Print comparisons from 12 vs. 21+MP

jamie_robertson2 · November 17, 2009

Just for completetion, here's a 100% crop. Granted, it's not exactly a high-res image but it's still nice.<div>

00V267-191637684.jpg.ed86f5f6e0664e7446c249440e000191.jpg

</div>

oskar_ojala · November 17, 2009

If you want to use high ISO then the benefits of a larger pixel count are dimnished.

If you don't use a tripod then the benefit of increasing pixel count is not as large as when using a tripod.

13x9" is a small to medium size print, it can be handled by 12 mpix without problems. A lot depends on your printing materials and technique; the greatest benefits can only be achieved with very good printing. Now larger prints such as 16x20" are a different matter, there larger pixel counts start to show their benefits.

Why is Canon the only option? Sony has 24 mpix full frame for around $2000, I would say that's the best deal at the moment.

david carver · November 17, 2009

Oskar, I agree and I did list Sony earlier in the thread. The Zeiss lenses are the draw with Sony. I also handled a Sony at a store and was very impressed with the quality and feel of the camera. It felt very good in my hands.

Jamie, the crop is very impressive

jamie_robertson2 · November 17, 2009

"If you want to use high ISO then the benefits of a larger pixel count are dimnished. " Depends which cameras you compare. Compare a 21MP 5D2 to a 10MP Canon 1000D and the high ISO of the 5D2 is way better. Compare the 18MP 1.6x crop Canon 7D to the full frame sensor of the Canon 5D and the 7D has far better high ISO performance.

"Why is Canon the only option? Sony has 24 mpix full frame for around $2000, I would say that's the best deal at the moment." It's a good deal if you only shoot at ISO 100 or 200, anything above that doesn't compare with the D3x or 5D2. It's also a good deal if you don't want a wide range of lenses to choose from. It's also a good deal if you want a completely different flash hotshoe design that's incompatible with virtually every flash accessory available.

Sorry for being so sarcastic... no offence intended... just my weird humour ;0)

don_cooper · November 17, 2009

I think if you'd search about 5-6 years back and look for the same type discussions about the difference between 6-7mp and 12mp sensors you'd find pretty much the same feelings. Some saying the difference is negligible and not really worth it, and others feeling differently.

dan_south · November 17, 2009

Speaking of proper technique, I hear that Mirror Lockup is difficult to use on the 5D2. I guess you have to go through some menus to access it. Would anyone like to share their experiences with the MLU?

jamie_robertson2 · November 17, 2009

"

Speaking of proper technique, I hear that Mirror Lockup is difficult to use on the 5D2. I guess you have to go through some menus to access it. Would anyone like to share their experiences with the MLU?"

MLU has always been a pain on Canon cameras. However, there is a workaround with the 5D2. The camera has a customisable "My Menu" system so you can program the MLU function into that and then be able to activate it at the turn of a dial. More importantly though, many people now just use Live View when using a tripod. Obviously, when using Live View on any DSLR, the mirror is already flipped up and out of the way so MLU isn't an issue at all. I use Live View when on a tripod so that I can get perfect focus (using the 10x mag feature). It also enables me to get a more accurate framing of the shot compared to using the viewfinder, not to mention saving my poor bad back from bending down so far! Then when you fire the shutter there is no risk of mirror slap.

ShunCheung · November 17, 2009

I think if you'd search about 5-6 years back and look for the same type discussions about the difference between 6-7mp and 12mp sensors you'd find pretty much the same feelings. Some saying the difference is negligible and not really worth it, and others feeling differently.

</blockquote>

Don, the difference between 6 and 12MP is more noticable than the difference between 12 and 24MP. 12MP is more than enough for a lot of applications for perhaps 95% of us in this forum. And when you have even denser pixels, there are disadvantages as we have already discussed. Essentially you are facing diminishing returns.

