ND64

hate to repeat myself but its about letting your loyal customers think other brands are handling the video better, even if its just a perception made by very local internet commenters, not that D5500 or D750 output is totally useless. People are comparing and vote with their money.

They know market is going to be saturated very soon (anyone who needs a video camera, has already one, and dedicated video cameras are not like smartphones that you upgrade every year. Its hard to convince professionals to ditch their two years old gear, you know), so they want to be the one who saturated it. Its way better than just watching :-)

Canon Inc said on Tuesday its third quarter operating profit fell 48 percent, hurt by a strengthened yen following Britain's vote to leave the European Union. and its market leader with 45% market share. imagine Nikon situation.

Yea, D5 is the only flagship DSLR with 4k 4:2:2 HDMI output at the moment, but its not enough in 2016. Moreover, its not like the expected features need rocket science. Its already done by others. I'm pretty sure Nikon engineers are capable of delivering the best 10bit log profiled image with most refined color science, efficiently encoded. The problem is their boss doesn't ask them to do it.

A dedicated video camera is out of question for Nikon. They don't make sensors. Sony has its own fab(s), Canon hss its own, Panasonic has its own (plus Towerjazz), so Nikon has to ask someone to make a video optimized sensor for them, and because its very low volume market, that will be hugely expensive. Df is niche, but its sensor is not. Besides, even if they find a reasonably priced sensor, the mount problem remains. Most serious video shooters are in PL camp, or using Canon lenses, or already adapting their old lenses to Sony's open mount. These people have no interest for F mount. No, it clearly makes no sense, businesswise. The thing is they are letting their customers to experience and finally embrace other options. When a DSLR shooter realizes he needs a secondary camera for video jobs, which is probably a mirrorless, he slowly discovers that secondary camera can also take pretty high quality stills in situations he wouldn't expect it to perform well, so asks himself Why I should keep using my bulky DSLR in every shoot? Its not even dangerous for DSLR business, its disaster.

"There are two types of high end users [using video]’ explains Kikaota: ‘those who started [their careers] shooting still pictures, but also there are a lot of customers who started with video. In our company we have two types of customer. It’s difficult to make one solution to fit for both of them.’ ..Shooting video with current cameras can often be quite complicated, we suggested. ‘[This] complicated operation is fit for the customer they started shooting from video. On the other hand there are professional [stills] photographers: they are not yet used to using video, their demand is easy operation. So there are two types types of people. It’s a big problem." ------- We've always been wondering why Nikon, that has no Cinema line to protect, literally resists to offer pro video features. and now, in this DPR interview, I can see why. They fear the pro video features confuse their conservative still shooter user base! Really unbelievable! Yea, maybe they're right, many of their customers have no clue what video is, let alone stuff like log profile, focus peaking, color sampling, codecs, LUTs, grading,... and even if they need to record a short clip they want it to be just Start/Stop operation. But how on earth this can be justifiable excuse to not delivering features serious video shooters are asking for years? Just add a dedicated menu, or charge for Premium Firmware if you wish, whatever. Why it should be a "big problem" for a company that brought the video to the world of DSLR?

The main problem with smartphones is their lenses. While Pixel lens is way better than other flagships but its still far behind even a P&S camera lens. and they can't do much to improve it though, I mean its ridiculously tiny plastic!

Even a DSLR body like D810 can't handle it alone sometimes. Definitely a lot of stress for A7 kind of body. But if you can keep your left hand under the lens all the time, and I mean literally all the time, then its ok.

Yes, and thats why it needs different algorithms. in your sample, binned version isnt soft. There is really 2k worth of details in the data, its just imprecisely rendered. but Sony is not much interested in spending R&D money on this software puzzle while they are capable of delivering ever faster hardware, though with heat issues. Here is the thing: 30-40mp still is the lowest resolution people are going to demand, and they want 60fps 4k in the same device, which should remain light and compact. With your preferred full-sensor-readout-and-downsamplig method, a lot a lot of data should travel through system and be processed that will generate extra heat, consume extra power, and affect overall reliability.

in a video camera, yes, but we are talking about small hybrid cameras that are supposed to take high resolution still images. in that case best solution should be the best balanced compromise. (as a side note, 2x2 binning need different demosaicing algorithm than conventional bayer, and Sony is not good at making those algorithms)

Nobody said its not possible to downsample 4800 pixels to 3840, but that would be a huge CPU task for very small unnoticeable benefit. Sony chose similar approach in A7r2 (though that's 5168 pixels) but its not an efficient way to do 4k. The best solution could be 40mp sensor to do 2x2 binning with full field of view of FF, and because sensor outputs only 8mp frames, adding 60fps capability would be much easier.

