When Will Nikon Counter Canon's 5DsR?

Discussion in 'Nikon' started by 25asa, Jun 20, 2015.

  1. I heard rumours Nikon was working on a 54 megapixel SLR using a Sony sensor. Yet Sony just announced a 42 MP camera in the A7RII. So I'm wondering if a 54 megapixel sensor is actually in the works? That said when do you think Nikon will counter Canon's 5DsR camera to reclaim its claim to highest megapixel DSLR?
     
  2. When? Soon. And that raises the question when Canon will beat Sony/Nikon's still to be released sensor? Answer: not
    too soon.
     
  3. The lenses we have don't fullyr resolve 36 megapixels so honestly what is the target market for even higher megapixel cameras? It seems the age old mantra of lenses first has been forgotten.
     
  4. ShunCheung

    ShunCheung Administrator

    In the history of this forum, nobody who askes about any rumor has ever gotten any concrete answer. You are better off
    asking your source of such rumors. Otherwise, I would like to think the mega pixel race has ended. Back in 2012, I bought
    a 36MP D800E, and I notice that diffraction begins to set in at f8 and starts to degrade the images seriously at f11. With
    even denser pixels, I am not sure there is much to gain other than some huge image files, and of course bragging rights.

    Last year I added a 24MP D750 for a bit faster frame rate, better high ISO results and smaller image files. I haven't switched my image printer on for a year. Even 24MP seems too many nowadays.

    I would rather see how Nikon improves its AF system, which is an area Nikon has fallen behind for a while.
     
  5. I'd say : Ask Nikon..
    I also know the answer from Nikon already :
    "It is our policy to never disclose information about future products or future developments " or some thing like this .. :)
    Sigma does a better job in marketing, Sigma anounces half a year beforehand that they will be announcing a new product or development in the near future, so that all interested people get a wake up call, and start anticipating on and discussing about the future product so that a lot of people really start waiting until it becomes available. this way Sigma creates a market for their products before they are even starting production ...
     
  6. When.... Rather ask yourself: do they really need to? Instead of looking at the total megapixel count, see how many real world pixels you actually really gain on the long end and short end, and how that translates into real-world additional printing size. It's neglible. Even disregarding the (correct and valuable!) point on whether lenses actually can keep up, wonder whether ever-increasing resolutions do bring real world advantages. And it seems we're bottoming out - no news, but larger sensors are the only serious answer if you really need to scale up.
    Will Nikon respond? Probably yes, as the marketing department has its needs to; but whether all this excess is really bringing something useful to photographers, I have severe doubts. I get excellent 13"x 18" prints from a D700 still; the pixel count really isn't the issue.
     
  7. Lenses can indeed keep up. That's not something to worry about.<br>Perhaps not many of the more recent crop of lenses, that seem to stress the assumed importance of a wide zoom range in a fast lens more than image quality. And which are made knowing that many sensors are too 'dumb' to begin with, and are hidden behind a Bayer pattern and a soft focus filter to boot. So why bother with making a lens good? But lenses used to be good enough, and those oldies that do rather well are still around.<br><br>The worrying bit is that 24x36 mm is too small for that many pixels. Larger sensors (larger area) are what is needed. Get a PhaseOne back and put it on a MF camera if you want or need (and yes, there can be a need for many pixels. I also still scan film and the 120+ MP i get from a 6x9 frame sometimes still are not quite enough) a high pixel count.
     
  8. Who cares? 20, 30, 50 mp it is still 24x36.
     
  9. I don't want or need 50 mp. I would like a higher dynamic range and a true 16 bit. Give me better pixels not more.
     
  10. I was given the d810 for my birthday last year as a gift from my lovely wife. I love it to death but I agree with previous posts that I don't need to go higher for a lot of reasons. The d810 may well be my last camera.
     
  11. I second Michael B, what is missing is better HDR and improved creaminess, even broader ISO without degradation, yes it is technologically more difficult, than just having more smaller pixels; maybe 3 orders of magnitude between resolving a bright sky and the bark of a dark tree in the same shot very hard to capture the full range and tones. i was of the understanding that 35mm velvia resolution was more or less equivalent to a 50MP camera with 16bit color. The new 42MP Sony alpha 7R would appear to be moving towards both bigger and better, a very attractive camera, one would expect Sony not to waste the huge investment in tooling up for such a sensor and we should await what Nikon can do with it.
    "I will finally be able to take good photos when I get my hands on the next generation of gear"....Ansel Adams (or not)
     
  12. I'm with Glenn - the A7R 2 sensor - if it can do what Sony claims - is far more interesting than Canon's. It took Canon a very long time to feel that 22MP wasn't enough to compete with 36; Nikon may feel the need to do something for bragging reasons, but remember that the 12MP D700 was the competition to the 21MP 5D2 for a long time (excluding the low-volume D3x).
    Nikon's(/Sony's) sensor is already market-leading in terms of dynamic range. I'm not sure how much better we'll get, although the D7200 is impressive given its size. The lack of dynamic range is why the 5Ds sensor has absolutely no interest for me (other than only having legacy Canon gear these days); the dynamic range abilities of the D800 are what sold me on upgrading my D700, not the resolution.
    Of course, the A7R 2 sensor would still leave Nikon lagging Canon is the resolution stakes (and I've always felt it a bit unfortunate that you can't buy one camera with the reach of a 24MP DX sensor and FX coverage - not that I suspect Nikon mind selling a D7200/D810 pair). The resolution boost over the D810 is negligible, but it does mean you have enough pixels to shoot 8K video - or, more usefully, 4K video with full colour sampling at each pixel. I suspect that might be behind Sony's sensor development. Nikon and Canon haven't rushed to be at the cutting edge of video in their DSLRs (recently), but I'd really hope they get at least 4K within the next few releases. We'll have to see how good the latest generation of on-sensor AF is - and, generally, how the A7R2 behaves. Sony didn't have an established full-frame body (the A900 barely counts) and selling the D800 sensor to Nikon as an exclusive had to be in their interests; I'm not sure how fast they'll be to hand over the crown jewels now that the A7 series is a little more established.
    If we're going to have Canon envy, I'd rather talk about how old the MultiCAM 3500 is, and how it would be nice to have some cross-type sensors a bit farther from the centre (however good Nikon's tracking is these days). Maybe we'll see in the D5, when(/if) such a thing appears. For now, my D810 is a pretty darned good camera, and I don't feel I'd miss much with an alternative.
     
