Lotusm50 wrote:
Absolutely. This "normal viewing distance" and other equivalent "concepts" about image quality and size are just rubbish. I've been annoyed by this nonsense for a long time.
I've been annoyed by the absolute obsession people have over extremely minute things. Do some people go up to a print? Sure...but the fact is that 16 MP still looks really good up close even on a 30" print. Can you see every tiny minute detail at that size up close? No, but you can see a LOT of detail, and it looks very good.
People have gotten a bit crazy. "Oh my gosh, there's some GRAIN at ISO 1600? What are we to do?" Nevermind that the shot still looks good. People have become so obsessed with these tiny insignificant differences in image quality that I frankly think they are failing to appreciate photography for what it is...expressing a vision of the photographer.
Show me a single instance where a 5D II shot printed at 20" wide looks poor up close, if the photographer did what they were supposed to do. It's gotten crazy.
Jman13 wrote:
I've been annoyed by the absolute obsession people have over extremely minute things. Do some people go up to a print? Sure...but the fact is that 16 MP still looks really good up close even on a 30" print. Can you see every tiny minute detail at that size up close? No, but you can see a LOT of detail, and it looks very good.
People have gotten a bit crazy. "Oh my gosh, there's some GRAIN at ISO 1600? What are we to do?" Nevermind that the shot still looks good. People have become so obsessed with these tiny insignificant differences in image quality that I frankly think they are failing to appreciate photography for what it is...expressing a vision of the photographer.
Show me a single instance where a 5D II shot printed at 20" wide looks poor up close, if the photographer did what they were supposed to do. It's gotten crazy. ...Show more →I disagree. People have become very relaxed with print quality. There are large format photographers who will not take a 4x5" negative past 16x20. There are others who will not enlarge at all. And then you have people who are ok with 5mp images printed to 40x50". There is no right or wrong. But I disagree that people have become crazy, at least in regards to final print quality. Grain / noise I agree with.
Zaitz wrote:
I disagree. People have become very relaxed with print quality. There are large format photographers who will not take a 4x5" negative past 16x20. There are others who will not enlarge at all. And then you have people who are ok with 5mp images printed to 40x50". There is no right or wrong. But I disagree that people have become crazy, at least in regards to final print quality. Grain / noise I agree with.
Anyone working on individual images will find the bigger file size is of little consequence, I believe, especially if you use layers in PS. I spend much more time honing the image than fretting over already adequate response times, and new PCs are much faster than just a couple years ago. People have been working successfully with huge files (300-600Mb) for many years on inferior computers, from high re scans from MF and LF film frames. PCs will not get slower...
Agree re the 'normal viewing' meme, it's so last century, an excuse for inadequate imaging. You don't just want print resolution to equate with the best normal human vision but to surpass it somewhat as the difference is perceptible. From memory, doubling the pixel count on the same format equates merely to a 40% increase in res, so for A850/900/D3X/5DII users this is no big deal. The 24Mp sensor in the APS-C Sonys is well over 50Mp on 35mm and people are not throwing out their lenses, old or new...quelle surprise.
Nikon, like Sony in the coming year, have done us all a giant favour, as there are numerous reasons to want many more Mps. Just straightening some (A900) images I end up with 5DII file dimensions, and anyone who sees the world exclusively in 3:2 aspect is seriously missing out on some great imaging opportunities. Can't imagine the medium format makers are too happy with the direction of this wind.
Jman13 wrote:
I've been annoyed by the absolute obsession people have over extremely minute things. Do some people go up to a print? Sure...but the fact is that 16 MP still looks really good up close even on a 30" print. Can you see every tiny minute detail at that size up close? No, but you can see a LOT of detail, and it looks very good.
People have gotten a bit crazy. "Oh my gosh, there's some GRAIN at ISO 1600? What are we to do?" Nevermind that the shot still looks good. People have become so obsessed with these tiny insignificant differences in image quality that I frankly think they are failing to appreciate photography for what it is...expressing a vision of the photographer.
Show me a single instance where a 5D II shot printed at 20" wide looks poor up close, if the photographer did what they were supposed to do. It's gotten crazy. ...Show more →
+1
People imagine things on their own. It's imaginary tho! ：）
Just read above. "Prints that are above the range of human vision look better" (paraphrased). Hhehehehee....
Look, I'm not saying that no one has a need for resolution of this sort. There are plenty of photographers who need (and more who want for some reason), insane levels of resolution. Large format photographers of course are not going to feel that a DSLR generally gives them what they want, but I view it this way.
Plenty of tests have been done that show that a 16 MP DSLR image outresolves medium format film, even at low ISO. So we're talking 4x5 as being the benchmark above. I'm just curious as to when medium format film images became 'not enough resolution' for a 13x19" print.
