My NX1 vs D7100 High ISO test

Started Mar 22, 2015 | Discussions
Horshack Veteran Member • Posts: 7,921
My NX1 vs D7100 High ISO test
7

Both using same tungsten lighting and absolute exposure. Both shot using Zeiss 35mm ZF.1 (NX using F->NX adapter). Shot raw, converted in LR 5.7.1 all defaults except WB. Luminance NR = 0, Chrominance NR = 5 (ISO 100-6400), Chrominance NR = 10 (ISO 12800-25600). All downsampled to 8MP to normalize for the resolution differences.

The ISO 100 -5EV and -7EV are a test of base ISO DR - the exposure was adjusted back to normal in LR using the exposure slider (-5EV) and exposure slider + local adjustment brush (-7EV).

I selected the D7100 because it has the best High ISO performance of any APS-C camera at the time the NX1 was released.

I can provide raws if anyone would like to work these in their own PP tool.

100% Crops of 8MP images

Full 8MP images:

NX1 ISO 100

D7100 ISO 100

NX1 ISO 100 -5EV

D7100 ISO 100 -5EV

NX1 ISO 100 -7EV

D7100 ISO 100 -7EV

NX1 ISO 3200

D7100 ISO 3200

NX1 ISO 6400

D7100 ISO 6400

NX1 ISO 12800

D7100 ISO 12800

NX1 ISO 25600

D7100 ISO 25600

Nikon D7100 Samsung NX1
If you believe there are incorrect tags, please send us this post using our feedback form.
Menneisyys Senior Member • Posts: 1,458
Re: My NX1 vs D7100 High ISO test

Fabolous test, thanks for publishing it!

The NX1 doesn't seem to be worse than the Nikon. Quite the contrary: the Nikon heavily bands even at -5EV, let alone -7EV.

Assuming the NX500 behaves in exactly the same way, it's a no-brainer.

tecnoworld
tecnoworld Veteran Member • Posts: 7,228
Re: My NX1 vs D7100 High ISO test
1

At insanely high iso the d7100 seems a bit better, but up to 3200 (or even 6400) I tend to prefer nx1. Anyway it's a close call.

It would be interesting to do the kind of test that dpr did with a7s, ie shooting underexposed and then taking up the ev, to evaluate the dr at any iso.

 tecnoworld's gear list:tecnoworld's gear list
Samsung TL500 Samsung NX200 Samsung NX100 Samsung NX300 Samsung NX1 +15 more
Melidonia Regular Member • Posts: 228
Re: My NX1 vs D7100 High ISO test

Thanks for this test, the NX1 has nothing to be ashamed of. 
And indeed, if the NX500 has the same ISO quality, it's a bargain.

 Melidonia's gear list:Melidonia's gear list
Samsung NX1 Panasonic GH5 Sony a7 III Sony FE 50mm F1.4 ZA Samsung 50-150mm F2.8 S +9 more
Raw Jaw
Raw Jaw Senior Member • Posts: 2,662
Re: My NX1 vs D7100 High ISO test

I really appreciate your very objective test.

Your being Objective should not preclude you from expressing your conclusions.

I would like to know your thoughts about the results of your comparative test.

Thank you.

captnj2 Forum Member • Posts: 86
Re: My NX1 vs D7100 High ISO test

We're the nx1 pics downsized to 24 mp? That's the most fair right?

marks123 Regular Member • Posts: 113
Re: My NX1 vs D7100 High ISO test
1

We're the nx1 pics downsized to 24 mp? That's the most fair right?

Looks like both were downsized to 8mp.

otto k Senior Member • Posts: 2,099
Re: My NX1 vs D7100 High ISO test
1

And this so shows how advanced Nikon image processing algorithms are, sensor is basically putting out banded garbage but ISO 12800 still looks quite acceptable. NX1 sensor seems even better but sensor signal cleaning (I'm trying not to say noise reduction, since this is "raw", but it's hard) of Nikon is clearly more advanced. It also shows where samsung could be in next model or even firmware (at least a stop better with same sensor).