  13. It depends on what you're shooting. My old D7100 (12 MP) captured very nice portraits but when taking scenics in the Eastern Sierra Nevada, it showed its MP weakness. While a 24 MP D7100 is good for wildlife, it want't quite as good as my D810 in Jackson Hole for capturing detailed scenics. Of course if Nikon introduced a 50+ MP body, I'd be very interested. Then I could stop thinking about a Pentax 645 Z.
     
  14. ShunCheung

    ShunCheung Administrator

    If for whatever reason you need 50MP and beyond, a larger sensor area so that you get high-quality pixels is the real
    answer. The problem is that sensor cost goes way up along with real estate, and you'll need some medium format lenses.
    You'll still need to update that expensive sensor/camera every few years as the technology improves.

    Meanwhile, with the advent of iPads and other tablets, as well as PCs, images are not printed that much any more. At my
    home we mainly read magazines on the iPad and read news on web sites. I think the demand for high quality images is
    quickly diminishing.
     
  15. Actually, I'd argue pixel coins are starting to matter
    more. On a print-out, you need a huge print to make the
    difference between 24 and 50MP obvious, and not that
    many of us fill art galleries. On screen, you can zoom in -
    all the detail that the sensor captures can be seen. Not
    that I'm arguing for Nikon to try to get the resolution
    record off the 80MP medium format backs.
     
  16. I'm happy to shoot at 16 or 24 mp all day. Normally I'd say it doesn't matter to me whether Nikon makes a 50 mp camera,
    but remembering being stuck with the choice between keeping my D700 or upgrading to a D800 with too many mp and
    not enough fps, I'm hoping they don't. Maybe in a few years, when I have a new computer with a couple terabytes of SSD
    and a 30" retina display, and Nikon is using Sony BSI full frame sensors, and using my 150-600 lens has taught me much
    more about keeping my camera stable, I'd be ready to think about more mps, but to me these super high res cameras
    don't do anything right now.
     
  17. C.P.M. van het Kaar, Jun 21, 2015; 04:57 a.m.
    I'd say : Ask Nikon..
    I also know the answer from Nikon already :
    "It is our policy to never disclose information about future products or future developments " or some thing like this .. :)
    Agree, this would likely be the outcome. But I find it rather funny when Nikon responded with: "why would you need something like this ?" (not exact quote)....when I asked about any upcoming FF rig when only the D3 was out and the D300....and 5D on the other side. Around two months later D700 arrived. Hello ?
    To me, if I need more pixels, I'll rent it. For time being, I can't even get my D610 to perform like a real camera w/o issues (camera is in the shop for oil spots as if was D600). No need for me to derail this thread, but it would be nice that Nikon was more concerned with the product they currently make and enhance that, IMO.
    Les
     
  18. "why would you need something like this ?" (not exact quote)....when I asked about any upcoming FF rig when only the D3 was out and the D300....and 5D on the other side. Around two months later D700 arrived. Hello ?
    They were doing their job, which is to keep the sales of existing products from collapsing. E.g. Nokia's mobile phone business was utterly decimated by a guy who slipped that their existing OS was a burning platform and they would switch to a different OS which would only be available in a year or so. Near total loss of market share and arguably the products that eventually came out were worse in many respects than what they had developed mostly in house (N9). Only companies with no market share (in a given market sector) and nothing to lose pre-announce products well in advance. Notice e.g. that when Canon had an IS 70-200/2.8 already, and Nikon was some time from being able to manufacture theirs, they pre-announced the VR 70-200/2.8 to convince their existing users that they would eventually have a product in this crucial segment. They didn't strive to protect the sales of their existing 80-200/2.8 lenses because at that point probably they had no sales to protect. Same thing with 800/5.6 which was pre-announced an shown in London for testing ... the last 800/5.6 Nikon had had was a manual focus lens of another age. When Nikon does have a competent existing product that is still selling acceptably, there is not going to be any pre-announcement since their legal obligation to their shareholders is to make money instead of intentionally sabotaging their sales.
    To me 24MP is close to the golden compromise but I believe the D810's successor will use a 50+MP sensor. I agree with Shun that autofocus improvements are key in realizing better image quality in many real-world situations than even increasing pixel counts. One advantage of higher pixel counts is that the base ISO dynamic range increases (per sensor area) without the need to go to higher bit converters. However for me this is no longer important as the D810 is more than good enough in this aspect and I rarely get to shoot it at base ISO apart from landscape photography on tripod.
    I make quite a lot of prints (thousands of) per year but most of them are around 12x18cm or 20x30cm in size rather than the full capacity (A2) of my printer. I do make some large prints for framing on a wall, and to be honest I have no qualms about making such prints from 12MP FX images of old. 24MP and 36MP make the prints a little bit more detailed but this difference requires rather close observation and typically large prints are for mood and decoration and the details are not seen at the typical viewing distances. I do like the 24MP-36MP sensors and the added detail but the benefits mostly are in fine details when viewing the print at close reading distance, and in the flexibility to frame action situations a bit loosely and then crop to finalize the composition (I do this a lot with the D810). This technique is very valuable to me and the main benefit of high resolution sensors. However, 36MP is more than enough and next I would like to buy something faster and a bit better in the high ISO department.
    However, I understand there are more picky people who want to maximize the detail in their images and no doubt the D810's successor will move into that direction, although the relative improvement is likely not to be all that great in the real world. I think to make a noticeable improvement over the D810, 100MP would be a good target to go for (the next step 200MP, if anyone cares at that point). However, in practice we know that the camera industry prefers to make small changes and sell them frequently to make more money than they could by introducing only cameras that present large improvements. There is nothing wrong with this approach of course, and no one says that you must buy a new camera every generation (or even every second generation). I would like to see that people use their cameras longer and buy less frequently since that would result in less e-waste. Since the technology of digital image capture has almost reached its theoretical maximum in a given format size (in terms of signal-to-noise), improvements to sell new cameras will likely take place in other areas, such as improved AF and better integration with the online world of communication.
     
