Dynamic Range

The Battle Of Low Light Vs High Resolution

Guest Post By: Cemal Sagnak

Nikon Df - ISO 12000

With the Launch of the Sony A7s the company is offering 3 almost identical  Cameras with a different Sensor design for matching Customer needs. This is unique and brings the Consumer into the conflict selecting from 3 Full Frame sensors to match his personal need. You can select between the resolution King A7r with 36mp and the moderate and affordable A7 with 24mp or the new Low Light Monster A7s.

Nikon Df - ISO 1000

It’s a clear trend driven by the technology leaders Nikon and Sony to launch Cameras which became “night vision devices”. Maybe it’s a Trend requested by the Users but maybe, “because they can”.

After a strong trend towards increasing megapixel the trend goes to achieve higher image quality by downsizing to enhance the low-light capability on the cost of resolution.

Nikon Df - ISO 12000

Low Light Monster Hitting The Mainstream

We had those cameras for years in the professional segment like the Nikon D4 with 16MP. The latest Innovation by Nikon used  a slightly modified sensor and more affordable Retro Body called Nikon Df. Many expected it with a 24 MP sensor but came out with “only” 16 MP , a resolution which is more common in APSC than in Full Frame Cameras.For a while it was the King of the night until the new “Lord of the darkness”, the Sony A7s with just 12 MP was launched. I followed the initial reaction by claiming the 12 MP is not state of the Technology today, but what is the state of the Technology ,or better say the state of the Art and how many pixel does someone really need.

Nikon Df - ISO 12000

Leica M9 - ISO 800

Size Matters

Decreasing the amount of pixels, allows manufacturer to build larger photosites which are better capable of capturing light.

Let’s do some maths and just a little physics. When we compare size and megapixel , an APS-C with 16 MP Camera would have similar sized of Pixel as compared to a FF Sensor with 36 megapixel. Now taking this to the next level would mean you are more than doubling the pixel size if you have 16 MP on a FF camera , and even bigger with the Sony A7s.

Pixelsize comparison :

Nikon D4  -  7.3µ

D800  -  4.9µ

Nikon D7000  -  4.8µ

Sony RX100  -  2.4µ

Sony A7s  -  8.4µ

Leica M9 - ISO 800

When comparing image sensors, either CCD or CMOS, the system is essentially a box where the input is light and the output is an image based on the light that is seen. The service provided by the sensor is the conversion of light to a digital image measuring light energy. And here hits physical limitations the abilities. Increasing the count of the pixel does not increase the measureable light. In other words: The larger the size of a pixel the better the ability catching (available) light (and the dynamic range).

Sony´s engineers confirms this physical fact, when they were asked why to put a 20.2 MP sensor into the relatively small 1” sensor of RX100 :

“It’s true that increasing pixel count increases noise. But since we manufacture our own sensors, we can easily tweak sensor specs to suit specific needs…..”

Knowing the correlation between Pixel count and Sensor size I reviewed my archives to find high Iso images when having the Nikon Df for a period of time.

There were some occasions like in the church or shooting night shots, but statistically I shot maybe 20 pictures out of 3000 with ISO higher than 3200.

I have to admit shooting with ISO 12800 is a nice feature to have but its like taking a sledgehammer to crack a nut. Some might need it but I am a streetshooter and I am okay with max ISO 1600.

Sony A7 - ISO 3200

Sony A7 - ISO 3200

The Leica M240 has a better lowlight capability than its predecessor the Leica M9 and in fact I was much more relaxed using Auto ISO on the M240 than with other Leica cameras, For me it has a perfect balance between high resolution and low Light capability. Having said that, my Preference Camera is and will stay the Leica M9. Its look and apeal is the closest thing to analog Film, in fact when Leica designed the M9 in cooperation with Kodak, analog Film stand model for the KODAK KAF-18500 Image Sensor.

Leica M240 - ISO 3200

Leica M240 - ISO 3200

Printing with Low Pixel Counts

Most Print Services require a quality of 120 dpi to get good to very good printing quality.