Another issue is that with 12MP on DX and 24MP on FX, we are approaching the limit for a lot of lenses, especially the cheaper consumer zooms. That was no a factor a few years ago when you compared 6 vs. 12MP. Therefore, in many situations more pixels are somewhat meaningless or can potentially even be a disadvantage.

However, clearly, more pixels still seems to sell cameras, at least to some customers.

dan_south · November 17, 2009

Obviously, when using Live View on any DSLR, the mirror is already flipped up and out of the way so MLU isn't an issue at all.

</blockquote>

It's not "obvious," because that's not how Nikon cameras implement Live View. After having been raised for Live View, the mirror closes and opens AGAIN when you press the shutter release to make the exposure (really, really stupid design).

ilkka_nissila · November 17, 2009

After having been raised for Live View, the mirror closes and opens AGAIN when you press the shutter release to make the exposure

Right; so the way one would use it is to turn Live View on only for focusing and then switch to either self-timer with exposure delay mode or M-UP mode for taking the actual exposure. I read that this issue has been fixed in the D3s, though just from one source.

ilkka_nissila · November 17, 2009

I think if you'd search about 5-6 years back and look for the same type discussions about the difference between 6-7mp and 12mp sensors you'd find pretty much the same feelings. Some saying the difference is negligible and not really worth it, and others feeling differently.

Pretty much everyone was very impressed by the results when Nikon made their first 12 MP camera (D2X).

I'd love to have a 24 MP DSLR when the signal quality gets approximately in the same ball park with the 12 MP cameras (before calculating, average 2 pixels of the 24MP image together for every pixel from the 12MP model to make it even playing ground). From the looks of it the 12 MP FX cameras are increasing the lead.

jamie_robertson2 · November 17, 2009

"It's not "obvious," because that's not how Nikon cameras implement Live View. After having been raised for Live View, the mirror closes and opens AGAIN when you press the shutter release to make the exposure (really, really stupid design)."

Sorry Don, I didnt realise that was the way the Nikon system worked. I just assumed the Nikon system worked the same as Canon's.

"Another issue is that with 12MP on DX and 24MP on FX, we are approaching the limit for a lot of lenses, especially the cheaper consumer zooms." That's not actually the case. Of all DSLR cameras currently in production, the Canon 7D is theoretically the most demanding on lenses because it's pixel density is the greatest at 5.4MP/cm3. In contrast, the Canon 5D Mk2 has a density of just 2.4MP/cm3 = less than HALF that of the 7D. The performance of the 7D sensor is excellent and does not go beyond the limit of current lenses. Going by that data you can naturally work out that a full frame sensor with the same pixel density of the 7D would equate to 46 megapixels and would put no more demand on lenses than the 7D already does. In fact, I wouldn't mind betting that the upcoming Canon 1Ds4 will have a megapixel count of around 40MP.

"I'd love to have a 24 MP DSLR when the signal quality gets approximately in the same ball park with the 12 MP cameras " The 21MP 5D2 is very close to the performance of the 12MP Nikon D3 and D700 already when it comes to ISO performance/signal quality. The new D3s is, however, yet another leap forward in this game and leaves the 5D2 behind in terms of high ISO quality. The new Canon 1D4 seems to be relatively similar to the D3s from initial previews but we will have to wait and see to be sure.

stephen_asprey2 · November 17, 2009

Its an invalid comparison. You can buy a D90 for a third the price of a 5D2. Buy a D700 and compare that. Even then its not valid. Just because canon decided to squeeze lots of little noisy photocells onto the 5D2 sensor to make up 21mp and claim superiority, that does not make the image better than the 12mp of fat, low noise cells of the D700. Even the older 5d may still be superior. The end result would be the same to the eye. Even the 12MP sensor of the D300/D90 would still be comparable after post.

