Canon Announcement

<p>It said that Nikon fans are still sensor envy. Even many said they don't need 50MP it's still a form of sensor envy. The Canon guys aren't talking too much as they are busy finding ways to buy the 5Ds.</p>

 

 

<blockquote>

<p>The Canon guys aren't talking too much as they are busy finding ways to buy the 5Ds.</p>

</blockquote>

 

 

 

Nah, just busy shooting with what we have. Far better than thinking about spending money.<a name="pagebottom"></a>

 

 

<p>I am afraid that the real "problem," at least for the camera manufacturers point of view, is that for still photography, digital cameras have reached a plateau. There are now far more pixels than 99% of us can use, so are the high-ISO results, dynamic range .... It took Canon three years to catch up with the D800, and Nikon is releasing the D5200, D5300, D5500 without a whole lot of improvement from one generation to another ....</p>

<p>That is exactly why we have threads such as: <a href="/casual-conversations-forum/00d6j4">Digital camera sales continue to decline</a></p>

<p>So unless there is some new innovation in video capture or 3D still capture to generate fresh demand on new cameras, a lot of us are already happy with what we have and won't be buying new cameras, certainly not at the same rapid upgrade pace as we used to do between 2004 to 2012 or so. Nikon has already put their top-of-the-line AF module on the very much consumer grade D7100 back in 2013, and the D7100 is now selling below $1000. What do they need to do to stimulate the next round of upgrades?</p>

Andrew, being able to see the edge of the aperture, or not, is not important. The only thing is that the aperture has an edge and diffracted light will spread over the projected image. The more edge there is relative to the size of the image, and as known, both do not grow (or shrink) equally fast with a change in radius or diameter. Etcetera.<br><br>When digital photography first arrived we were told that we needed special lenses, that the good ones we used on film would not do. That may have been so, due to the restrictons of sensors. But it was, of course, marketed such that 'special' had to be understood as 'better'. Meanwhile the demands on image quality digital capture posed were laughable. Terrible quality, helped to be bad quite a bit by Bayer patterns and low pass (i.e. soft focus) filters.<br>I found it not surprising, but still somewhat disheatening to see how now that sensors have become much better, people appear to believe that they need 'even better' lenses to do justice to the quality of the sensors. No, we don't. We need the good lenses back again.<br>And no, we did indeed not only have good lenses back then. We had the same balanced mix of good and expensive, cheaper and less good, and cheap and pretty awfull. No change there.<br>The same peculiar notion (that we enter into something new, never seen before, I.e. that we now have a new set of "modern standards". We don't. We are leaving the era of widely accepted but terrible to merely bad digital capture quality) is present in this diffraction thing. "f/8 and be there" has always been bad advice (for a number of different reasons). Good lenses and a good capture medium always have revealed that stopping down traded image quality off for depth of field (and though it would be better if image quality did not suffer, that is not a bad thing).<br><br>That the 50 MP Canon gets more attention in the Nikon forum than in the Canon one could mean a number of things.<br><br><br>Shun, there is a need for more pixels. Perhaps not in the (computer screen viewing) 'mainstream', but they can certainly be put to good use, or even be quite simply necessary. Remember that (even though Andrew still says that 6 MP is just about what 35 mm film could deliver) film had quite a lot to give. And there was a real need for medium format (you can get about twice what Canon now promises out of a well scanned bit of MF film) and large format. People did not spend what they now do on a decent 35 mm based DSLR and lug arounbd those heavy machines just for the heck of it. They delivered what was indeed needed.<br>35 mm based DSLRs have indeed reached a plateau. Cramming 50 MP on a sensor that size is no good. We still need bigger cameras for that more demanding work. Alas, i might add.<br><br>Interesting notion (also peculiar to the Digido) by the way, that camera sales only depend on upgrades to the even more new and marvellous, such that a decline in sales has to be attributed to a lack of that even more new and marvellous.

<p>Shun: Thom Hogan would claim that (decent) connectivity is the "big thing" that would make cameras sell again. Not being a social media person, unless you count this forum, it actually wouldn't do much for me, and nothing much is going to make a full-frame sensor and a 70-200 f/2.8 easier to carry (though I have vague thoughts about diffractive optics/fresnel lenses and collapsible mounts - fingers crossed for the 24-70 rumours).<br />

<br />

QG: I'm afraid you've lost me on the argument about the edge of the aperture. If you'd like to educate me off-line, I'll be interested. But to your other points...<br />

<br />

There were three factors to "digital" lenses. One was multicoating the rear element, because digital sensors are (and especially <i>were</i>) more reflective than film, and you get light bounces off the back. I don't think it was ever claimed it made much difference otherwise. There were changes to the telecentricity of lenses, because digital sensors tend not to like light entering at an oblique angle - although in the absence of infinitely-thin film, that ought to help sharpness a little even with traditional media. And the crop-sensor lenses were never more than a cost-saving.<br />

