfortable using the new cameras in just about any situation. The secret is to understand how they can be set to ensure optimum operation. Used well just about any modern camera can excel. But few can match the GH5 for it's versatility (vis-a-vis the combination of video and stills).