The main news with the Lumix GH4 is of course the video features. However, the GH4 also adds a lot of features that make it a viable upgrade for those who are mostly into photography, e.g., better flash sync speed, faster burst rate, also while doing continuous autofocus.
Over the recent years, we have seen significant improvements to the image quality coming from digital cameras. We get a better dynamic range, and less noise, especially at higher ISO. I expect this development to continue, of course.
At this time, it looks like the Lumix GH4 is roughly equivalent to the GH3 in terms of image quality. Early comparisons show that the GH4 produces sharper images at high ISO, however, it appears to be mostly due to more aggressive sharpening. This is also the case with the GM1, which appears to apply more sharpening to the out of camera JPEG images.
While the current image quality performance is very good, future generations will still improve this.
Dynamic range in Video
When improving the video output from previous generation M4/3 cameras, there are essentially two approaches: Increase the resolution, as the GH4 does by adding 4K video, or increase the dynamic range of the 1080p video stream.
Increasing the dynamic range requires using higher bitrate video and other codecs that support more bits per pixel. An example video camera that does just this is the Blackmagic Pocket Cinema Camera. It accepts Micro Four Thirds lenses, however adding an additional crop factor. With the crop factor of 2.88, it makes a 14mm lens into a 40mm equivalent short normal lens.
The camera supports ProRes 422 high dynamic video output, and also lossless CinemaDNG coded. While this sounds all well, the downside is that using all this dynamic range requires professional processing of the video output, called "grading". Without competent grading of the video stream, the video is just going to look grey and boring.
So it is probably not surprising that Panasonic chose to increase the resolution over increasing the dynamic range when creating the Lumix GH4. While getting the real world benefits of a higher dynamic range requires a lot of competence, everybody understands that higher resolution is better. Selling 4K is easier.
Still, future Micro Four Thirds cameras may very well go down the high dynamic range path as well, adding ProRes 422 or CinemaDNG as optional output file formats.
4K video without a crop factor
A downside of the Lumix GH4 4K implementation is that not the whole sensor area is used:
So in 4K mode, a 14mm lens behaves like an 18mm lens (36mm equivalent) in terms of field of view. For people using the GH4 4K video feature, a wide zoom lens therefore becomes valuable, e.g., the brilliant Lumix X 12-35mm f/2.8.
The reason for this implementation is simple: Using the whole sensor for 4K video would require downscaling from 4608x2592 (16:9 sensor area) to 3840x2160 (4K) for each frame. High quality scaling of images is very computer intensive, and cannot be done with current technology. Here, Panasonic have an improvement potential.
As a side note, Sony is releasing a new version of their Sony A7 capable of 4K video. Somewhat strangely, this Sony A7S has a sensor with only 12MP. However, the reason for this is very easy to understand in the light of this article: It is to be able to produce 4K video without cropping, and without rescaling the sensor output. They could have sampled the 4K video stream from a larger sensor too, but that would typically lead to line skipping artefacts.
Higher video framerate
So far, no Micro Four Thirds cameras support higher video framerates than 60FPS.
Many competitors do better. E.g., Nikon 1 cameras support video framerates up to 1200, albeit at a reduced resolution.
Perhaps some will say that this is largely a gimmick, but I think it would have been a fun gimmick to have.
Traditionally, cameras have required a mechanical shutter. This curtain shutter starts and stops the exposure. However, a mechanical curtain shutter is infeasible for video use. Hence, the exposure is started and stopped electronically.
This will be the future of cameras: An electronic shutter which is silent and doesn't wear out. So far, some Panasonic Micro Four Thirds cameras include an electronic shutter, however, it is not very useful yet. The GH3 electronic shutter has a 1/10s readout speed, which is very slow and has a lot of side effects. The GM1, on the other hand, has a somewhat faster electronic shutter at 1/25s, which is better but still not perfect.
Other manufacturers have come further in this respect. For example, the Nikon 1 cameras have an electronic shutter readout of 1/80s, and some models have gotten rid of the mechanical shutter completely. Using this electronic shutter feature, the cameras, can take 60 full resolution images per second, which is very useful for sports applications.
An electronic shutter capable of instant readout is called a "global shutter". This is the aim for the development, but we will probably not get there during the near future.
Current Micro Four Thirds cameras have a very fast autofocus for static still image use. This hardly needs any improvement.
On the other hand, continuous autofocus of moving subjects is still not very impressive. The same goes for autofocus during video recording.
Other manufacturers try to address this by employing on-sensor PDAF detectors, e.g., Nikon 1 and Canon 70D.
Panasonic have said in interviews that they don't believe in on-sensor PDAF. They rather want to improve the autofocus by improving the image processing technology. This requires improvements in both the processing algorithms, and in the processing speed.
While the Lumix GH4 is certainly an impressive camera, there are many ways in which the cameras can continue to improve. Can you think of other improvement areas?