<br />

I don't recall anyone really saying that these were "better" than film lenses, just that they deal with limitations of the sensor technology. But I'm interested in these "good old lenses" that are good enough now. There are always exceptions - e.g. a traditional 50 f/1.8 is pretty sharp by any standards at the right aperture, though the "right aperture" is absolutely not "wide open" - but the lenses I mentioned (80-200, 180 f/2.8, 135 DC) were, as far as I know, regarded as "good" lenses. Certainly the 500 f/4 AI-P I own was thought of as pretty good, but I still had to stop down quite a bit on my D800. Nobody's going to claim that an 18-200 is optically all that good, just as I never claimed the 28-200 I used on my D700 was more than "better than you'd think", but the reason there are new lens designs is that the old ones really don't hold up.<br />

<br />

I didn't say you couldn't get more than 6MP from 135 film. I've heard 35MP quoted as a plausible figure, and I vaguely believe it, at least of the best monochrome films in test shots, and depending on where you define the contrast extinction. (On the other hand, some films that have been scanned at more really don't show the resolution the owners think.) I've struggled to get much useful information beyond about 6-12MP out of a consumer or drug store scanner, but certainly you can get more from a drum scanner or a decent dedicated device. But I stand by my argument that, along with the detail, you tend to be scanning lumps of grain at that point, making noise reduction much more of a pain than on digital - at least with something like Velvia (I've never scanned PanF). Larger formats are somewhat better, not least because the grain size is smaller relative to the film area (at the same speed) - which is partly why I felt happy getting more detail from my Pentax 645 (even in a 17MP cheap scan) than from my D700. I think the D800 keeps 645 very honest, though (but I still use Velvia for the spectral response sometimes). If anyone wants to give me a Mamiya 7, I'll happily do more film shooting - I do still have my eye on a 5x4.<br />

<br />

But I'm not really prepared to claim that a 50MP 135-format sensor is "no good", at least until there's more testing. The D8x0 series are <i>very</i> good; maybe not quite 80MP-medium-format-back good, but for convenience versus image quality, there's no argument. And if convenience weren't a factor, there would be more scanning backs out there.</p>

<blockquote>

<p>By the way, do you think it says something that there have been more posts in the thread about the 50MP Canons on this forum than on the thread on the Canon forum? :-)</p>

</blockquote>

<p>Haha, Andrew I was thinking the same thing! I guess it means we have more gearophiles over here than in the Canon forum?</p>

<p>If you look at the Photo.net Community page.. http://www.photo.net/community/</p>

<p>You can see that the Nikon Forum has more new posts (~gas-bags...;-0) than all the other equipment fora, not just Canon!</p>

<p>Spot count Nikon 351 v Canon EOS 217.</p>

<p>Mike: Well, I do my part. But I've been posting in that Canon thread too, to balance things out. :-)</p>

Andrew,<br>Diffraction is caused by the edge of the beam, making light go in all directions, overlaying the image produced by the unaffected part of the beam. As long as there is little diffracted light compared to the not-diffracted light, the effect is not great. So as long as the edge is relatively small compared to the total area of the beam, the effect is relatively small too. When you decrease the diameter of the opening, the edge length (the circumference) decreases linearly, the area exponentially, creating an increasingy worse edge to area ratio, i.e increasing the effect of diffraction. Smaller holes, more diffraction, less resolving power.<br><br>All lens elements were coated in all coated lenses already. Makes no sense to do it any other way. So no, there was nothing special done to coatings.<br><br>'Telecentricity' used to be called retrofocus, having the exit pupil further away from the film. It does indeed produce a more evenly illuminated image, but at the cost of increased distortion and other aberrations, not increasing but indeed decreasing resolution/sharpness. Nothing new, though indeed a plus when using digital sensors back then.<br><br>Yes, it was widely claimed that lenses made for digital were better (even though they very often were the very same lenses as were available before, just rebadged "digi[something]". Doesn't matter much though, since the lenses were not the limiting factor. The sensors and that dumb soft focus filter were. Things changed and the D800e is quite good. But it still is just about on par with good old film (though not quite as good still as some films were).<br>In the days before sensors became the limiting factor, you could indeed tell that lenses weren't all good (even expensive ones) and that they behaved different at different apertures (though there were a few that were at their very best wide open already). What i find a quite bemusing is to see how people are starting to discover this again now that sensors have caught up, as if it was a new phenomenon introduced by the current crop of digital cameras.<br><br>Using a good scanner, you will indeed be scanning lumps of grain (the bigger ones first, the smaller ones only when you scan really well. Grains aren't pixels, but come in a variety of sizes, and there is no such hard limit to resolving power as sensor wells present). And when you do, you discover that film captures detail at that level. So it's not what it might appear to suggest "scanning grain": it really is not that futile an exercise. There is still image detail at the grain level.<br>More convenient than using film digital certainly is.<br>Perhaps interesting re what you say about 'noise [i.e. grain] reduction' is <a href="http://www.photo.net/casual-conversations-forum/00d7Pm">the thread about the look of film</a> and products like DxO's FilmPack.<br><br>Re 4x5": as a user of cameras in that format i can't tell you how much i long for an affordable 4x5" digiback. it will never come, but it would be a very welcome thing! ;-)<br>Meanwhile, for anything from 50 MP and up, MF digibacks are also far too expensive, but the thing to use.

