Panasonic Lumix DC-GH5S Review
The Panasonic GH5S is a video-focused Micro Four Thirds camera built around what the company markets as a 10.2MP sensor. It's best understood as an even more video-centric variant of the GH5: it can shoot either DCI or UHD 4K footage natively (one capture pixel = one output pixel) at up to 60p.
Panasonic wasn't the first company to introduce high quality video to what was otherwise a still camera, but with its GH series it has been constantly expanding the range of professional video features appearing in consumer stills/video cameras. The GH5S takes this logic one step further, by lowering the sensor resolution and omitting image stabilization to make a more single-minded video tool, rather than an hybrid intended to be similarly capable at both disciplines.
The ability to shoot DCI 4K at up to 60p with no crop is the most obvious distinction between this and the standard GH5, but the differences run deeper:
- Oversized 'Multi Aspect' sensor with dual gain design
- 10.2MP maximum usable area from at around 12.5MP total
- DCI or UHD 4K at up to 60p
- 10-bit 4:2:2 internal capture at up to 30p
- 8-bit 4:2:0 internal 60p or 10-bit 4:2:2 output over HDMI
- 1080 footage at up to 240p (with additional crop above 200p)
- Hybrid Log Gamma mode
- ISO 160 - 51,200 (80 - 204,800 extended)
- AF rated down to –5EV (with F2 lens)
- 3.68M-dot (1280 x 960 pixel) OLED viewfinder with 0.76x magnification
- 1.62M-dot (900 x 600 pixel) fully articulated LCD
- 14-bit Raw stills
- 11 fps (7 with AFC) or 1 fps faster in 12-bit mode
- USB 3.1 with Type C connector
As well as the ability to shoot DCI 4K at higher frame rates, Panasonic also claims the GH5S's larger pixels and 'Dual Native ISO' sensor will mean it shoots significantly better footage in low light.
Differences vs GH5
- "10.2" megapixel oversized sensor (vs 20.2MP Four Thirds sized sensor)
- Dual-gain sensor design with two read-out circuits
- Fixed sensor (no internal stabilization) for use with pro stabilization systems
- DCI 4K available in 59.94, 50, 29.97 and 25p (GH5 is 23.98 / 24p only)
- 1080 mode
- AF rated to work in lower light (–5EV vs –4EV)
- 14-bit Raw available in stills shooting
- VLog-L enabled out-of-the-box
- Time code in/out
- 'Like709' and 'V-LogL' color profiles available in stills shooting
- Mic socket offers Phantom Power and Line-level In options
- LUT-corrected display available in playback as well as capture
- 120fps viewfinder mode
Beyond these changes, the GH5S keeps the rest of the GH5's capabilities, with matching codec options and the same support tools, such as vectorscopes, wave forms and preview modes for anamorphic, Log and Hybrid Log Gamma shooting, for instance.
As on the GH5, Panasonic recommends the use of V60 rated cards or faster for shooting 400Mbps video. However, the V60 standard itself seems to be vague enough that even some nominally V60-compliant cards are still not fast enough. The company says to use either its own brand V60 or V90 cards or to stick to well-known manufacturers with a proven history of producing fast cards (and, ideally, to buy from a source with a good return policy).
The GH5S uses a chip that natively shoots DCI or UHD 4K, meaning one pixel on the sensor is used to produce each pixel in the final footage. The sensor, like that on the GH1 and GH2, is oversized. This means it can shoot different aspect ratios using the full extent of the imaging circle projected by the lens, rather than simply cropping down from the 4:3 region.
As well as using the maximum amount of pixels and silicon for each aspect ratio (with consequent image quality benefits), this also means that the diagonal angle of view is preserved, whether you shoot 4:3, 3:2, 16:9 or in DCI 4K's roughly 17:9 aspect ratio.
It also means that the GH5S should offer a fractionally wider angle-of-view than the GH5 when shooting video, especially when capturing DCI footage.
The only downside is that the use of a larger region could limit the use of APS-C and Super35 lenses in conjunction with focal length reducing adaptors, such as SpeedBoosters. A 0.71x reducer needs to capture a roughly 30.5mm image circle to cover the GH5S's larger video region, while a 0.64x reducer needs a 33.8mm image circle, both of which are larger than is guaranteed to be projected by an APS-C lens. You'll almost certainly be OK with the 0.71x adaptor, since that's been shown to work with the majority of APS-C lenses but with the 0.64x versions it's likely you'll have to check on a case-by-case basis.
