Jammy Straub wrote:
I think a lot of people are looking at things from the wrong way when they say things like 'I wish they'd focus on still photography' or 'I don't care about video'.
The still imaging parts of Nikon's cameras are already incredibly refined, the D3s had best in class high ISO quality, the D3x had best in class resolution. Their ability to make improvements at this point insofar as still images is concerned is one of small steps of refinement as they improve their R&D and make hardware refinements.
At this point many upgrades to the video aspect of the cameras are software tweaks to the existing hardware or small changes to how existing things are already built. Things like smooth aperture control, the ability to make 1080p out of various sensor crops, and things that have not yet been introduced like focus peaking are minor compared to improving a sensors dynamic range or high iso quality.
The video improvements we are seeing now are 'low hanging fruit'. Benefits to the sensor tech, ISO quality etc will benefit both still and video capture, they are not mutually exclusive....Show more →
i know they are fine cameras, but can sure be much better if the resourced be focused on stills. progress could been much faster and many product which we want would not be delayed for so many months and updated so slow.
Let's concentrate on making the cameras actually REACH normal HD (1920x1080) before fretting about the 4K format.
The difference between a REAL full-HD signal and the crap that comes out of a DSLR is still quite a lot. You could double the information amount in what is called "full-HD" in cameras like the 5D2, 60D, D7000 and still not fill the HD format with real image information.
Having 8K on the long side means something entirely else: That you can bin 8:4:4 and get a perfectly sampled 1920 resolution. IF that's the solution they've chosen. It would take some serious work to make the binning a production reality though.
BTW - the 500$ J1 sensor/readout combination can already today output 4K/60P.... :-) -It just can't stream the data out fast enough to a recording device to make anything useful out of it. It's limited at a few tenths of a second until the write buffer fills. It would take some quite serious processing horsepower to actually do the video comprimation. But the limitation is on the -out/stream/save side, not on the -camera/in side.
What Nikon and Canon thinks is supposed to be called "1920P", compared to a real 1920P resolution:
Why bother with higher resolution video when the video we have isn't even XGA?
theSuede wrote:
Let's concentrate on making the cameras actually REACH normal HD (1920x1080) before fretting about the 4K format.
The difference between a REAL full-HD signal and the crap that comes out of a DSLR is still quite a lot. You could double the information amount in what is called "full-HD" in cameras like the 5D2, 60D, D7000 and still not fill the HD format with real image information.
Having 8K on the long side means something entirely else: That you can bin 8:4:4 and get a perfectly sampled 1920 resolution. IF that's the solution they've chosen. It would take some serious work to make the binning a production reality though.
BTW - the 500$ J1 sensor/readout combination can already today output 4K/60P.... :-) -It just can't stream the data out fast enough to a recording device to make anything useful out of it. It's limited at a few tenths of a second until the write buffer fills. It would take some quite serious processing horsepower to actually do the video comprimation. But the limitation is on the -out/stream/save side, not on the -camera/in side.
What Nikon and Canon thinks is supposed to be called "1920P", compared to a real 1920P resolution:
Why bother with higher resolution video when the video we have isn't even XGA?...Show more →
Quoted for truth.
The only reason I brought up the 8k spec is in hopes that they could (with such a simple bin) perform that step on chip.
problem is not that much of resolution (comparing 24p video to stills will always lose) but the system as it is.
they are focusing on market which nobody wanted really. sure they get some additional sales but it affects some people. for example i was driven to media format because of slow progress in 135 world. which means i spend a lot less on lenses and accessories when i use the system for only part of my work.
but what do i know, they make most sales on coolpixes anyway, so we just have to carry with what is available. i can at least vote with my and friend's wallets.
To say Video DSLR's are a waste of time or resources, check out the Oscar nominated film, Hell and Back Again on Blu-Ray. Shot on a 5DII and a 24-70 with a homemade steady cam. It is amazing the capabilities of still cameras to be able to shoot video at 1080p (ya I know it isn't real 1080p but its still good) and be so compact. Story is crucial here, he was able to capture great moments and be able to embed with marines in combat. He also had the chance to use some 1.4 primes and shoot in very dark situations. These cameras are a very exciting welcome!