Same is probably true for AF performance, given time to perfect the algorithms, NX could get very close to DSLR while keeping all of the advantages (seriously, tiny centered AF diamond in FF Nikon can be very annoying).

 otto k's gear list:otto k's gear list
Samsung NX1000 Canon EOS 600D Samsung NX500 Samsung NX 20-50mm F3.5-5.6 ED Samsung NX 30mm F2 Pancake +6 more
Kaziklu Senior Member • Posts: 1,301
Re: My NX1 vs D7100 High ISO test
1

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100. 
I think this is fairly impressive though that NX1 holds it own..  the noise isn't significant but slight.

 Kaziklu's gear list:Kaziklu's gear list
Samsung NX1 Samsung NX20 Samsung NX 30mm F2 Pancake Samsung NX 60mm F2.8 Macro ED OIS SSA Samsung NX 50-200mm F4-5.6 OIS +1 more
OP Horshack Veteran Member • Posts: 7,921
Re: My NX1 vs D7100 High ISO test

Kaziklu wrote:

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100.
I think this is fairly impressive though that NX1 holds it own.. the noise isn't significant but slight.

All images I posted were downsampled to 8MP, which normalizes the noise differences related to pixel count/density.

OP Horshack Veteran Member • Posts: 7,921
Re: My NX1 vs D7100 High ISO test

Raw Jaw wrote:

I really appreciate your very objective test.

Your being Objective should not preclude you from expressing your conclusions.

I would like to know your thoughts about the results of your comparative test.

Thank you.

They have about the same midtone noise profile. The D7100 has slightly less shadow noise at the ultra High ISO levels (ie, slightly better High ISO dynamic range). Color saturation and fidelity is a bit higher on the D7100, esp the blues and reds - this might owe to a more selective CFA on the D7100. The NX1 has significantly better base ISO DR in qualitative terms - much less banding and shadow tinting.

Randy Veeman Contributing Member • Posts: 799
This matches (confirms?) another test I saw
1

The A6000 which I think uses similar sensor as the Nikon exhibits severe banding.   The NX1 was even better in this test by Jordan of admiringlight.com:

http://www.fredmiranda.com/forum/topic/1344209/3#12829565

Just my opinion, but I prefer the NX1 over the Nikon up to ISO6400.

This part is not opinion based, for pulling detail from the shadows, the NX1 is the obvious winner.

 Randy Veeman's gear list:Randy Veeman's gear list
Samsung NX1 Samsung 16-50mm F2.0-2.8 Samsung 50-150mm F2.8 S
Randy Veeman Contributing Member • Posts: 799
Re: My NX1 vs D7100 High ISO test

Horshack wrote:

Kaziklu wrote:

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100.
I think this is fairly impressive though that NX1 holds it own.. the noise isn't significant but slight.

All images I posted were downsampled to 8MP, which normalizes the noise differences related to pixel count/density.

I was about to ask this.  Thanks.

 Randy Veeman's gear list:Randy Veeman's gear list
Samsung NX1 Samsung 16-50mm F2.0-2.8 Samsung 50-150mm F2.8 S
jennyrae Senior Member • Posts: 2,690
Re: This matches (confirms?) another test I saw
2

if lowlight picture quality is put in test, I do not believe a6000 should be even deserved on the same category. maybe under good light, but lowlight is not a6000 strongest point.

Kaziklu Senior Member • Posts: 1,301
Re: My NX1 vs D7100 High ISO test

Horshack wrote:

Kaziklu wrote:

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100.
I think this is fairly impressive though that NX1 holds it own.. the noise isn't significant but slight.

All images I posted were downsampled to 8MP, which normalizes the noise differences related to pixel count/density.

it doesn't though. it just makes the images the same size.

100% at 24mp and 100% at 28MP have a different number of pixels in it. Which is why Larger MP equals more noise at 100%

the same number of pixels from the 24mp sensor vs the same number of pixels on the 28mp sensor represent a different size of the image. As such a 1:1 comparison isn't a true comparison. More so a 8mp image is compressed to 33% of the original data for a 24MP image and 28% of the original data for the 28MP image. The heavier compression potentially can cause issues. More so you are sill trying to show an area of noise from the sensor that actually held more pixels.

There is no perfect ways to do it.