15.4 is like 4800x3200, so gonna crop (1.25x) to get UHD?

Shooting film is stupid right now.The less a medium is used, the more is the cost of keeping it alive and usable. Kodak would stop the production few years ago if it wasn't the cash support from cinephile snobs like Tarantino. Shoot 4k, with a HLG-like profile, ProRes 10bit 4:2:2, and store it in enterprise HDDs.

a year old sensor with even 12mp more pixels with even higher speed read out, aka A7R2, has better performance than this, but Ok, its fine. 13 stop of usable DR is enough for %90 of situations.

His problem is he couldn't even say the chip is not fast/efficient enough to support log processing, because it is. It can process 60fps 4k plus full time AF in 1DX2 case.

Having a virtual 6 stop ND without paying a dime and with absolutely no colorshift is so a big benefit shouldn't be ignored, by anyone. Just because it was near impossible in 1920 to make a rotary shutter fast enough to achieve 1/4000 exposure doesn't mean we shouldn't use it today. But, with such high shutter speed you lose lots of event-data that could be used by your brian to understand the movements in the image. At 1/4000 and 24fps, you capture only %0.06 of every second, while at 1/48 you get half of every second. This is huge difference. Of course that lost %49.94 may only contains blurry pixels of displaced objects, but those pixels help our brain to think "this thing is moving". If you give it a flashing sequence of appearance and disappearance of objects, that will be harder processing task to render it as normal action. I get headache after watching a video with bad motion cadence for longer than five minutes, especially when I'm close to screen. Maybe the combo of myopia and astigmatism I have exacerbate the problem though.

I dont know where this "nice image" came from, I dont consider Canon colors "nice". Its not natural at all, and contain lots of hue shift. it was just better than what other "photo" camera makers were delivering. in raw cinema game, it means nothing.

Nobody push 6ev, but its what log profiles do, somehow. You expose to the highlight, camera lift the shadow, and put the whole thing in 8bit range. This test shows that D750/D810/K1 with log could be very capable 2k video cameras, because you could have near 14 stops of usable DR at clean ISO 6400. AND in 5Dmark4, if it had log, you shouldn't go beyond ISO 3200.

Not gonna happen. Nikon's main problem is they don't have their own fab. Having Sony as your sole supplier makes you encounter with some consequences. I think they will deliver a 1.56x crop (6000/3840). Do you really think Sony would give them a 6k/30fps sensor, while cheap enough to put in a $2k body? I just fear they don't realize how important is D750 line for Nikon shooters who need a hybrid system.

Actually 2x crop factor of upcoming GH5 will have higher quality than almost 2x crop of a FF sensor. Because in FF, its 1:1 readout, but in GH5 it will be downsampled. These days we dont have much problem with read out noise, we are now limited to shot noise (randomness of light), and that kind noise will be lower in downsampled data. I mean its not just about obsolescence of DSLRs, its about format advantage.

Apparently every pixel has two subpixel which makes it a potentially 60mp camera. so it has 6500x2 pixels in horizontal line. With skipping every 3 subpixels, you get 4.5k resolution, but probably with more noise.

Nikon used a thinner CFA in D5 for achieving higher S/N, yet I haven't seen users noticing its color be inferior to D4 or D3. if there was meaningful difference between A7RII and IQ180, assuming both use same optics (same formula, same coating,..), I'm sure it would be presentable. its not aesthetic, its about numbers.

not-good-as-ccd-color mindset is made because usually something is wrong with raw processing. here is CCD IQ180 vs. your A7RII tonal range according to DXO

seems you think with changing optical formula you can change gemoetry laws :-) Never mind

Nope. some optical designs, correct the perspective or any distortion better than others. in case of your mentioned lens, distortion at close distances is apparently better corrected than longer distances (maybe because it was used for 1:1 reproduction). You can make a lens with similar design for any format, even m4/3.