Panasonic describes the GH5S as having 'Dual Native ISO,' which is standard video terminology for a dual gain sensor design. Such chips have two read-out modes, one that maximises dynamic range at low sensitivity settings and a second designed to minimize noise but at the cost of dynamic range, at higher settings (the second mode changes the 'conversion gain': essentially increasing the pixel's voltage output). It's something we first encountered in Nikon's 1 Series cameras but that's become increasingly common over the past few years, resulting in visible improvements at high ISO settings.
The only difference we can see between the approach taken by Panasonic is that it lets you limit the camera to either one of the sensor's modes, whereas other brands just change mode in the background, without the user ever knowing.
From a stills point of view, the two circuits are used from ISO 160 - 640 and from ISO 800 and upwards, respectively. You'll see talk of the camera having 'Native ISO's of 400 and 2500' but this is perhaps best completely ignored.
One of the only concepts fuzzier than 'ISO' sensitivity itself is the videography term 'Native ISO,' which essentially appears to mean 'setting at which the quality is good but that gives room to move either up or down from.' This should not be confused with the idea of base ISO, which is the setting with the minimal amount of amplification, which usually results in the widest dynamic range.
Lower pixel count
The other thing Panasonic says contributes to giving the GH5S a performance boost in low light is the adoption of fewer and therefore larger pixels.
In general terms, there's no significant advantage to large pixels over small ones: individually they have access to more light (which usually means less noise when viewed 1:1) but once you scale things to a common size, the noise and dynamic range levels tend to be similar. Instead, using more but smaller pixels can have a resolution benefit, even if you then downsize. This is because pixelated systems can only capture a certain percentage of their nominal resolution, but sampling at a higher resolution then downsizing (oversampling) can preserve some of the higher frequency detail it initially captures.
However, readout speed and processing/heat constraints mean very few cameras currently offer oversampled video, instead sub-sampling their sensors to find the ~8.5MP needed to capture 4K footage. This creates a tension between the needs of high-res stills photographer and lower-resolution of video capture. By concentrating on video capture, Panasonic is able to pick sides in this struggle.
The most obvious benefit is that it's quicker to read out fewer pixels. So, while the latest processors are fast enough to generate oversampled footage from high pixel counts, the sensor read-out rate risks creating significant rolling shutter. Having fewer pixels means the GH5S should have less rolling shutter than the GH5.
Having a lower pixel count also means the GH5S is also able to include an anti-aliasing filter that reduces the risk of video moiré, without having to worry about limiting the stills resolution.
Just as we expect to see better pixel-level noise from larger pixels, logic would also lead you to expect greater pixel-level dynamic range (even though again, this advantage tends to disappear when you compare images at the same size). This additional pixel-level dynamic range is the reason the GH5S needs to offer 14-bit Raw files: because you need the extra bit-depth to provide room for that additional dynamic range.
From a photographic perspective it may seem odd to remove image stabilization from the camera but for high-end video shooting, Panasonic says it makes sense. Sensor-shift IS systems operate by 'floating' the sensor using a series of electromagnets. Even when they're 'off' they're not locked in place, they're simply set so that the electromagnets aren't attempting to correct for movement. This has the side-effect that, which mounted on a professional stabilization rig, there's a risk of the sensor being shaken around.
For high-end video work, Panasonic says its users would prefer to use dedicated gimbals and dollies, rather than internal stabilization, and that means physically locking the sensor in place to avoid unwanted interactions between these systems and a floating sensor.
However, regardless of what Panasonic says, there's also the limitation imposed by the oversized sensor: since the camera captures right out to the edge of the image circle there's simply no room to shift the sensor without risking capturing footage of the inside of your lens barrel. This is highlighted in the one situation in which the GH5S does offer digital stabilization: when combined with a lens offering optical stabilization. When engaged, the video has to crop-in slightly to provide room to pan and scan around the sensor.
|Review Publication History|
|January 8||Introduction, video specifications, video features, first impressions|
|January 29||Raw Dynamic Range & Log and DR in video sections added|
|February 5||Image Quality, Video Quality and Conclusion added|
|Llyod's Building Elevators by Jonathan Shapiro|
from Your City - Elevators
|Fire Lake by bbadgett|
|Tail Fins...1961 Cadillac Sedan DeVille by J Warren|
from Car Shows 2018
Canon's new normal is seriously sharp wide open. After shooting with it for a few days, we've prepared a gallery of real-world sample images.