That being said I don't doubt the D7100 with Nikon's Experience and smaller Pixel Density has a slightly better noise profile in low light. Samsung is very new to this by comparison

Which just Highlights the power of BSI, the noise of a NX1 is exactly what one should expect for a 28MP of that Nikon Sensor if it could be upped that much and keep the same noise profile.

ie Yes the D7100 is slightly better, but the NX1 is very good being it is new technology for a sensor of this size... and it is the largest Pixel Density of any APS-C sensor.

The real test is in the images and how much they can be cleaned up... and want can be viewed by people not pixel peeping.

 Kaziklu's gear list:Kaziklu's gear list
Samsung NX1 Samsung NX20 Samsung NX 30mm F2 Pancake Samsung NX 60mm F2.8 Macro ED OIS SSA Samsung NX 50-200mm F4-5.6 OIS +1 more
OP Horshack Veteran Member • Posts: 7,921
Re: My NX1 vs D7100 High ISO test

Kaziklu wrote:

Horshack wrote:

Kaziklu wrote:

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100.
I think this is fairly impressive though that NX1 holds it own.. the noise isn't significant but slight.

All images I posted were downsampled to 8MP, which normalizes the noise differences related to pixel count/density.

it doesn't though. it just makes the images the same size.

100% at 24mp and 100% at 28MP have a different number of pixels in it. Which is why Larger MP equals more noise at 100%

the same number of pixels from the 24mp sensor vs the same number of pixels on the 28mp sensor represent a different size of the image. As such a 1:1 comparison isn't a true comparison. More so a 8mp image is compressed to 33% of the original data for a 24MP image and 28% of the original data for the 28MP image. The heavier compression potentially can cause issues. More so you are sill trying to show an area of noise from the sensor that actually held more pixels.

There is no perfect ways to do it.

Naturally there is no perfect way to do anything but there is a least-imperfect way to do something. The reason I downsample both images to a common size instead of one to the other is so that both undergo the same transformation even if the magnitude to reach that common resolution is different (and 28MP vs 24MP for the trip down to 8MP is not a large magnitude difference regardless). Lastly, this is same methodology photographers would use to render their photos, as most don't print at 1:1, so the noise results from this process models the noise they would achieve in their real-world handling of images.

HFLM Senior Member • Posts: 1,947
Re: My NX1 vs D7100 High ISO test
2

I initially expected the NX1 to be better since
a) BSI sensors have better QE and light sensitivty (http://www.samsung.com/global/business/semiconductor/news-events/press-releases/detail?newsId=13683). Although you have more pixels, the 30% higher sensitivity should compensate for this.
b) The NX1 has a newer generation sensor.
So it is amazing how well the (Toshiba?) sensor in the D7100 performs.

 HFLM's gear list:HFLM's gear list
Sony a7R III Sony a7 III Sony a9
Kaziklu Senior Member • Posts: 1,301
Re: My NX1 vs D7100 High ISO test

Horshack wrote:

Kaziklu wrote:

Horshack wrote:

Kaziklu wrote:

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100.
I think this is fairly impressive though that NX1 holds it own.. the noise isn't significant but slight.

All images I posted were downsampled to 8MP, which normalizes the noise differences related to pixel count/density.

it doesn't though. it just makes the images the same size.

100% at 24mp and 100% at 28MP have a different number of pixels in it. Which is why Larger MP equals more noise at 100%

the same number of pixels from the 24mp sensor vs the same number of pixels on the 28mp sensor represent a different size of the image. As such a 1:1 comparison isn't a true comparison. More so a 8mp image is compressed to 33% of the original data for a 24MP image and 28% of the original data for the 28MP image. The heavier compression potentially can cause issues. More so you are sill trying to show an area of noise from the sensor that actually held more pixels.

There is no perfect ways to do it.

Naturally there is no perfect way to do anything but there is a least-imperfect way to do something. The reason I downsample both images to a common size instead of one to the other is so that both undergo the same transformation even if the magnitude to reach that common resolution is different (and 28MP vs 24MP for the trip down to 8MP is not a large magnitude difference regardless). Lastly, this is same methodology photographers would use to render their photos, as most don't print at 1:1, so the noise results from this process models the noise they would achieve in their real-world handling of images.