  19. Well, the 'camera industry' is already serving photographers needing more than what the 35 mm format based DSLR users are using. Has been for quite a while, but apparently unknown to most DSLR users. You really do need larger sensors. And that means larger cameras. Leica pumped up their camera to get to what is needed. But the market is still served best with digital backs for (existing or new) MF systems. And those have been available, with high pixel counts, for quite a while. Sony, Canon nor Nikon are driving this through their 35 mm based camera systems. So if you want to see what is happening, what is possible and what is available, don't look at an A7r II or 50 MP Canon.
    In other words: we do not need them to put 50, 100 or 200 MP sensors in tiny cameras. The industry knows that. Yet (and alas) the industry also knows that people will be paying lots of money for such sensors that make little sense in a 35 mm format based package. So it might happen anyway.

    Shun, we also do not " need to update that expensive sensor/camera every few years as the technology improves.". The thingies work, and will keep working until the day they break down due to being overworked, i.e. wear.
     
  20. It is fairly safe to say that no current Nikon lens challenges the potential of the D810 sensor. On the other hand, the D810 (and Sony A7ii) make Nikon lenses look about as good as possible. When compared back to back on the A7 agains Leica, Zeiss and Sony lenses, Nikon lenses do not fare well.
    A rule of thumb is that the lens must have four times the resolution of a sensor to take full advantage of it. At that point, any deficiencies in resolution of the lens constitute less than about 5% of the overall results. We are a long way from that, outside of special laboratory tools and monochromatic light. I have a fair amount of experience using Nikon and Leica lenses on a Sony A7ii, as well as lenses made specifically for that camera. Now that I have more time after to the end of the school year. I will try to assemble comparisons on a practical level, rather than resolution charts.
     
  21. we do not need them to put 50, 100 or 200 MP sensors in tiny cameras. The industry knows that. Yet (and alas) the industry also knows that people will be paying lots of money for such sensors that make little sense in a 35 mm format based package. So it might happen anyway.

    Irrespective of need, people will continue to buy higher and higher resolution small format cameras. It is what they can afford. True medium format (i.e. full frame 645 or larger) is prohibitatively expensive for most photographers (including professionals) and the crop medium format sensors lack wide angle options. Personally as I've said I would like about 24MP FX as a compromise as that's about consistent with my print needs and practical for handling volumes of images. But this won't stop people from craving more.
    It is fairly safe to say that no current Nikon lens challenges the potential of the D810 sensor.
    To see relative performance of lenses, see www.lenscore.org and click "Scores" and then "All lenses". First on the list is 85mm Zeiss Otus, the second is the 400/2.8E Nikkor. Third is Leica 50/2 ASPH, fourth the 55mm Zeiss Otus, fifth Nikon's 200mm f/2G II. Sixth the Canon 400/2.8 IS II. The tests are performed using a high resolution (200MP) custom sensor since so many lenses outresolve current camera sensors. Notice how of the top lenses if you require AF, Nikon and Canon do quite well for themselves; in fact you have to go down to position 22 to find an autofocus lens that is not made by Nikon or Canon. It's not for no reason why these are the top brands.
    So, please let me know which Zeiss, Leica, and Sony 200mm lenses beat the current 200/2 Nikkor. I would like a recommendation of one of each. I hope your lens recommendations are autofocus since I prefer to to work with AF when photographing people.
     
  22. I mostly agree with your short market analysis, Ilkka. With a few notes on being prohibitively expensive though.<br>Yes, MF digital is very expensive compared to 35 mm format based DSLRs. But professionals (usually) do not get stuff because of the wow factor, but because it makes economic sense. If it doesn't do that, obviously no expenditure. Stuff has to pay for itself. And not only for itself, it also has to pay for the bread and butter you live on. Else you are not in business, not a professional.<br>And thoughMF digital is expensive in comparison to even a D810, photography using decent MF digital actually isn't an expensive business to set up compared to other types of business. For instance try starting a simple bakery, with all the apparatus you need for that. Though a digital back may cost as much as a modest new car, it is not that expensive. (Yes, i know that there other types of business that are much cheaper to set up.)<br>The next thing is that it is a growth thing. You (mostly) do not have to start from scratch, pay for everything at the same time, but add what you need when you need it. The beauty of professional MF equipment is that it is modular. If you want to switch over to MF digital capture, all you need is a digital back to add to your existing MF kit. So though that obviously doesn't make anything less expensive, spreading the expenditure makes it all a bit easier.<br><br>Yes, people will be blinded by pixel count numbers and the believe that more and new = better and must-have. This thread's topic is a good example. Instead of asking and discussing when Nikon will catch up and overtake Canon, the thing that must be (and luckily is) discussed is whether anyone would really want or need that to happen. I'd say that 24 MP or 36 MP is not a compromise, but just about the pratical limit of what the type of equipment discussed (those 35 format based DSLRs) can do. More would not make sense for these machines, so if you really need more, you also need a different class/sort of equipment. Hoewever if it serves its intended purpose well, i.e. if there is no need for more/better, there is no compromise.<br>But yes, i agree that people will want to buy a 50 MP Canon anyway.
     
