Att Slangas: I miss some aspects of current 4K limitations in this article:
1. Limited Framerate: I always shoot 1080@60fps. I have trouble shooting hand held in 30 fps or lower without getting unbearably strobing videos. To get smooth output in low framerate (30fps or lower) you need to pan around very very slowly and keep the exposure time at around1/60s. In 60fps I don't get particularly annoyed even if I shoot at 1/1000s. Which leads to:
2. You need to decide, when shooting video to extract stills, if you want to use the output as stills (then use fast shutter times and get stroby video) or as video (1/60 or 1/50 s)
3. What is the dynamic range of the 4K output? What are the compression artifacts in a 4K video.
Would be interesting to see some side by side comparisons of these aspects.
Agree... 4K/30fps is definitely choppy for fast motion, though improved if you use slower shutter (1/60). Also, the combination of higher def and low frame rate make hand-held shooting difficult for still scenes (you get a jiggley image), even with lens stabilization. A tripod works a lot better.
The studio comparison looks a bit out of focus in the center. The RX10 sample looks better than the RX10m2.
It appears that the A7RII has slightly lower noise than the D810 in RAW at high ISO. OTOH, at low ISO, the D810 looks a bit sharper than the Sony.
As is pointed out in the article, the stage 2 compression (a run length encoding process) is the one to create artifacts near high contrast edges. This is the part the worries me the most - it might well interact with USM and cause issues similar to Jpeg near edges (i.e. "mosquitos").
I think Sony sticks to compression mostly out of stubbornness and the need to save face. There is no excuse NOT to give the user the option of uncompressed RAW on any camera that offers RAW. Other than larger file sizes (and maybe slower writing). Let the user decide what they prefer. Sony could still offer compression as an option, like Nikon.
As to Sony not doing firmware updates, they do sometimes (like the improved CODEC in the RX10m1). Certainly anyone with A7xxx should complain to Sonny support asking for them to do an update.
I believe that DPR, by escalating attention to the issue, has done us a service and perhaps will get Sony to act.
Jabs767: Great shot with a very minor point.
Why is it that a lot of non-aviation specialist photographers shoot prop-driven aircraft at too high shutter speed thereby freezing that beautiful blur of a prop with a contrasting coloured tip that would otherwise describe a beautiful coloured arc?
Having a frozen prop detracts from what otherwise would have been a FANTASTIC shot. Sorry!
And what shutter speed would you suggest to get that, without otherwise getting too much motion blur?
nicoboston: For less than 400 USD, we can get the tiny EOS 100D body, which can be used with dozens of amazing lenses and generates infinitely better images that this pixel jello.Whatever the brand, all these "premium compact cameras" are just a joke IMO. They are not small, not versatile, not cheap, not convenient, very slow, and generates files that are technically put to shame by any entry level DSLR, as well as a growing number of smartphones.Save your money and have fun with a good camera! Take time to study what you can get for... $999.
Show me any smart phone image of equivalent quality.
Claude Ganter: Interesting article. Thanks!
However it would have been worth mentioning that the higher resolution images from New Horizons spacecraft are actually coming from another camera called LORRI, working in black and white.
As for RALPH you get multispectral color information but the resolution is in fact much lower.
LORRI is a 1K X 1K CCD array (13mm wide). RALPH is 5KX32 TDI CCD (65mm wide). The focal lengths are 2630mm and 657mm respectively. It would seem that the spatial resolution is about the same. RALPH operates in a scanning mode and is used during close in flyby. LORRI is a frame capture and used for observations at greater distances.
CameraLabTester: Impressive output for a 9.5 year old imaging device.
If this were launched today, I reckon that the same sort of camera would be used (TDI/CCD) as it is well suited to this application.
olypan: Remember the weeks of hype over the Canon 750/760D. We had pre previews, hands on, preview, review, post review, Canon littering the top of the news page day after day. DPR are dropping the G7 like a bad smell.Oh I'm imagining all this of course?
The RAW still image quality of the canons is distinctly better than the G7 in the DPR studio scene. Even more so at higher ISO.
So it depends on how much you value such characteristics.
This has happened on my LX5, and mt lx7 is showing signs of it.
Serial numbers are bad enough, but the labels on the various buttons on several camera bodies (and lenses) also wear off. I have taken to covering them with clear coat nail polish (very carefully). This is disappointing on otherwise expensive and well made equipment (like a D800 or GH4).
In the not so distant past, labels were engraved and filled with colored paint - even on inexpensive products (look at old Nikon manual lenses). At one time, plastic buttons were made by a (more expensive) two-shoot molding process, where the lettering integral to the part. Early computer keyboards were made this way.
Seems like manufacturers are cutting costs in trivial ways. I am sure that with modern technology, making wear resistant labeling is not impossible.
munro harrap: I am still trying to understand, yet I cannot, as the format itself is so poor. The published samples here are worse than any 16MP APS-C camera, and are certainly not as good IMHO, as I get from an FZ1000: they do not have a photographic quality.
I try to be generous, I try to make allowances, but in a medium in which the lenses now cost as much as full-frame lenses do, why should anybody?
There is no comparison between the results here and those you can get from ANY full-frame SLR, and my ancient 20D and D200 both do better.
The mediums limitations have it stuck at 16MP as well, so it appears that nobody possessing excellent Olympus, Leitz and Panasonic lenses has any possibility of enlarging their work to match other more competent media, and I feel sorry always for people who have bought into the system when I see them.