Here an example :

A uncropped file from a Leica M8 with 10 MP (3916 × 2634) can be printed :

  • in a good quality with (111 dpi) on 70×90 cm
  • very good quality with 166 dpi on 50×60 cm

In this case the pixel count makes a difference but looking into the size of prints you can get with a 16.2 MP file is maybe larger you might need. Don’t get confused by your home printer with +1000 dpi, this number represents overprinting a single “dot” with different colors to create the color that is needed.

Fuji X-T1 - ISO 3200

Dynamic Range

One additional advantage, and maybe for me the biggest advantage of modern Low-Light Monsters is the Dynamic Range.Simply said its the ratio between lightest and darkest regions (contrast Ratio). The ranges increases with the size of photosites of each Pixel. Actually the human eye has a very wide DR as it can easily adapt to different light situations. the DR of a Human eye is seen in the range of 24 f. The Sony A7s shines also here where the dynamic range is measured similar to the high END Video Cameras , one which is used in professional and cinematic Genres. In this area a high DR is critical capturing night and dark Scenes.As long you are not planning to become the next Steven Spielberg this topic is less important for you.

Nikon Df - ISO 1000

CONCLUSION

Here is my conclusion doing my math and research, I personally don’t need a “Lord of the darkness” as I really prefer some grainy look on available light photography but I can also live very well with lower pixel count in a FF sensor and would only decide for the A7s because of its dynamic range. The perfectly balanced Camera for me is the Leica M240. But thats just my personal choice. The A7r is demanding regarding lens selection and its size of the files. The Sony A7 has the same MP vs Sensor size ratio as the M240 still performs less good on high ISO. But A7 has a great image quality and the best performance / price ratio as its the only FF camera hitting the 1000 Euro line. As I don’t shoot DSLR anymore the Nikon Df is not an option for me but delivers one the best IQ and dynamic range.

Thanks for reading.

Cemal Sagnak - http://cemalsagnak.wordpress.com/

D4S TESTING - DYNAMIC RANGE COMPARISON

Is The D4S The Worst Performer?

I have been shooting with the D4S for the past month and have had mixed opinions on it. Since I currently own the Nikon DF and D800E, I have been making comparisons between the 3 very different camera bodies. Being that the D4S is the flagship camera, and most expensive option currently offered by Nikon, my expectations were pretty high. I love my Df, and with the D4S having the same sensor just tweaked a little differently via software, I expected image quality between the two to be similar with a slight advantage going to the D4S.

My initial thoughts on the D4S were that I loved it. I loved the speed it offered. I loved the solid feel, button layout, and…well, that’s about it. The extremely high ISO is nice, but I only care about the ISO ranges that I actually use, and how well it does when in that range. I had initially planned on purchasing the D4S once my month of testing this particular loaner was up, but I have changed my mind. Not only have I changed my mind, but with my last in-studio dynamic range testing, I’m actually pretty disappointed in Nikon’s flagship $6500 camera. I haven’t had the opportunity to shoot the new D810, but now seeing how well the D800E stacked up against the D4S in dynamic range and ISO performance I will be giving it a try.

I usually don’t do any “brick wall” testing when I do reviews, I try to only use real life examples. Because of a few shots that I had under exposed with the D4S, I had some concerns with banding when bringing up shadows. In real life, when exposure is on point, dynamic range shouldn’t ever really have to be pushed this much. But, when reviewing a camera, especially one like the flagship D4S, I like to push them to their limits to see how they perform. I brought the D4S, Df, D800E, and even the 5D MarkIII into the studio to test the dynamic range and ISO. I shot all of the cameras at ISO 100, 1/100sec, 0EV, and f/1.8. The D4S and D800E with the 85mm f/1.8G and the Df with the 58mm f/1.4G. I shot the 5D with the same settings, just with the 50mm f/1.2L. All were shot at f/1.8. The main test was between the D4S and the D800E, I threw the Df in once I seen how much the D800E outperformed the D4S.