And Normal viewing distance comes into play for most large prints, but the fact is that 16MP images look good even up close on a 30" print. I was looking at a few of my prints from my 1Ds II today that I've got printed at 24x36". If I put my nose to print, I can see areas where there is very minor loss of detail. At 12", though, it looks fantastic. That's 12" away on a 36" print...I was looking at roughly 1/6 of the image. Trust me...NO one other than other pixel peeping photographers will ever view an image at that level of detail, that close and say "wow, this stinks." Can it be better? Yes...but it's a very incremental step. For you, it might be worth that step, but frankly it's akin to audiophiles paying $4,000 for speaker wire for that extra 0.2 dB increase in S/N ratio, that comes mainly in the audio spectrum that is outside the threshold of human hearing.
Note that this does not apply if you are printing HUGE prints...where if you do go up close and want to see super fine detail you'd notice a big difference between 16MP and 36 MP.
Clever idea as to why the "negating" AA approach on the 800E. From Joseph Wisniewski (who usually knows what he's talking about) in a DPR comment:
I'm guessing Nikon is using a trick first seen on the Canon 5D II. The clear components in front of the sensor, because their index of refraction is higher than air, increase spherical aberration, and therefore decrease resolution, an annoyance that is most noticeable with high resolution sensors.
So, they eliminate one optical component by replacing the clear sensor cover glass with one of the two LiNbO3 filters. Because of that, you can't eliminate that filter without making up a new batch of sensors, which is expensive.
So, instead of eliminating it, you negate it. Cheaper than another sensor run.
Jman13 wrote:
Look, I'm not saying that no one has a need for resolution of this sort. There are plenty of photographers who need (and more who want for some reason), insane levels of resolution. Large format photographers of course are not going to feel that a DSLR generally gives them what they want, but I view it this way.
Plenty of tests have been done that show that a 16 MP DSLR image outresolves medium format film, even at low ISO. So we're talking 4x5 as being the benchmark above. I'm just curious as to when medium format film images became 'not enough resolution' for a 13x19" print.
And Normal viewing distance comes into play for most large prints, but the fact is that 16MP images look good even up close on a 30" print. I was looking at a few of my prints from my 1Ds II today that I've got printed at 24x36". If I put my nose to print, I can see areas where there is very minor loss of detail. At 12", though, it looks fantastic. That's 12" away on a 36" print...I was looking at roughly 1/6 of the image. Trust me...NO one other than other pixel peeping photographers will ever view an image at that level of detail, that close and say "wow, this stinks." Can it be better? Yes...but it's a very incremental step. For you, it might be worth that step, but frankly it's akin to audiophiles paying $4,000 for speaker wire for that extra 0.2 dB increase in S/N ratio, that comes mainly in the audio spectrum that is outside the threshold of human hearing.
Note that this does not apply if you are printing HUGE prints...where if you do go up close and want to see super fine detail you'd notice a big difference between 16MP and 36 MP. ...Show more →
Film has more resolution than you think: http://www.onlandscape.co.uk/2011/12/big-camera-comparison/
The D3x turns to mush. Large images printed from digital cameras, especially low MP cameras, have a plastic-wrap look. You may be happy with your 30" prints from a 1ds. But I am not close to satisfied with my 20x30 prints from a D300s.
About the same as I wrote earlier. I'm quite sure Joseph is right, and it's nice to know someone else is thinking the same thing - that's often a good sign. They've rotated the front filter layer by 90º and removed the phase plate.
One thing that I've never really thought about earlier, but was shown in practice in some Ricoh M-module test, is that depth of field actually gets more shallow - by a tiny bit - as a result of the increase in SA you get by stapling more filter depth in front of the sensor surface. The rear falloff happens just a little bit quicker, and DoF at really small apertures (like F11 or smaller) never gets as deep on a camera with AA filters as on a camera without AA filters...
So, for a landscaper or a macro-specialist, the "double-inversion" filters on the E model unfortunately doesn't give them the same result they would have gotten by removing the filters altogether.
(well, it doesn't actually matter that it's an "AA filter" material, what matters is that you introduce a thicker layer of high-refraction material in front of the microlenses)
The D3x turns to mush. Large images printed from digital cameras, especially low MP cameras, have a plastic-wrap look. You may be happy with your 30" prints from a 1ds. But I am not close to satisfied with my 20x30 prints from a D300s.
Yes, but that special part of the test is made on T-max, one of the highest resolving b/w films there is (that has an almost usable density latitude). Any color reversal or negative film will have a resolution that's less than half of what the test shows. Which you very clearly see in the other parts of the test, where they compare landscape targets - and then with color film of course. To get "better" than a D3x you have to move beyond the medium format cameras, up into large format. The D3x equals the Mam7 in most cases - at least it's not noticeably worse.