<p>Ah, thanks for the explanation, QG. That's certainly not the interpretation of diffraction that I was taught in physics class, but I see where you're coming from.</p>

<p>On the lens coatings, I was referring particularly to Sigma's "DG" lenses, some of which were (I think) largely identical to their predecessors except for the rear element coating. At least, that's what they claimed - I can't speak for whether this was entirely a marketing difference.</p>

<p>I concur that higher resolution sensors are putting more demands on lenses than older ones. I used to shoot a lot more wide open on my D700 than I did on my D800e, because the D700 hid most of the softness that resulted. Good modern lenses are generally good lenses on film as well, and can be seen as such - I'm just doubtful that there was a large stash of "good old lenses" that we've somehow all forgotten about. In most (not all) cases, lenses I've heard of that had a good reputation from the film days are turning out to be disappointingly soft.</p>

<p>I don't dispute that there's detail into the grain on film - just that the graininess of even something like Velvia is distracting to someone used to digital, at least for low ISO, and the technical ability to resolve detail was less important to me than the look of the result. I also accept that some films can resolve extremely fine detail, although usually at very low contrast compared with a digital sensor with a more direct drop-off in resolution. And yes, the best black and white films have very little grain (I have some in my fridge), and you can make a good argument about the resolution merits of not having a Bayer sensor. But - on a format-by-format basis - I'd generally expect to get more detail, more easily, from a modern digital sensor. Apart from anything else, I've not met many 35mm film cameras with AF fine tuning (though I realise that Leica sort of have it).</p>

<p>Having said all that, I can't claim I've ever shot with a 50MP sensor, and I'm not disputing the merits of digital medium format backs (even if most of the sensors are much smaller than 645). So I'll take your word for it that there's an advantage in switching format. I'll be interested to see whether the results of testing a high-end lens (by which I mean something like the Zeiss 135mm) on a 5DsR hold up well - the jump over 36MP is small enough that I'd actually be surprised if there's an issue, but I'm here to learn. Of course, it'll have to wait until June when the cameras are actually released - although since Nikon barely announced anything (of interest to this forum) at the show, a delay in shipping isn't the worst crime.</p>

<blockquote>

<p>Having said all that, I can't claim I've ever shot with a 50MP sensor, and I'm not disputing the merits of digital medium format backs (even if most of the sensors are much smaller than 645). So I'll take your word for it that there's an advantage in switching format. I'll be interested to see whether the results of testing a high-end lens (by which I mean something like the Zeiss 135mm) on a 5DsR hold up well - the jump over 36MP is small enough that I'd actually be surprised if there's an issue, but I'm here to learn. Of course, it'll have to wait until June when the cameras are actually released - although since Nikon barely announced anything (of interest to this forum) at the show, a delay in shipping isn't the worst crime.</p>

</blockquote>

<p>On the contrary, you could have determined the effects of this pixel density for yourself at any time in the past couple of years using any of several 24MP DX bodies. 50MP FX does not represent a new pixel density.</p>

<p>Nick: Only for the frame centre. To an extent, the trade-off between depth of field and pixel count is relevant. If pixel density was the only factor, my V1 would be a better indicator of lens quality than my D810. (Okay, sometimes it is, but it's not a complete test without some kind of shift adaptor added.) But you make the quite valid argument that full frame pixel densities aren't necessarily all that scary compared with, say, the V3.</p>

<blockquote>

<p>aren't necessarily all that scary compared with, say, the V3.</p>

</blockquote>

<p>I'm occasionally looking for a cheap J4*. It seems to manage 20fps @ 18 Mpix OK....or 60fps with AF locked. The Expeed 4A has promise! Trouble is, you've got to get that <strong>1</strong> second just right!</p>

<p>*They seem to be about £200 now!</p>

<p>as resolving capacity is subject and contrast dependend, this discussion, mostly, is pointless.<br /> i would recomment reading about spatial and angular resolution as a starting point and go from there.</p>

<p>cheers</p>

Is it, Norbert? Contrast and transfer, diffraction, resolving power are all players in the same game. The way an optical system is limited by 'hard' thingies (such as aperture diameter) sets the maximum achievable resolution, no matter what the subject is. The subject then can limit it further.

<p>Wait til Nikon or Canon start installing micro T Grain shape pixels that randomly regenerate themselves across the sensors for each exposure, with 18 stop dynamic range - totally mimicking film. All current digital SLR cameras will be found at Goodwill for $50 each.</p>