  23. It's true that making a significant difference in resolution requires quite a big jump. A 50% increase in pixel count is significant but not huge - I can see someone preferring to buy a D810 over a D750 because of it, all else being equal, but not so much upgrading from a D750 to a D810 just to get more pixels. On the other hand, the 3x improvement from a D700 to a D800 was extremely visible, especially with the strong low-pass filter on the D700. A 100MP successor to the D810 would certainly have a visible resolution advantage, but a vanishingly small number of my images would significantly benefit from it; on the other hand, my D700's 12MP were actually sometimes a problem.
    I maintain that the best justification for a jump in sensor resolution over a D8x0 is to get 4K/8K video. A D8x0 sensor doesn't have the 7680 horizontal pixels needed for FUHD (7680x4320). The A7R II sensor does. Arguably, better than the 5Ds sensor, because it fills more of the sensor area doing so. The ability to do perfect 2:1 sampling for UHD video is arguably more important than FUHD for the foreseeable future. The 5Ds sensor does have enough spare pixels to do cinema 4K (4096x2160) with 2x oversampling, though it wastes some sensor area in doing so; the A7R II sensor can't quite do that. I could understand Nikon making the jump to one of these resolutions once they finally get their act together about shooting UHD video - or just going with the minimal 3:2 7680x5120 (39MP) or 8192x5462 (45MP). A 100MP sensor (let's call it 12247x8165) would actively be worse at FUHD video resolutions due to the non-integer scaling (or reduced area if cropped), and would have uneven colour sampling even at UHD resolutions. 1080p televisions tend to look better than cheaper (and mostly obsolete now) HDTVs because the lower resolution panels tended to be 1366x768 rather than 1280x720 (for historical reasons), and were therefore incapable of regular sampling - or displaying a 720p input signal properly.
    So. Don't expect me to be shocked if there's a tiny bump from the 36MP D8x0 sensor in Nikon's future. I'd be a little more surprised if Nikon compromised the sensor in other ways just to chase after the 50MP figure. If they want more resolution, looking at the Hasselblad/Pentax multi-exposure sensor-shift approach gets them some of the way there. A 36MP camera is manageable for consumers, speaking as one. The same is probably true of a 50MP camera, though I think a good argument could be made that most people should stick with the 22MP 5D3. Going significantly higher is well into the realm of diminishing returns for most people.
    As for the necessary compromises... well, eventually people are going to start using faster flash storage cards, and prices continue to drop on quantity, but it's still not free. Sony's sensor changes indicate that read-out speed may not be a huge problem with future high-res sensors (though I'm reserving judgement until I see an A7R II tested).
    I only ever saw the D800 speed disadvantage compared with the D700 as being a problem to those shooting with a grip (who had the D4 upgrade path too...) - otherwise, the 1.2x crop mode got you back to 5fps and 25MP, and still competitive high-ISO behaviour with the D700. The D810 gets to 6fps with the same trick, which is near enough to the D750's speed that I don't consider the latter to have a significant advantage. I rarely find the need for 6fps, but I realise those spoilt by the "D3 on the cheap" D700 + grip are going to feel otherwise. I'd really hope Nikon can make a speed jump with their next sports camera - the 1Dx always had an on-paper advantage, and there are now a number of consumer cameras that make 11fps seem ordinary (notably the NX1).
     
  24. Really, why does anyone care about having the highest MP count? People seem genuinely concerned for no real reason that has to do with their photography.
     
  25. The LenScore resolution rankings for Nikon are dominated by focal lengths of 200 mm or longer. Zeiss does not have any offerings in this category, nor Leica or Sony except for SLR (ZA) cameras. We can get a better comparision if we restrict the inquiry to "normal" focal lengths for general photography, say 28mm to 105 mm. On this basis, 8 of the top ten in the LenScore list are Leica or Zeiss, with positions 8 and 9 occupied by Nikon - (Nikon 85/2.8 PC) and (Nikon 45/2.8 PC), respectively. None of the Zeiss or Sony-Zeiss lenses designed for the A7 were listed.
     
  26. Well, Leica has the older R lenses lenses that were amazing, but they are all manual, I believe.
     
  27. The counter would be in spirit to keep these threads going. Other than that, why?
     
  28. i have heard theyre even going for 73mp in the next nikon d500x with 37 fps!
    oh my
     
  29. we do not need them to put 50, 100 or 200 MP sensors in tiny cameras
    Personally i agree to that, but the market works different...
    In General the old saying still goes :
    - "The availability creates the need"
    And the industry knows that verry well...
     
  30. All things being equal wouldn't we would all choose a camera with more pixels than less? If there was a D610 with 50MP or a D610 with 24MP at the same price with the same everything else I would have gone with the 50MP version. However if there was a D610 with 24MP with better dynamic range and 16 bit color vs a 50MP with 12 bit color and normal dynamic range, I would choose the 24MP version.
    Maybe the market for future cameras will split into the a hypothetical Nikon D900 "Art" and the D900 "Sport". The art series , slow, but increased HDR, 16 bit color with real creaminess, optimized for best quality portraits and landscapes printed to paper vs a sport optimized (mirrorless) for high speed shooting ( 32fs in 8K), AF capture and files optimized for online jpeg conversion ( 8 bit color) . It will be interesting the if the next D5 vs the D800 successor shows this dichotomy of photographic outcomes. That being said no company is going to willingly tool up two very expensive production lines just to fabricate two sensor types if they can possibly avoid it.
     
  31. All things being equal wouldn't we would all choose a camera with more pixels than less?​
    No, because things inevitably will not be the same. Suppose that on a pixel level there is no penalty to pay for having physically smaller pixels, and we're comparing same generation silicon and software, then more megapixels will leave me with larger files, increased requirements for processing etc. Given that my uses never ever come close to needing all the extra megapixels, I would be perfectly happy with the lower MP version too.
    And given that things really aren't equal, the lower MP camera makes more sense for most of us; I do wholeheartedly agree with QG analysis. There is a niche market that needs more, and they're better served with physically larger sensors to get there. The rest of us.....let's be realistic there: nearly all people I know (in person) with higher megapixel cameras than my D700 do not print, or print less than 10 photos per year at a size larger than A4. So, really, there is a point where more pixels really stop being useful for the mass market.
    split into the a hypothetical Nikon D900 "Art" and the D900 "Sport".​
    Like the D3 versus the D3x; the D4 versus the D800? Nikon already does this.
    What is this extra creaminess that is mentioned a few times in this thread? I know some lenses to which it would be an applicable adjective, but a sensor? What does that have to do with it?
     
  32. It is hard to explain creaminess with examples from a computer monitor, what I am meaning is the seamless "creamlike" tonality you see in the normally professionally printed photography at exhibitions etc, normally from medium or large format film cameras. I dont think it is only a lens factor, more a factor of the media/sensor that receives those nicely aligned photons.
    Of course anyone can argue that difference that doesn't make a difference is not a difference: if there is no visible difference between photographic outcomes in selecting a 50MP sensor over a 24MP sensor then why bother? But of course reality is that there will be a difference with more pixels, the ability to crop/zoom, extract fine details etc that make me want a 50MP sensor over a 24MP sensor (all things else being equal).

    The "compute" factor however I don't think is such an issue, disk is phenomenally cheap a 50MP image is going to consume around 50MB, 2TB disk sells for US$80 or US$160 for two so you have a backup, around 40,000 images or 1cent each. My laptop has 8000MB of RAM and 4 CPU cores, more than memory more then enough to process a single image or even to blend a panorama, as would any computer produced in this decade and likely to be owned by a person who buys a top end camera system..
    Processing 4K at 60f/sec however is the desktop challenge.
     