Yup not noticeably worse. But far from a 16mp camera besting MF. In the bill examples the D3x renders the letters to look like Arabic. There are also lower grain films and better lenses they could have used for MF.
1. Tariq's correct on transfer rates of modern HDD's, they're 60+MB/s these days, In fact the best consumer drives peak out at 140MB/s+ for sustained transfers. albeit those are high-end drives. Average retail drives are 60-80MB/s. There's a good reason why current HDD interfaces just hit 6GB/s with the latest SATA spec, they need that speed to avoid saturation from burst transfers.
2. With regards to processors today's processors are significantly faster, clock for clock, on a single core than anything equivalent from 5-6 years ago. The current kings, Intel's Sandy Bridge i5 and i7 processors have per-cycle performance higher than the old Pentium III and Pentium M processors. That makes them 2-3 times as fast as a Pentium 4 at the same clock speed on a single core and 1-2x faster than a Athlon of that era clock-for-clock. This is due to significant improvements in execution efficiency Intel developed after they figured out the hard way that the Netburst architecture in the Pentium 4 line was a dead end and in fact hobbled performance. Even the first Core processors were faster than the equivalent Pentium 4's despite running at nearly half the clock speed, and the current Core i5/i7's are faster at a given clock speed than an original Core processor.
And due to their ability to dynamically balance clock speed and power consumption, they actually ramp up the clock speed when you are using less cores. My i5-2500K, the current price/performance king, has a base clock of 3.4GHz, which is what it runs at when all 4 cores are running and it's out of power-saving mode. When only one core is in use it runs at 3.7GHz (2 and 3 cores in use clocks in at 3.6 and 3.5GHz respectively). And that's before you get to the issue that Intel currently underclocks their processors due to a lack of competition from AMD at the moment, the i5-2500k is clock-unlocked and will run solidly at 4.4GHz base clock on the stock retail cooler. INtel is in fact selling a 4.4GHz part as a 3.4GHz part.
Assuming that no "special" problems have been included in teh design, expect D7000 levels of DR, or a few tenths of an Ev less. Scaled to 9MP you get 15Ev.
Romain wrote:
Wait, what? You say the dynamic rang increases when downsampling?
The lower boundary of the DR is determined by a certain level of SNR, and since you lose some noise when downsampling, you're gaining some DR I suppose.
Yes, exactly so. You can also see it as "looking at detail on a larger scale", and then noise grain finer than the detail you're looking at becomes less important. It's the same thing, but expressed in a different way.
But in reality the 15Ev at 9MP that I stated is more than a bit exaggerated, since you never gain the "perfect" 1Ev per halving in the first halving of resolution. It may be technically true, but not in a real picture. The raw conversion process doesn't work properly at the pixel level.
alundeb wrote:
I guess we will have to agree to disagree. 4 MP images do not look good up close in a 15" print IMO.
I guess we will. I can't imagine what distance you're looking at these prints from where you would actually think that looks bad, but perhaps your eyes see things mine do not, or you are just innately more bothered by a lack of insane detail at all viewing distances. There's nothing wrong with that, but I guess I can't understand where you're really coming from, and we'll have to leave it at that. I did try this morning, though to see if I was just fooling myself.
I stared at a 12x18" print this morning from my old 6 MP digital rebel and it looks outstanding, even from 12" away (which is way closer than 'full viewing'), which is of the image below. Well, at least to me it looks outstanding.
In the print, up close, you can easily see the the cross beam boards of the tiny little gate (indicated by the yellow arrow). It is an almost invisible detail when viewing from a more normal distance of 2 feet..but up close you can see that detail. The gate on the print is 2.7mm wide (I measured the width in Photoshop), and I can discern when looking closely, the four horizontal planks and the diagonal cross brace. And the thing is, this photo was taken with a Sigma superzoom from a moving train handheld...something tells me I didn't get every single ounce of sharpness possible out of this system. I can't really see what reasons I would need more detail than that in a print of that size. But I guess that's just me. I'm sure it would look a little better up close if it had been taken with a 5D II and a 24-70L II, but is it even necessary, I guess is the point for me. For you, it might be.
theSuede wrote:
Yes, exactly so. You can also see it as "looking at detail on a larger scale", and then noise grain finer than the detail you're looking at becomes less important. It's the same thing, but expressed in a different way.
But in reality the 15Ev at 9MP that I stated is more than a bit exaggerated, since you never gain the "perfect" 1Ev per halving in the first halving of resolution. It may be technically true, but not in a real picture. The raw conversion process doesn't work properly at the pixel level.
Interesting, I never read that before that's going to make me like this camera even more...
that's going to make me like this camera even more...