As noted in the review several pictures show shake they ascribe to shutter shock I dont get ever with 10 year-old equipment. And certainly not with an FZ1000
Agree with you. I guess I just don't like 4/3 either. The inherent aspect ratio is lousy for landscapes; you can crop, but you loose pixels in the vertical direction, and the 16MP leaves less to work with. The kit lenses start at 28 mm (equiv), and are really closer to 30mm in the horizontal field of view. 24 is better, but you really need to invest in an (expensive) wide angel zoom. Less margin to bring up shadows due to less dynanic range and higher noise levels.
The shutter shock issue should never have arisen - Panasonic should know how to make shutters by now (and we are not talking about 36-50 MP camera here).
Other issues I have with this camera is the poor design of the 4-way controller, which is difficult (for me) to operate by touch, and the flush video record button.
I am inclined to go for the new Sony RX10m2 for a "light" weight general purpose camera.
Jim Hully: Once again Panasonic USA forces you to buy this camera with the kit lens., no body-only option. DPReview, why isn't that a "con" in your conclusions?
I don't want the hear the usual witless suggestions about just selling it. Tell you what, next time you have your car serviced, let the dealership sell you a new wheel. Not something you wanted but hey you could always sell it...
I agree, a con. Do not like to have to use 14mm as my "wide" end. Much prefer to have 12mm, which means buying a 12-XX zoom. The 14-42 kit lens appears to sell for very little on the second hand market.
breivogel: Very poor ergonomics on the G7 in two critical areas. This was so bad that I returned the camera, which I otherwise liked.
The Video button is totally flush with the top surface - it is impossible to tell by feel where it is. They should have at least have put a raised ring around it or made it stick out a bit. The 4-way controller is also very hard to use, as the buttons don't stick out enough either and are hard to actuate. Maybe these work OK with small Japanese hands, but not mine, They would be totally useless if wearing gloves (as in wintertime).
I tried and Olympus OM-D E-M5 II - the difference in the same controls was night and day. The M5 4-way controller much better, and you could locate the video button.
I don't think you should have to memorize where the button is on a video-centric camera like this. Should not have to add a stick-on either. An obvious design flaw not present on other cameras.
Hard to believe that someone with big hands could use that 4-way controller. I tried at night and it drove me mad - always hitting the wrong button. And my fingers are not not SFFS - work just fine on the LX7 and RX10. Of course, this is a personal perference, YMMV.
Very poor ergonomics on the G7 in two critical areas. This was so bad that I returned the camera, which I otherwise liked.
mgatov: Thank you for posting this. It led to me purchasing the Panasonic LX100... as I was really on the fence between these two cameras.
I agree that the LX100 has some issues compared with the RX100.I had two samples of the LX100, and both had very poor focus under normal daylight conditions for certain distant subjects (trees!) The EVF on the RX is much nicer - no rainbows, good black level, good optical quality, although a bit of bother to have to pop it up. Sony menu layout is also better.
QuarryCat: I have done tests with EOS 5DSR - Sony has a lot to do, to deliver such a good low noise performance and such an incredible sharpness... we will see what reality brings. Canon's ISO 6400 is realistic and even Alpha 7s is at 6400 ISO just 1 1/2 Steps finer at 12 MP compared to 51 MP.You can use 51 MP at 1600 ISO with best sharpness on the market and very low noise. Even 3200 is realy good and surpasses EOS 5D3 or Nikon D810.But for now, the EOS 5DSR is the king of the crowd - by far!
I would say that the noise performance of the 5DSR appears, at best, similar at 3200 to the D810 - not better. Dynamic range is another matter. If you are going to lift shadows a lot, the Canon leaves a a lot to be desired. The Sony sensor design is superior.
At higher ISO (12,800) the 5DSR lags other cameras.
It appears as the low end compact camera is taken over by smart phones, that the manufacturers see premium compact (RX100, LX100, G7X) and bridge cameras (RX10, Lx1000, G3X) as the place to make money - fewer units but a higher ASP. I wonder what the profit margin is on such cameras?
mosc: I really don't get this type of product at all. It's nothing against Think Tank Photo's particular offerings here and I'm sure they make other products that I would prefer but I really just don't get this offering.
These are not sturdy enough to protect a 70-200 in the overhead bin of an airplane let alone checked baggage. These are not sturdy enough to put in the luggage area of a greyhound bus (with somebody famous's name on the side). These are bizzare to think about as camera-only backpacks on a hiking trip compared to a more general purpose bag. They certainly don't offer any type of holster-like ease of access middle ground either.
Anybody want to explain why I'm crazy?
Looks like a reasonable backpack (the 10) for short hikes where 3 lenses would be enough. Provides better organization and padding than a normal backpack used to carry camera gear. I don't know if well padded enough to be tossed in checked luggage, but should be OK in an overhead bin. Like that includes a rain cover. Seems like it should have a water bottle pocket holder though, and I am not sure that the side zippers are necessary.
It appears that the active area, number, and pitch of the pixels remains the same as before. The peripheral circuitry that previously sat around the pixel array is moved underneath it on a separate layer. The big advantage lies in the fact that the fabrication processes for the photo sensor and the logic layer can be individually optimized. This probably results in the ability to run the logic at much faster speeds (hence the ability to do much higher frame rates as well as 4K). I might have a concern about possible thermal issues, as the underlying circuitry might heat the sensor array unevenly. OTOH, charge injection from the logic could not influence the photosites.
It is really quite an amazing technology, actually bonding two chips on top one another and interconnecting them - all with acceptable yield for a consumer device and the potential noise sensitivity for photosensors.
I love your prose (you must be from the UK). Yanks seem to have lost the knack.