  33. more a factor of the media/sensor that receives those nicely aligned photons
    the ability to crop/zoom, extract fine details etc that make me want a 50MP sensor over a 24MP sensor (all things else being equal).
    Mm I think there is a little issue here.... , it is virtually impossible to create a lens that will always transmit an image to the sensor where all channels ( be it RGB or CMYK) will be nicely alligned within 1 pixel, this becomes more difficult whn the density of the sensor increases, Meaning , with current technology, that cropping from a 50mp sensor might not always give you a better quality image then it would when utilizing a 24mp sensor, especcilly when the crop factor increases . THer is a break-off point somewhere i guess. Possibly the minimum aperture will play a role here too, diffraction from aperture will be more visible when using a higher density sensor ...
     
  34. I think the problem is not having good enough lenses (expensive), but I know that in 99,99% cases it`s very difficult to capture 100 lp/mm (equals 7200 x 4800 pix in 24x36 sensor). You have to have a strudy tripod, no AA-filter in front of the sensor and optimal conditions.
     
  35. Esa: That's really not true. Is the pixel-level contrast as high with a 36MP FX body as it is with a 12MP DX body? No, of course not (although the strong AA filter on the D700 really compensated for that). Can you resolve detail with "good" contrast? Yes. Pick a suitable aperture and a fast enough shutter speed, and all is more than acceptably sharp. Run it through mild image processing (usually DxO, in my case) and there's a heck of a lot of detail there. Yes, I tend to stick to the f/4-f/5.6 range more than I would with the D700, and I use good lenses, but daylight is plenty to get you to a suitable fraction of the shutter speed for sharpness, especially with VR. Could I get slightly better sharpness with a long exposure on a solid tripod with good lens support? Of course. But not by enough to be typically visible.

    The D8x0 cameras are more demanding of optics and technique than the 24MP FX bodies if you want pixel-level sharpness, just as they're more demanding than the D700/D3. Not much, but a bit. The 5Ds R will be a little more demanding still - but still only a bit. Nikon and (now) Canon make 24MP consumer bodies, and they're capable of decent detail resolution as well. If it was that hard to take a decent photo, all consumer bodies would still be 2MP. Getting pixel-level detail out of a 100MP body will be harder than it is with a D8x0, but only about as much harder as a D8x0 is compared with the D700.
     
  36. ShunCheung

    ShunCheung Administrator

    All things being equal wouldn't we would all choose a camera with more pixels than less? If there was a D610 with 50MP or a D610 with 24MP at the same price with the same everything else I would have gone with the 50MP version.​
    Of course not, because things will never be equal. As I pointed out earlier, I added a 24MP D750 in addition to the 36 D800E precisely for fewer pixels, which is what I prefer in most situations. For the occasions that can benefit from 36MP, I still have the D800E.
    As I pointed out earlier, when the pixels get smaller and smaller, diffraction will set in earlier and earlier, wiping out all advantages from the extra pixels. You will have more pixels and therefore much larger image files and frequently slower frame rate, but diffraction will prevent you from getting more details in your images. As a result, you will be carrying the extra pixels and large image files. You'll have bragging rights for having more pixels, but in reality, your images will have no real advantage over those from 36MP cameras.
    Additionally, whether your subject indeed have more details to be captured in the first place, photographer technique, lens quality ... are all issues.
    I used to recommend the 17-35mm/f2.8 AF-S over the 14-24mm/f2.8 AF-S due to the 17-35 being a more useful zoom range. However, I noticed in 2010 that on the 24MP D3X, the 17-35's edge performance was no longer acceptable near 17mm. Back then it wasn't a serious problem since very few people had the $8000 D3X. As soon as Nikon introduced the popular D800 and then D600 in 2012, every time I have to quality my recommendation of the 17-35mm/f2.8 nowadays. Some lenses are not up to the job for even 24MP, at least in some part of the zoom range and in some apertures.
    If you indeed need more pixels, you cannot keep squeezing them into 24x36mm. That is why you need larger format so that you can once again use f8, f11 without diffraction problems.
     
  37. "Sigma does a better job in marketing, ......this way Sigma creates a market for their products before they are even starting production ..." Maybe this is why I buy more Sigma lenses than Nikon.
     
  38. More pixels mean larger prints, which may not actually mean much in a practical sense. It also means fewer problems with Moire on cloth and other repetitive patterns. Finally, you can crop more and still have enough pixels left for larger prints. This reduces your dependency on large, heavy and expensive zoom lenses. Of course smaller pixels tend to have more relative noise than larger ones, hence smaller dynamic range and lower maximum ISO. These are design limits, rather than intrinsic, so not all pixels are equal, and the game is continually improving.
    Diffraction in a lens is proportional to the relative aperture. A 50mm lens and 80mm lens have the same propensity for diffraction at the same aperture (f/11, or whatever). Longer lenses have larger openings, but the exit pupil is further from the focal plane, resulting in the same spread. Sensor resolution has no effect on diffraction, except you may see problems with the lens more clearly.
    https://en.wikipedia.org/wiki/Diffraction-limited_system
     
  39. Shun,

    “If you indeed need more pixels, you cannot keep squeezing them into 24x36mm. That is why you need larger format so that you can once again use f8, f11 without diffraction problems”

    If you are having the following scenario:

    Someone wants f8 on 50MP FX camera but doesn’t like the amount of diffraction softening at pixel level (because f8 airy disc is 10.7 microns and the pixel size is 4.2).

    Are you proposing they should switch to a larger format for larger pixels so they can take their photo at f8 and reduce the f8 diffraction impact as pixel size is bigger (for example MF 54mm x 40mm pixel size would be of 6.6 microns for 50MP)?

    But this f8 on a larger format has less depth of field (assuming the same field of view in the final photo) so more out of focus blur in your final photo.

    So you stop down to get your DoF back. You would need f12.3 on the medium format example. That gives you an airy disc of 16.6 vs pixel size of 6.6, that’s the same ratio as airy disc of 10.7 vs pixel size of 4.2.