But to answer your question...up to 13x9, there will be little difference...certainly not worth the price difference. And with the D300/D700 you get the D3's superior AF and colour matrix metering. And we won't start again on the 5D2's cheap plastic body, will we folks? Especially if you go out in the rain or dust.

dan_south · November 17, 2009

The 21MP 5D2 is very close to the performance of the 12MP Nikon D3 and D700 already when it comes to ISO performance/signal quality. The new D3s is, however, yet another leap forward in this game and leaves the 5D2 behind in terms of high ISO quality. The new Canon 1D4 seems to be relatively similar to the D3s from initial previews but we will have to wait and see to be sure.

</blockquote>

The technology that we have available today is astounding!

sebastianmoran · November 17, 2009

I haven't compared the camera systems, but I did do a little test to see if lots more pixels would make a difference in print image quality.

I wanted to know if more pixels than my D200 would make a difference visible to the naked eye in an 8x12 print.

An 8x12 print is 96 square inches. The very interesting gigapxl.org project says roughly "beyond eight square inches per megapixel, the human eye will notice a loss of sharpness." We're talking about close inspection by naked eye, not looking from the 'normal' viewing distance. By this rule of thumb, my 8x12 would be just beyond the largest optimally sharp print from a 10MPx D200.

To research this, I shot a demanding test scene with my D200 and Nikkor 35mm f/2 prime lens, using a tripod. It makes an excellent 8x12 print with pretty standard processing, and with a small size reduction to 300ppi for the print (2400x3600 pixels). Sharpening makes a difference of course; I follow Bruce Fraser's approach.

Then, for comparison, I mounted a Nikkor 105mm f/2.5 prime on the same camera, used a panorama head and shot 16 images to stitch together for a 54MPx image. Pretty standard processing, then reduce it in Photoshop to the same 2400 x 3600 pixels for a comparison print, same size, 8x12. Same sharpening for print.

The result:

<ol>

<li> Both prints are excellent. I judge both as critically sharp. Other viewers do so as well.</li>

<li> The print from 54MPx is better to the naked eye. This was a surprise to me. Details are sharper, and surface textures are better. I think it gives a richer visual experience; when you can move up close and the detail holds, I think it changes the viewer's connection to the image. At least, it does so for me. </li>

<li>You have to look closely to see this; the prints at arms length are not visibly different, and differences are barely visible at the 'standard' viewing distance of 20" for an 8x12 print (1.5x the diagonal measure). On the other hand, it doesn't take pixel-peeping or a magnifying glass to see the difference. </li>

<li>You do need good viewing conditions to see the difference in the prints.</li>

<li>Just a little motion blur or poor focus would wipe out the advantage of more pixels. I suspect you need a tripod and excellent glass for more pixels to make a difference at this print size.</li>

<li>I conclude that 10MPx is sufficient to make a fine 8x12 print, and that more pixels help make a really excellent 8x12. I think the rule of thumb "8 square inches max per megapixel" is pretty good for critically sharp prints. (Of course, successful images are made with far fewer pixels. I'm just talking about sharp prints that convey detail and texture.) </li>

<li>I'm not about to buy a 24MPx SLR or a MF digital back, but when I'm serious about the image and the scene permits, I'll make a stitched composite. This has worked very well for me in making excellent large prints up to 50" of buildings, for example.</li>

</ol>

If anyone is interested, I'll link to the files and you can draw your own conclusions.

I wish everyone good shooting!

ShunCheung · November 17, 2009

Then, for comparison, I mounted a Nikkor 105mm f/2.5 prime on the same camera, used a panorama head and shot 16 images to stitch together for a 54MPx image.

</blockquote>

Richard, I am afraid that your experiment represents a totally different situation than cramping a lot more pixels onto the same (FX or DX) sensor area with a single capture with one lens.

When you stitch 16 images together, all of a sudden your sensor area is effectively 16 times as big. Ok, there are probably plenty of overlaps, so say your sensor area is 10 times as big. Your pixels are not proportionally smaller because they are not cramped into an FX or DX area. But more importantly, you don't have the limitation that one lens has to resolve for so many little pixels in just one capture.

