Nikon has just announced the Z9, its flagship camera. The Z9 is Nikon's flagship camera. It is Nikons first mirrorless camera geared specifically towards professional photographers and hybrid shooters. Nikon's $5,500 Z9 camera is the first to be without a traditional mechanical shutter. This allows it to shoot at a new level of speed and focus performance.
Speed is important, especially for sports photographers. It is interesting to consider where this technology might be used in the future to take traditional-style camera. This could be the first step in larger format cameras adopting the computational intelligence that smartphones cameras have embraced for years.
Nikon did not mention computational photography, which allows HDR-style photos and cyclical buffering. Smartphones can simultaneously capture nine to ten frames and then combine them with every press of the shutter. The new 45.7-megapixel, full-frame backside illuminated stacked CMOS sensor doesn't differ from the older phones. At least not in terms of its core design. This type of construction employs a sandwiched architecture that includes a sensor, logic board, dedicated RAM, which allows for extremely fast readout speeds.
The Z9 can use an electronic shutter that is full-time at 1/32,000th of a second, allowing for fast burst shooting and a shutter speed of 1/32,000th of a second. The Z9 can capture 20 frames per seconds in RAW/JPG at full resolution, or up to 120 frames per second at 11 megapixels. There are no audible shutter sounds. Optional fake shutter sounds can also be enabled for an audible cue. Dual CFexpress / XQD slot slots and the new Expeed 7 processor give the Z9 a 1,000 shot buffer in RAW / JPG at full resolution. However, it is the stacked sensor's fast readout speed that could make the Z9 a key to computational photography.
It is the first major camera manufacturer to abandon the mechanical shutter. This puts Nikon ahead its rivals in the race for computational photography. Sony's A1 and A9 line already use stacked sensors to provide fast readout speeds. This makes electronic shutters suitable for full-time duty and Canon's forthcoming R3 will also utilize the same technology. The next evolution in cameras is to use an electronic shutter. However, Nikon will have to demonstrate that its electronic shutter can handle the daily tasks and demands of professional photographers.
Camera manufacturers have so far only attempted to implement computational photography with features such as Olympuss Live ND, Panasonics post-focus and in camera focus stacking. These are useful features, but they don't compare to the potential paradigm shift that fully computational photography could be implemented with each press of the shutter. Olympus' newly rebranded OM System recently stated that it would use computational photography technology in its next camera. However, we will need to wait to see if this is the main focus of the camera or just another feature.
Deep Learning is used in the Z9s object detection system. It has been used previously by Olympus and Panasonic. Although it improves autofocus tracking performance, mirrorless cameras still capture a single image, which is limited by the dynamic range.
Cameras like the Z9, and other mirrorless cameras with stacked sensor may be hindered by the image processing pipeline and data throughput. A 45-megapixel full frame sensor will capture 10 frames simultaneously. The combined files will be exponentially bigger than the same set of images from a smartphone sensor that is only a fraction of its size.
Cyclical buffering is also necessary in order to constantly write and rewrite images to the camera buffer in the background before pressing the shutter. These tasks might be too difficult for the Z9's new processor. While CPUs in smartphones are well-suited to this type of processing, they are sometimes used with dedicated hardware. Cameras, however, are not made the same way. Manufacturers of cameras may need to innovate at the CPU level.
The obvious benefits of computational photography are numerous. A majority of modern smartphones can produce a balanced exposure that includes subjects well lit, shadows with visible detail, and clouds all within the same frame. Google's Night Sight and Night Modes allow you to do things that are difficult with a standard camera. Google also continues to innovate computational tricks to keep subjects sharp in motion. Apple even allows RAW files containing computational data.
However, even with the best mirrorless cameras today, there are still some compromises to be made. For example, you might have to blow out the highlights or crush the shadow details in high contrast daytime scenes. To achieve the same look as smartphones, you will need to do some post-processing and editing. Ideally, this will be done from a RAW file. The JPG or another universal format must then be exported. Although dedicated camera systems may be able to perform computational photography, it could also re-energize camera sales. However, this might mean that camera manufacturers will need to create Wi-Fi apps that aren't too bad.
Professional photographers may find the Z9 camera to be a bridge to this pathway. This might make full-size cameras more interesting, even though it could blur the lines between what constitutes a photograph.