    If you want the same final photo it looks like you have gained nothing in terms of the diffraction softening artefact. So I’m not sure going to a larger format will get you round the diffraction problem
    00dM2E-557295584.jpg
     
  40. We're ignoring quite a lot of effects here. Firstly, we're assuming that the pixel quality drops off as the pixel count increases. Historically, that has absolutely been true. With the 36MP Sony sensors, it's only arguably true (the per-pixel dynamic range is very good, although the high-ISO performance is slightly worse). With the stacked approach of the A7r II, there is a chance that even this difference will have gone away. Or not, we'll have to wait for reviews.

    File sizes are bigger if you record the full sensor resolution. Of course, maybe you don't need to do that - if the sensor resolution is high enough, you can bin down to smaller sizes. Nikon (and to an extent Canon) made such a pig's ear of small RAW that the concept seems unappealing, but logically there's no problem, at least if your debayering software knows what it's doing (and if Phase One don't sue you because of Sensor+ - though I maintain what it does is obvious). Do this and moiré is better than using a camera at lower resolution.

    For a fixed resolution of output, sensor resolution can currently hurt. If you're capturing a 1920x1080 signal using a D800's sensor, and if you're doing so by point sampling the sensor at the relevant locations (I don't believe it's that bad, but the argument holds), you're reading something like one in 14 pixels from the sensor (7360/1920 is about 3.8), ignoring the aspect ratio crop. No wonder high-ISO video from a D800 sucks. On a D750, that number is nearer one in ten. But offer decent pixel binning (or more read-out) and you can use all the sensor area. You can also reconstruct video with full colour data at each nominal pixel. And, again, I think video is what's going to push the sensor resolutions, at least a little.

    Then there's the 'nobody makes 40" prints' argument. True, they don't. But most people don't peer at a 40" print from right next to it (unless it's been shot with an ultrawide and they're going for immersion). People do pan around images at the pixel level digitally, because they can. Yes, images on Photo,net are normally 700 pixels wide, and a lot of online snaps in social media are small (and look bad if they're shot bigger). It was only recently that I established that Instagram is now a social media network popular with kids, as opposed to a bit of software for unpredictably messing up a perfectly good photo. (Speaking of, I saw Jurassic World recently. One of the kids seemed to be trying to document everything with what looked alarmingly like a Holga. Seriously, you go to a park full of dinosaurs and your choice of recording equipment is a blurry thing with almost no exposure control? Bloody hipsters. Maybe I imagined it, and it was just a digital camera inexplicably styled like a Holga.) Anyway. There are plenty of online ways to share big images. Sometimes very big images. I've seen a wall-sized Ansel Adams large format print - very nice, but certainly not all that sharp. Even UHD and 5K monitors are starting to become more common.

    So, all else being equal, let's have more pixels. Let's just make sure that all else really is equal. It can be.
     
  41. Jon: Larger formats don't fix diffraction (at the same depth of field), although I did have to get over a brain fade to persuade myself of this. What they do is let you use a smaller relative aperture to do it, so a lot of optical aberrations go away. Given a 5x4 camera with a 150mm f/2.8 lens, you'll get an image that's in the same ballpark as a 50mm f/0.95 (or maybe 0.8). The 35mm lenses that are that fast? Kind of specialist, and really not very good at wide apertures. It's much easier to build an optically good f/2.8 lens, even with large coverage, than it is to make an f/0.8 lens. The same applies even at more reasonable apertures. I presume lens availability has historically been a factor in the look people expect - people tend to think of medium format cameras as having a shallow depth of field and little film grain because that's what you get if you use the same relative aperture as you would on 35mm. If both lenses are f/2.8, that may be all you could do. But since there aren't so many f/1.4 lenses with medium format coverage, medium format no longer has the advantage on this - though 135/FX still tends to have the edge on DX and below.
     
  42. Nikon routinely telegraphs new products by reducing the price of existing products which would compete for the same market niche. In general, they don't announce products months ahead of their release, but they seldom introduce revolutionary changes either. Consider that 36 MP (the 800/810) has only 20% more resolution than 24 MP. the 50 MP Canon is only 40% "better" than 24 MP.
    I would have jumped for the Sony A7Rii for increased high ISO performance, better AF and a silent shutter even if the resolution had remained at 24MP. For me, 42 MP is a bonus not a deal maker. (I won't turn it down, considering it doesn't have an AA filter, which sweetens the MP advantage considerably.) I was not turned on to the D800 or D810 because of the size and weight, and the comparison of Nikon lenses with my experience with Leica (later Sony/Zeiss) beginning last summer. If the rumored "D900" with 75 MP were announced next week, I would not change course unless the package included some metaphysical features.
    Early marketing is a sound business practice, as long as you keep your reputation for delivering on time. When Microsoft failed to deliver a new Windows version, they became an industry icon for "vaporware." It establishes your place in line, should the competition have a similar product available sooner. If you like Nikon (Leica, Zeiss, etc.), and accept the price, earlier availability from Sigma or Tamron will probably go unheeded.
     
  43. Andrew, I was really only pointing out that larger format doesn't appear to fix the diffraction problems.
    I agree that for the shallowest DOF large format may be easier but obviously there are plenty of other considerations when choosing a system especially if going to the extremes of the focal lengths and apertures on a system (size, weight, cost, build quality, plus the modern additions like auto focus systems, VR etc).
    I don't know much about large format but it looks like 150mm large format lenses are often f5.6 (so a 50/1.4 in FX?) and a 2.8 one not so common. The only one I came across (not that I looked hard!) was a 150mm Xenotar f2.8, "It weighs seven times as much as the Fujinon-W (f5.6), and if you work with it on your camera you get the feeling that you are budging a spacecraft through a swamp"!
    I guess there is always a compromise needed somewhere depending on your own priorities.
     
  44. Very good thoughts, opinions and answers. Personally I would ask "When will Nikon counter the amazing Sony (Alpha) A7R II" Incredible camera...Sony is really on track with the shape of things to come!
     