So as interesting as it may be, the result is for a totally different situation that cannot at all be compared to using a 20+MP camera in non-stitching situations.

steve_wagner1 · November 17, 2009

Shun said

Unless your technique and optics are superb, going to 20+MP is largely meaningless. For example, if one shoots landscape, always uses a tripod and stops down a high-end lens, 20+MP makes sense if you need to make large prints. Otherwise, you'll frequently suffer from the lower-quality photosites due to high pixel density and slower frame rate simply because the camera has so many pixels to process.

</blockquote>

I find this to be very dubious, like saying unless you do everything perfectly medium format film is meaningless compared to 35mm film. It's just not true. And the majority of serious photographers know what they're doing and do most everything right, and are using high quality optics.

Having almost twice as much image info make a big difference in quality with digital just as it has always done with film. Go to the bookstore and you can pick out the nature photo books shot with a 20+mp digital slr compared to a previous generation right away. Even at the smaller print size the difference is quite pronounced, just as was the case with 35mm vs. mf at smaller print sizes.

steve_wagner1 · November 18, 2009

Dan asked

Speaking of proper technique, I hear that Mirror Lockup is difficult to use on the 5D2. I guess you have to go through some menus to access it. Would anyone like to share their experiences with the MLU?

</blockquote>

Same as it's always been, same place, not difficult

steve_wagner1 · November 18, 2009

Stephen said

Just because canon decided to squeeze lots of little noisy photocells onto the 5D2 sensor to make up 21mp and claim superiority

</blockquote>

Boy, that's a pretty grumpy and cynical perspective. Have you shot with the camera? Noise and overall IQ are absolutely phenomenal.

jamie_robertson2 · November 18, 2009

Stephen said: "Its an invalid comparison. You can buy a D90 for a third the price of a 5D2. Buy a D700 and compare that." I was simply illustrating that the number of pixels and the size of the pixels does not necessarily relate directly to noise. In that respect I think it's a perfectly valid comparison.

"Even then its not valid. Just because canon decided to squeeze lots of little noisy photocells onto the 5D2 sensor to make up 21mp and claim superiority, that does not make the image better than the 12mp of fat, low noise cells of the D700. Even the older 5d may still be superior."

That is absolute utter nonsense. I agree that the D700 and D3 have the edge over the 5D2 in terms of noise but it's not exactly chalk and cheese. There is actually very little difference. In actual fact, due to the much higher resolution of the 5D2, it actually makes the noise much less obvious at large print sizes compared to shots from the D700/D3. You've obviously never used the camera, and as for your claim about the original 5D being superior simply because it has fewer pixels... that's just silliness. It's like saying the 2.7 megapixel Nikon D1 vastly outperforms the Nikon D300s at high ISO because it has less pixels.

"And with the D300/D700 you get the D3's superior AF and colour matrix metering. And we won't start again on the 5D2's cheap plastic body, will we folks? Especially if you go out in the rain or dust." Sounds like someone's getting really grumpy now. Yes, the Nikon AF is better, no doubt. Cheap plastic body? As far as I'm aware the 5D2 is made from magnesium alloy. And as for being cheap! £2000 ain't cheap my friend.

I think you need to take a chill pill. I love the D700 and the D3, awesome cameras, but you don't hear me slating them just because I use a 5D2.

oskar_ojala · November 18, 2009

Depends which cameras you compare. Compare a 21MP 5D2 to a 10MP Canon 1000D and the high ISO of the 5D2 is way better.

</blockquote>

We aren't comparing any random cameras here. Get a D700 for high-ISO work and 20+ mpix camera for low ISO work. As for the D300/D90, I don't think there's much to be won by increasing the pixel count when light gets dim just by increasing the amount of pixels. My high-ISO images tend to be resolution limited by either shake or lenses, not the pixel count.

It's a good deal if you only shoot at ISO 100 or 200, anything above that doesn't compare with the D3x or 5D2. It's also a good deal if you don't want a wide range of lenses to choose from. It's also a good deal if you want a completely different flash hotshoe design that's incompatible with virtually every flash accessory available.