My main concern with the D4S is the banding that I noticed over the past month of shooting with it, I noticed it several times when bringing up shadows. Banding has never been an issue with my D800 or D800E, I had only ever noticed it with my Canon bodies. I never noticed it with the Df until this test was done, and it’s very very slight, no where near as noticeable as with the D4S.

Here’s the thing, and don’t get me wrong, the dynamic range is still very good with the D4S. I am just surprised to see that as Nikon’s flagship camera body, it doesn’t perform better. The reason I even did this test was because I had noticed the banding in the shadows and was curious about the dynamic range performance. I also wasn’t overly impressed with the ISO performance. I don’t care how high the ISO can go if it doesn’t do a better job at the ISO ranges that I use in real life. In my ISO testing, it performed the same as the Df and D800E until it reached extremely high ranges that I just never use in real life shooting.

As you can see in the test results below, the D800E handled bringing up shadows at ISO 100 the best. There is barely any noise at all, and no banding. The Df did a slightly better job than the D4S, but still not as good as the D800E. How the Df does a better job than the D4S at cleanly pulling detail out of shadows seems a little odd to me since they are the same sensor. One thing that can definitely be concluded is that all 3 of the Nikon cameras do an awesome job of shooting in the dark, and all 3 are impressive. I just expected the one that is 3 times the price to be a little better then it’s cheaper siblings.

The only thing done in post was raised the exposure +5, exactly the same on all of them. No other adjustments made. Take a look…

D4S + 85mm f/1.8G

D4S + 85mm f/1.8G

Df + 58mm f/1.4G 

Df + 58mm f/1.4G 

D800E + 85mm f/1.8G

D800E + 85mm f/1.8G

5D MarkIII + 50L

5D MarkIII + 50L

Speaking with other Nikon shooters, a few had mentioned seeing this at ISO 100, but thought that it went away as ISO went up. I noticed it at all different ISOs. Here is another example of the banding I noticed with the D4S at ISO 800. This is a shot that I accidentally fired off during a shoot which is extremely underexposed. You can see the banding along the top once exposure is brought up in Lightroom. You can also see the color in the banding in this example. This is ISO 800 - 85mm at f/1.8 - 1/8000 sec

ISO 6400 COMPARISON 

The shots at ISO below 6400 are identical, but as you can see, when shot at ISO 6400, the D4S has a slight edge over the D800E. The Df is identical to the D4S at 6400. Even though the D4S has a slight edge here at 6400, that is still a little disappointing, I expected to see more of a difference than what you can see below. In real life shooting, I didn't notice any real advantage of high ISO shooting with the D4S over both the Df and the D800E.

D4S @ ISO 6400

D800E @ ISO 6400

CONCLUSION

This is a test, and just a test. I usually stay away from showing these kind of test results because they show what happens when these sensors are pushed to their limits. In real life shooting, you should never really have to push it this far and if your exposure is on point, all of these cameras are excellent. With that being said, there are times when we need to push these cameras, maybe not to these extremes, but they are pushed. The D4S is Nikon’s flagship camera, and it is touted as the best for many reasons…its price tag reflects that. The reason I brought the D4S into the studio and ran this test was because I noticed the banding while shooting in the field. Would I have noticed it if I nailed my exposure, absolutely not. The D4S has excellent dynamic range, but when pushed, I didn’t expect to see it out performed by its less expensive siblings. If you don’t need the speed of D4S, you are mainly concerned with its image quality. As a flagship camera, it should be the best, and for the price…it better be.

I had planned on purchasing the D4S, yet there were many reasons I decided against it. I don’t need the speed and it failed to impress me with image quality for its price when compared to the other cameras that I already own. As far as the banding goes, and how much importance I actually put on pushing a camera to is extremes like this, I can tell you this. Dynamic range is important to me, and its one of the reasons I shoot Nikon. With that being said, the 5D MarkIII performed the worst out of all the cameras, yet it is one of my favorite cameras alongside of the Df. The importance of these results are up to you, but if you are spending $6500 on a camera, you might expect to see a little better.