  45. Why did Sony elect to release a camera with 42 MP, not 50 MP? After all, Sony makes 40% of all the digital sensors on the market, including Nikon and medium format cameras. According to their senior manager of the digital imaging goup, 42 MP is the ideal size to record full frame 4K video with minimal binning, and video is a big part of mirrorless future. The A7Rii records 4K directly to memory, not to an external recorder like the A7S, 4,2,2 quality at 100 MB/s - broadcast quality.
    I have used my A7ii to record video in parallel with other HD video cameras, and found the results to be compatible and easily matched and mixed. It's a little tricky to handle the 29 minute maximum per clip (stop and restart) and batteries that last just under an hour, but I did it and will no doubt do it again when I need the footage.
    http://www.dpreview.com/articles/0717419525/interview-kimio-maki-of-sony-the-customer-s-voice-is-the-most-important-data-for-me
     
  46. it is a very well explained article and easy to cross reference wether or not it is good.
    it is damn good actually.
    i do not want to quote or write as if it was my knowledge.
    i got a part of it from him and i think he makes a very valid point,
    especially intersting to this conversation.
    read it.
    http://www.wildlifeinpixels.net/blog/sensor-resolution/
    at the end of the article he writes:
    • Greater sensor resolution enables you to theoretically capture greater levels of detail.
    but that extra level of detail is somewhat problematic because:
    • Diffraction renders it ‘soft’.
    and that softness i have seen in more than one exhibition.
    lots of people seem not to be capable of handling such a camera as the d8x0.
    now imagine the disaster with 50...
     
  47. Here is the thing I get confused, all things being equal , a sensor with even more pixels will be at worst no benefit, how could they be result in a deleterious soft image relative to a fewer pixel sensor?. Imagine a hypothetical "35mm" FX 250MP (18000x 12000) sensor in which clearly pixel size is so small diffraction will occur, how can the imagines is produces on my 14-24mm F2.8 be "softer" than a 25MP sensor as I have in my D610? I can accept an argument that with diffraction all those pixels may not render a sharper imagine, but I cannot possibly think that having 250MP would result in an overall softer image once RAW software developing crunches the file. The new samsung galaxy 6 has a 16MP camera, the sensor must be tiny relative to a FX sensor, if the galaxy 6 sensor with the same pixel density was expanded up to FX size it would have to be a 100MP sensor. I cannot imagine that such a supersized Galaxy6 sensor, all things being equal is going to give a worse image than a 24MP sensor.
     
  48. Glenn, I think that as the pixel count increases so does the size of each pixel decrease and therefore amplification must increase. That can lead to softer results due to noise reduction and other internal processing. Mobile phone images look quite soft at the pixel level - the added resolution is wasted and due to the additional processing end up looking worse.
     
  49. read the link i posted...
     
  50. Edward: I was completely guessing about the 4K/UHD/FUHD thing, but I was hypothesising about sensors in general. It's a shame that there's confirmation in that article that the A7R II can't read FUHD from the full sensor and use 2x downsampling to get 4K output - there's specific mention that it has to line skip. Mark Weir says it can read every sensel from the Super35 crop, but that's 5168x2912 - not enough to oversample UHD. I'm not really clear why more than FUHD (7680x4320) but less than 4K (8192x4320) is "ideal" for video, either. It's a shame that they confirm the A7s still has the low-light edge, and that they currently still have the lossy raw compression. The A7R II still looks like an interesting camera, but I wouldn't be surprised to see more things done in a similar sensor resolution in the future.

    Glenn: You're quite right. The issue with smaller sensor sites is noise in getting the data off the sensor, slower frame rates, and arguably some issues with well size (the sensor sites saturating). You tend to lose some sensor area to the electronics as well, but gapless microlenses, BSI sensors and stacked sensors mitigate that a lot more than was the case a decade ago.

    Norbert: Diffraction can be compensated for. You can either use a larger aperture, and potentially stack images to deal with the depth of field reduction, or rely on deconvolution (which I believe DxO does, otherwise I'd expect the last landscape I shot at f/22 to be much softer than it was). Like any sharpening, deconvolution does tend to make bokeh look slightly worse and it accentuates noise, but it can be handled in moderation. It's one benefit to Nikon sensors with a high dynamic range. Have a look here, here and here.

    Ian: As you say, the problem for mobile sensors is that, with a few exceptions, the sensor size is tiny, so it's not getting much light. The denoising applied is typically quite extreme, though there's a lot of work that goes on to produce the best mobile images (because all the reviews compare JPEGs!) and the resulting need for sharpening and denoising does produce the kind of per-pixel details best described as "ew". That said, give them enough light and most mobile phones can do a surprisingly decent job - as can many compact cameras. I'm not going to deny that, at pixel level, a bigger sensor helps. How much going above the 135 sensor size actually matters is another argument - even with a collection area advantage, economies of scale mean that there's usually more development going on in the consumer sizes (though Sony appear to have transferred some of that to the 50MP CMOS sensor seen in several medium-format devices). Of course, that might be another reason Sony went with sub-50MP in the A7R II - to avoid making their own medium-format sensors look bad!
     
  51. but less than 4K (8192x4320)​
    Isn't that 8K rather than 4K? Sorry, I am not familiar with the video terminology...
     
  52. Sorry Dieter, I was getting my UHD and FUHD mixed up with FUD. :) Yes, typo, I meant 8K (and convenient for downsampling to 4K).
     
  53. ShunCheung

    ShunCheung Administrator

    If you are having the following scenario:​
    Jon, I am not too concerned about that scenario altogether.
    My point is that when your pixels are so small that diffraction becomes a serious concern @ f5.6 and f4, and you typically need to stop down a good lens by a couple of stops to get the best out of it, for a lot of lenses, there is no more "optimum" aperture any more, regardless of depth of field.
    But even though you are willing to spend money, medium format lenses are typically not as great as 35mm-format lenses. Medium format used to get away with that due to the much larger film area.
    With digital, there are other solutions such as stitching and/or focus stacking multiple frames into one to gain more pixels or depth of field. They are not simple, especially when your subject is not completely still.
    But the question is still why you need so many pixels. Huge prints are usually viewed from far away such that the quality doesn't have to be that good. I tell myself that I need the potential to make huge prints, but after having a D800E for 3 years, it remains as potential, as I haven't made even one huge print. Since I sometimes test lenses, I also would like to have the most demanding camera body around. But for the most part, I am afraid that it is mainly about bragging rights.
     
  54. OK imagine the engineering gloves are off, but no magic physics, light waves/photos are defracted and optical laws must be obeyed, what would be the "ultimate" FX (35mm) sensor be? ie. the engineering end point as far as MP , ISO and dynamic range, above which would serve no advantage in incremental improvements?