I am VERY excited to work with the D800. I think it will become an all around great camera that would very nicely compliment the D4. Video will be great uncompressed via HDMI, shoot 36 MP for commercial or editorial clients who need it and (I hope) downsize the 36 MP to 18 MP for PJ applications and weddings. It is truly going to revolutionize the D700 and the capabilities of DSLR video. If the rumors are true, it will surely make the 5DII a joke until Canon comes out with their response.
Kittyk wrote:
i know they are fine cameras, but can sure be much better if the resourced be focused on stills. progress could been much faster and many product which we want would not be delayed for so many months and updated so slow.
I believe what Jammy was trying to imply, is that even with the implementation of video... the adjustments towards progression in improving the stills aren't mitigated whatsoever. Lets not try and misconstrue interpretations we "perceive" as correct, because I'm sure the photography aspect of Nikon is much more prominent than some trend video. Also, I'm glad someone caught the false equations and variables earlier.
Chris Langer: it doesnt matter what tools they use, sure some of thousands of films and music videos produced every day few use it when it is available, but that's not what driving force was and is.
innovis: well we see it on D4 how it is not exclusive to each other. I believe that many D3s STILL users really believed they will get more out of so delayed update.
Kittyk wrote:
Chris Langer: it doesnt matter what tools they use, sure some of thousands of films and music videos produced every day few use it when it is available, but that's not what driving force was and is.
innovis: well we see it on D4 how it is not exclusive to each other. I believe that many D3s STILL users really believed they will get more out of so delayed update.
It seems it was delayed because of the earthquake, not for extra dev time. I doubt 5ish months after the product has been finalised isn't enough time to create and test new capability.
Once live view was implemented on DSLRs, video was inevitable (just record the stream). And stuff implemented for video does benefit us - faster frame rates means driving the sensor faster which means contrast detect AF gets faster. Higher res video (and hopefully one day proper downsampled rather than lineskipped video) means data comes off the sensor and into buffer then card faster, which means a faster electronic shutter and better and faster sustained burst rate (for me it's shot to shot time that matters - once cameras get to 50ms, the camera can take some of the strain for capturing critical moments.
Wanting best in class video AF for consumers (combined with the rise of mirrorless cameras) means on sensor phase detect AF will be developed faster than otherwise - that will improve the useful %age of shots from a burst.
As far as improving stills without video goes, until we either get high enough resolution that rgbw is viable/supersampling to the extent that high and low sensitivity pixels can be used, or a completely new photodetection system that doesn't need light absorbing filtration for colour is developed, there isn't that much that can be done imagewise. There's a lot of ergonomics work that can be done, but the value of a pure ergonomics-based iteration would be lost on many.
The other big development will be the the sensor fabs completing the 200mm wafer move and starting the move to 450mm wafers. That will drive the cost of sensors low enough that we could see FF cameras at current APS-C prices, and MF at FF prices. That will be huge for us and the camera companies (them for the lens sales, us for the IQ jump - at this point a sensor size increase seems to be equivalent to about 6 years of rapid technology development).
So the video mode is 4:3? which is a very similar aspect ratio to 1.37 in film terms. wheras 3:2 is closer to the 1.85 crop in 35mm cinema. And then there is 2.35 Cinemascope .. my fav!
lxdesign wrote:
So the video mode is 4:3? which is a very similar aspect ratio to 1.37 in film terms. wheras 3:2 is closer to the 1.85 crop in 35mm cinema. And then there is 2.35 Cinemascope .. my fav!
Then again, all of this is speculation anyhow.
TV was originally in a 4x3 format, as was VGA (640x480) but not many people are still watching 4x3 TVs nowadays.
2.35/2.40 is by far my preference for cinematic films but 16x9 is by far the most prevalent video standard.
Both Canon and Nikon crop their sensors to 16x9 for 720p/1080p because that's the standard.
I think they both crop to 16:10, and then compress to 16:9 (or actually enlarge to fit 16:9). This fits the lineskip/single line "pseudobin" better.
There's no question video hurts still image quality. At least if you want to do video "right", ie minimize Qe to get a base ISO of 400-640, and then hurry the A/D process at twice the intended conversion rate.
The type of video we have in the D7000/60D is as far as you can go at the moment, without a noticeable loss of absolute performance.