</blockquote>

400 ISO and below is the best range for these high resolution bodies anyway whether it's the S, C or N brand. I think Nikon lenses could be adapted to Sony, but I'm not sure. That would be a viable option for tripod work. And there are plenty of Minolta lenses around. Why would you use a hotshoe flash without TTL anyway? In any case, I'm not saying that the Sony full frame is a solution for everything, just a viable camera to consider. The handling of the camera seems to be very good.

jamie_robertson2 · November 18, 2009

"400 ISO and below is the best range for these high resolution bodies anyway whether it's the S, C or N brand."

Surely 400 ISO and below is the best range for ANY DSLR is it not?

Oskar, I am not knocking the Sony, I was being a bit sarcastic... just my little bit of fun.

As far as I am concerned, the only high resolution body (20+MP) that can genuinely excel at both low ISO work and high ISO work is the 5D2. At present, nothing else compares if you want the best of both worlds. The D3x is awesome but at high ISO it loses out rapidly, not to mention the price of the thing. The D3/D3s are amazing cameras at high ISO but if you actually need 20+MP they're no good. The 5D2 sells so well because, as far as image quality goes, it's got the best of both worlds. It's price, in comparison, is also very attractive. Unfortunately it loses out in the AF and build quality department but I don't suppose you can have everything. The best value all round camera if you don't need 20+MP is still the D700 but in my position I couldn't afford to switch brands.

ilkka_nissila · November 18, 2009

due to the much higher resolution of the 5D2, it actually makes the noise much less obvious at large print sizes compared to shots from the D700/D3

This statement is a bit ambiguous. According to dxomark.com tests, averaged per print area the D700 has about 1 stop edge in 18% gray SNR (a bit more below ISO 3200, less above it), 1 stop more DR at low ISO settings (100-400), fraction of a stop above that, tonal range about 1 stop advantage, but in color sensitivity, it has almost two stop advantage. These measures were all averaged per area, yet still the D700 remains ahead. The difference in richness of colour and tones is nontrivial. 5D Mk II images from e.g. weddings and club photography that I've seen support this; the images look "thin".

Also, high ISO is normally used to photograph moving subjects, since a tripod and low ISO would yield better results for still subjects). With moving subjects, good autofocus is ... well, useful. For some potentially high ISO action work one might actually prefer Nikon's D300s due to its coverage of AF sensor points all over the frame, unlike in other cameras e.g. Canon 7D. Like you said, you can't have everything, at least now, in an affordable camera, but you can get specialized cameras for different tasks. I would not mind having an ISO 100 camera for landscape and close-up work, but currently my brand doesn't offer one at a price point within my reach. Similar things could be said about action cameras; Nikon seems to have an excellent choice in their D700 and D300s at reasonable prices, both cameras including top-of-the-line autofocus. I think Nikon and Canon just ball us around, people "switching" is very profitable for them.

desmond_kidman · November 18, 2009

Look at Jamie Robinson's beautiful shot of the woman in the crowd....that's certainly the advantage of higher megapixel cameras.

Using the entire image you really won't see any difference in that print size.

If you start cropping to the extent you are losing 40% in any direction you most likely will see a difference.

sebastianmoran · November 18, 2009

Shun wrote:

your experiment represents a totally different situation than cramping a lot more pixels onto the same (FX or DX) sensor area

</blockquote>

Yes, of course. My experiment is just about whether additional pixels would matter, not about the implementation in today's Canon or Nikon offerings.

Besides the camera systems, the OP asked about more pixels in 13 x 9 prints; my experiment showed me that more pixels can make a difference, for fine work, at this print size.

Sign In

Print comparisons from 12 vs. 21+MP

Recommended Posts

Link to comment

Share on other sites

Top Posters In This Topic

Popular Days

Top Posters In This Topic

Popular Days

Posted Images

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Create an account or sign in to comment

Create an account

Sign in