E-Vener wrote:
Who said anything about "magical sensors"? The EXPEED processor in any Nikon camera is programmer to work with the specific sensor being used. It's a whole package.
The samples I have seen from the D7000 do not make me want to rush out and buy one. One should compare full resolution files on a good monitor or make relatively large prints to compare. Ho0wever, it does look like a really good intermediate type body....
Since the D7000 has a slightly higher resolution, it will respond to chroma NR a lot better (when you compare same-size presentations).
So not only does the D7000 have a lower "from-the-sensor" noise sum over the frame compared to the 12MP crop Nikons, it will also clean up better. This was mainly a way to show what the camera REALLY can do - it's a base-line. But if you gave the D5000 SOOC jpgs the benefit of the same in-camera NR as the D7000 SOOC jpgs, then the difference would be a lot smaller I guess.
But that the D7000 should come even close to the D700 in real-world noise per frame was nothing more than wishful thinking as I've said before. There just hasn't been enough evolution of sensors in the last three years to double the light-sensitivity and halve the electronic noise amount of large-scale sensors (larger than 8x5mm). That's what it would take to make the D7000 "equal" to the D700. And that particular "crop vs D700" situation won't arise for at least two-three years yet. Next years' new models won't achieve that goal either.
What most people claiming "equal to D700" noise-performance are seeing is mainly a better and smarter noise reduction - and those new NR algorithms can be applied to the D700 raw files as well....
Giving all three models the benefit of the SAME exposure and the SAME post-processing places the performance of the D7000 somewhere in the middle in between D90/5000/300 and the D3/700. And probably closer to the DX than to the FX models.
I think people are scrutinizing far too much and not shooting nearly enough. You can not compare the images from a D7000 to the D700 because they are two different formats. There are far too many variables that can not be equalized to give a proper comparison as to which produces better images (and there is far more to an image than noise).
In the end, the D7000 looks to take some very nice photos, even with high ISO's. The D700 is a great camera and produces fantastic results as well. Neither camera should disappoint so chose your format or your price point and buy away.
theSuede wrote:
Giving all three models the benefit of the SAME exposure and the SAME post-processing places the performance of the D7000 somewhere in the middle in between D90/5000/300 and the D3/700. And probably closer to the DX than to the FX models.
+1
From what I've seen so far, D7000's "magical sensor" is only slightly better than D5000. At ~33% more pixel density, if D7000 matches the performance of D5000, it's already a 33% sensor improvement.
But D7000 is 300% the pixel density of D700, asking D7000 to be anywhere close to D700 is day dreaming, not going to happen this year or next.
From the samples, I can see that D7000 is an improvement from the D90, IQ-wise. Body-wise, much more so.
If it can give consistently good results at iso 3200, I am okay with that. Don't really care if it is as good as D700 or not.
Because that's the point. It's an improvement on the noise-levels we had with the D90/300/5000, even if it has more resolution. That's all that matters. The improved ergonomics (guess that's personal...) and the better LV and video, the custom settings on the dial, the AIS metering capability and a new AF sensor - THOSE are the valid arguments for buying a D7000 if you're currently with a D90 or a D5000.
Coming from nothing (supposing the D7000 will be your first midrange camera, a step up from entry-level DSLRs or advanced compacts), I guess the D7000 is nothing more than a very strong contender in the mid-field crop camera ranges. The 60D (not for me...) and the Pentax K-5 look very strong too.
Tommy_D wrote:
I think people are scrutinizing far too much and not shooting nearly enough. You can not compare the images from a D7000 to the D700 because they are two different formats. There are far too many variables that can not be equalized to give a proper comparison as to which produces better images (and there is far more to an image than noise).
In the end, the D7000 looks to take some very nice photos, even with high ISO's. The D700 is a great camera and produces fantastic results as well. Neither camera should disappoint so chose your format or your price point and buy away....Show more →
Well said TommyD. If the camera meets an individuals' needs - go for it.
If I were still shooting my D50 and was looking to upgrade I would probably choose the D7000 over the D300. 3 years ago Nikon's lineup was not so expansive and your only option from the D50 was the D80 or the D300. I'm happy with my D300 and think it produces great photos, but at times I wish the body were a bit smaller.
I'm not seeing anything from the D7000 that makes me feel the need to "upgrade" from my D300 to it, but I can defiantly see why people are hyped about it. Its just like when the D300 came out. People went nuts. I would say its a good upgrade for the D40-D5000 crowd and a marginal upgrade for the D90 crowd. Not really an upgrade for the D300 crowd though.
I'm sure it's been said but gah high ISO tests in broad day light or with flash are pointless. They're just high ISO for the sake of being high ISO, not because they need to be that high.
If you want to shoot high ISO, shoot in a real world scenario where you need the ISO that high in order to freeze motion.
I can shoot ISO6400 at the sun with my dad's free gift camera and it'll look fine... (okay maybe not but you get the idea)
fusiongt wrote:
I'm sure it's been said but gah high ISO tests in broad day light or with flash are pointless. They're just high ISO for the sake of being high ISO, not because they need to be that high.
If you want to shoot high ISO, shoot in a real world scenario where you need the ISO that high in order to freeze motion.
I can shoot ISO6400 at the sun with my dad's free gift camera and it'll look fine... (okay maybe not but you get the idea)
The quantity of light doesn't affect High ISO performance; only the quality/color of light does. The sensor sees the same amount of light for a given exposure regardless of the EV.