    He who cannot be named in this forum (Ken RockW) has a mouse rollover comparison between detailed the 5DRS and 5D MIII images and unless it is a fabrication ( or more kindly "fictional" comparison) it does seem that there is a noticeable improvement between the resolution of images of a 5DR (50MP) and the 22MP.
     
  55. 4K video is the new standard for theatrical productions. It is still relatively rare in home theater settings, but that may change as prices drop. Resolution and binning are not the only concerns, nor even the most important factors in video quality. The A7ii cuts 4,2,2 HD (2K) video (10 bit) at 50 MHz bandwidth, which is the cusp of broadcast quality. The A7Rii gets the same quality for 4K up to 100 MHz.
    The Sony 7S is certainly a competent camera, but for a certain niche. Those who need high ISO performance, excellent video and a silent shutter find the A7S indispensable. Others I've known were not satisfied with the low (by Sony standards) 13 MP resolution, and either took a pass or traded it for an A7 or A7R. However the A7Rii records 4k internally to SDXC cards, whereas the A7S requires an external recorder via HDMI.
     
  56. Glen: Can we call the wavelength of "red" ~750nm? That's pushing it, but gives you 48000 x 32000 pixels on an FX frame. That's about 1.5 gigapixels. Somewhere around that point, you're mostly capturing interference patterns. Of course, "capturing interference patterns" means "recording holograms", so that's not necessarily a reason to stop...

    Shun: I'm going to keep saying that prints don't matter any more. With a digital image, you can zoom in as much as you like and look at whatever detail got captured. I've absolutely spent time panning around an ultrawide image to see who was there, and pixels help. It's not so rare, although I admit it's not relevant to absolutely every image.

    Edward: Yes. 4K is starting to matter, and UHD TVs are no longer all that rare. I'm interested to learn how much use people are getting from the A7S - I've been contacted by friends who wanted a camera for very low-light video, and it was my recommendation (partly because ISO is so relevant when you need a large depth of field for video). 4K recording in general seems to be a pain, and I'm not blown away by the external recorders - a Shogun is far from cheap, and you've already thown away some detail by the time you've used an HDMI cable.
     
  57. Norbert - great article.
    Everything in photography is a trade-off – you can’t have more of one thing without having less of another.​
     
  58. ShunCheung

    ShunCheung Administrator

    Everything in photography? Not only that; everything in life is a trade off.
     
  59. Besides large scale projection, the driving force behind larger sensors for 2K or 4K video is a shallow depth of field. This is both for popular dramatic effect (pulling focus) and because it more closely resembles 35 mm cinema in that respect. By the time you stop down to f/4 or f/5.6 on a 1/3" or 1/2" sensor, the DOF is so deep you hardly have to focus at all. I'd like to know how 4K is useful for local TV and ad production, much less home movies.
    A Sony A7S (Canon, or whatever) is an inexpensive way to get the large-sensor effect of an Arri or Red camera. A one-man operation with an harness is a lot cheaper than a three man camera crew plus grips. The next time you feel pinched buying a Canon L lens, take a look at what real cinematic lenses cost. Zeiss has an "inespensive" line of primes which start at about $5K. Zeiss zooms are really expensive, $16K and up, and they're considered the low end of this market. The Canon C300 ($20K) is becoming popular for TV and action movies, instead of a repurposed DSLR. It is configured for either EF or P lens mounts, and has lots of expensive production related accessories.
    Noise is not nearly as visible in video due to its random nature and persistence of vision. It's much more apparent if you stop the frame. Noise is further "averaged" out if you use a 3-chip camera. My Canon has three 1/3" sensors (each 1920x1080), and can be used at 24 dB gain (about ISO 3200) without obvious noise.
     
  60. Megapixels are easy to count, noise level not so easy.
    In many cases, it might be that low noise is more important.
     
  61. Andrew, displaying high resolution digital images on your computer (and zooming into the details) may work for you personally but how do you disseminate digital images in such a way that your audience can see the details, without making physical versions of the images? Also if you allow the high resolution digital image to be seen on the viewer's own device, you lose control over the future use of the image.
     
  62. Ilkka: URL?

    It's always possible to scan or recapture an image from a physical version, with varying degrees of inconvenience. Whether it's convenient or practical once we start talking gigapixel images is another matter, but a streaming dynamic loader/zooming browser that never stores the full-resolution version of the image on the viewer's computer is arguably in a better position, from a "hard to copy" perspective, than a large sheet of glossy paper. Copyright is copyright, and there's very little you can do against the determined infringer that doesn't have a much greater effect in annoying the legitimate viewers.

    Of course, I'm an amateur and don't sell my images (although I do ask for a copyright message to stay attached when anything ends up in the press). I may think differently if photography made, rather than cost, me money.
     
  63. Today I read the Amateur Photographer review - they tested both versions - and the cameras do give good detail but dynamic range was not up there with the current Sony sensors and at the end it said the 5D Mk3 was still the better camera for most users. The review said a lot more than that and was interesting to read so well worth picking up a copy.
     
  64. Thanks, Ian. I'm now subscribed, but I've been too busy to actually read an edition since December (my wife got me the subscription for Christmas). I'll make a point of catching up with this one - but I think we knew about the dynamic range already. On that note, I should have answered Scott's original question with "March 2012", when the D800 came out.

    And I've always maintained that the 5D3 is a very competent camera (not that I want one), though I believe the D810 and D750 make up for most of the disadvantages of the D800 relative to the 5D3 while maintaining most of the advantages. That said, at the same price, I believe the 5Ds has a number of (small) non-resolution improvements over the 5D3, so I'd actually claim it's the better camera than the 5D3 unless you care about the tiny fps difference or the storage size requirements from big images. But it's hardly clear-cut, unlike the 5D2/5D3 upgrade where everything improved. I'll be interested to see what the 5D4 looks like, when there eventually is one.
     
  65. This whole megapixel race is getting completely out of hand. When they put out sensors which exceed the resolution limits of the lenses, it is just pi**ing in the wind
     
  66. We do need sensors that exceed the resolution of our lenses. Then we need to worry about one thing less (untill they find a way to make better lenses).<br>We do not need sensors that exceed sensors that already outresolve our lenses. That would be silly.
     

Share This Page