Tommy_D wrote:
I think people are scrutinizing far too much and not shooting nearly enough....
Thats true. As someone who is currently doing most of his shooting with a Panasonic G1, I am sure any of these sensors will "blow it away"
However as sensors supposedly become better and better, one does get curious that how better are the sensors these days compared to past With that in mind I had recently done an experiment quite similar to what the theSuede has shown above. I downloaded the ISO1600 and ISO3200 raw files for this same scene for some "old" cameras as well as some of the more recent "high-iso kings". I used canon 30D, D90, k-x, NEX and T2i. I also threw in the 5D to see if the recent APS-C sensors are close to the old FF.
I converted them using lightroom 3.2 with sharpening and NR set to 0 (it is not as unprocessed as using dcraw, but since in real-world I use lightroom so this is closer to what I will expect in practical use).
Then I looked at these in two ways - first at 100% pixel level from full-sized images. Second at 100% pixel level after resizing everything to the 30D resolution (8MP).
The results that I got were interesting:
5D had lower noise as well as better detail then all APS-C's (as expected)
The 30D and D90 were the best of the APS-C cameras in terms of noise and detail (When looking at resized images, 30D actually look a little better but thats sort of unfair to D90 as it has been downsized. So overall I would say they were about the same)
The NEX had the largest detail loss at ISO1600 and ISO3200
T2i and k-x were somewhere in between these two in terms of noise and detail retention
In normal viewing conditions (full screen), the differences between ISO1600 samples were negligible. At ISO 3200 the differences were visible though still not excessive.
Obviously the performance could vary if we have a genuine low light scene with tungsten lighta etc. However it was still interesting to see that under these particular conditions the newest sensors seemed somewhat worse instead of being superior.
If you want to compare two cameras of differing resolution and pixel density, instead of comparing the physical area (h x w), you might wish to measure the area by pixel count.
svenjosh wrote:
Yes all the results from Thom Hogan to image resource are different from yours. I am not counting my results but they are too different from yours. Do some real life low light shots that really require high ISOs. Your results will correlate with others.
Svenjosh
Was just stating an observation.
I still stand by my opinion. I think the D7000 is not much worse than the D700 in terms of noise and keeps a lot of detail even at high ISO. It is a marked improvement of the previous crop sensors.
DavidWEGS wrote:
I find the two to be pretty close in real terms..... meaning, when I will use 6400iso, the light is low. With that in mind, I shot a couple (not too scientific) images of what would be normal for me in terms of low light. One is the D7k the other the D700. Each shows the full image followed buy a 100% crop.
Let me know what your thoughts are of the differences.
Here are the results:
Looks like the bottom photo is the D7000 since it's a larger crop. To be fair you'll need to shoot both at the same FOV.
Noise "effect" is scale-independant, so that doesn't bother me from a scientific PoV. What DOES bother me is the total lack of colour in the target. It's quite easy to NR a B/W target, and a lot harder to keep low-level colour detail. That's why the IR samples are so good, they're an intelligently chosen mix of high/low contrast targets, and also a good mix of strong/weak chroma differences.
The "easier" the target, the closer to the D700 the D7000 will perform. But compared on targets containing surfaces with fine / subtle colour differences and the difference will increase. So it can depend on what you generally shoot (and how important low-contrast detail is to you) how the "absolute" noise performance of the D7000 will affect your shots.
That's an enlightening excercise many more should try. It puts a quite effective muffler on the voice whispering -"must buy!!! must buy!!!" inside our heads... :-)
In many, many ways our software has evolved a lot more than our hardware for the last 5-8 years.
theSuede wrote:
Noise "effect" is scale-independant, so that doesn't bother me from a scientific PoV. What DOES bother me is the total lack of colour in the target. It's quite easy to NR a B/W target, and a lot harder to keep low-level colour detail. That's why the IR samples are so good, they're an intelligently chosen mix of high/low contrast targets, and also a good mix of strong/weak chroma differences.
The "easier" the target, the closer to the D700 the D7000 will perform. But compared on targets containing surfaces with fine / subtle colour differences and the difference will increase. So it can depend on what you generally shoot (and how important low-contrast detail is to you) how the "absolute" noise performance of the D7000 will affect your shots....Show more →
I was thinking same FOV for comparing both detail and noise instead of just noise. Agree with your points regarding color and tones.
That's an enlightening excercise many more should try. It puts a quite effective muffler on the voice whispering -"must buy!!! must buy!!!" inside our heads... :-)
In many, many ways our software has evolved a lot more than our hardware for the last 5-8 years.
My experience with 5D, D700, D300/D90/D5000 etc. concurs with curious80's observation.
D7000 is a very impressive camera, easily the best APS-C DSLR from Nikon and arguably the best APS-C DSLR from anyone. It's probably a better all around camera for many than D300 or D700, due to performance/features/size.
But the hype was mainly about the performance of state-of-the-art cropped sensors. D7000 is supposed to be able to match the performance of D700 with 300% pixel density, which is a revolutionary progress in sensor technology. People on the forums already started talking about the future of FX. It's pretty clear now, we are still not there yet.
5D is more than 5 years old. The IQ even at ISO 1600 remains unmatched by any APS-C cameras, that says a lot about the difficulty in trying to improve sensor performance within the same fundamental design, overcoming >2x the pixel density.