That is just it.. it doesn't..

No one displays photos like this unless pixel peeping.

a 1:1 crop isn't something that is done.. yet that is what you've done. an image that is over 6k either way zoomed all the way in.

More so when shooting RAW no one would print an image in this way.

I really believe that these kinda comparisons though, they have their place are pointless.

 Kaziklu's gear list:Kaziklu's gear list
Samsung NX1 Samsung NX20 Samsung NX 30mm F2 Pancake Samsung NX 60mm F2.8 Macro ED OIS SSA Samsung NX 50-200mm F4-5.6 OIS +1 more
OP Horshack Veteran Member • Posts: 7,921
Re: My NX1 vs D7100 High ISO test

F> Kaziklu wrote:

Horshack wrote:

Kaziklu wrote:

Horshack wrote:

Kaziklu wrote:

The issue that will happen in any noise comparison between a 24, 20 12, or 18mp sensor with a 28mp sensor is that 100% is different. As such the noise is different. Which is why more MP usually means more noise. This shows that to be true.. but the BSI sensor shows that even without the experience of Nikon the BSI holds well. The NX1 is exactly what it should be and performs exactly as I expect it to. As well as the D7100.
I think this is fairly impressive though that NX1 holds it own.. the noise isn't significant but slight.

All images I posted were downsampled to 8MP, which normalizes the noise differences related to pixel count/density.

it doesn't though. it just makes the images the same size.

100% at 24mp and 100% at 28MP have a different number of pixels in it. Which is why Larger MP equals more noise at 100%

the same number of pixels from the 24mp sensor vs the same number of pixels on the 28mp sensor represent a different size of the image. As such a 1:1 comparison isn't a true comparison. More so a 8mp image is compressed to 33% of the original data for a 24MP image and 28% of the original data for the 28MP image. The heavier compression potentially can cause issues. More so you are sill trying to show an area of noise from the sensor that actually held more pixels.

There is no perfect ways to do it.

Naturally there is no perfect way to do anything but there is a least-imperfect way to do something. The reason I downsample both images to a common size instead of one to the other is so that both undergo the same transformation even if the magnitude to reach that common resolution is different (and 28MP vs 24MP for the trip down to 8MP is not a large magnitude difference regardless). Lastly, this is same methodology photographers would use to render their photos, as most don't print at 1:1, so the noise results from this process models the noise they would achieve in their real-world handling of images.

That is just it.. it doesn't..

No one displays photos like this unless pixel peeping.

a 1:1 crop isn't something that is done.. yet that is what you've done. an image that is over 6k either way zoomed all the way in.

More so when shooting RAW no one would print an image in this way.

I really believe that these kinda comparisons though, they have their place are pointless.

I'm not understanding your point. None of the images posted in this comparison were displayed at a 1:1 magnification. They were normalized to a smaller resolution, the same that would be done for an image from any high-resolution sensor when rendering to a smaller-resolution medium.

Randy Veeman Contributing Member • Posts: 799
Re: My NX1 vs D7100 High ISO test

HFLM wrote:

I initially expected the NX1 to be better since
a) BSI sensors have better QE and light sensitivty (http://www.samsung.com/global/business/semiconductor/news-events/press-releases/detail?newsId=13683). Although you have more pixels, the 30% higher sensitivity should compensate for this.
b) The NX1 has a newer generation sensor.
So it is amazing how well the (Toshiba?) sensor in the D7100 performs.

I agree with the other people, the difference is Nikon is more experienced as signal processing.  When the ISO 100 image is raised +7 EV, the Nikon results are not as good as the NX1 due to the banding.  Nikon processing thinks it is ISO 100 and doesn't do what it would at ISO 6400 or higher (no banding fix or 'cleaning up').  I may be wrong but Nikon is doing something with the RAW files to clean them up a bit while still retaining detail.

 Randy Veeman's gear list:Randy Veeman's gear list
Samsung NX1 Samsung 16-50mm F2.0-2.8 Samsung 50-150mm F2.8 S
Keyboard shortcuts:
FForum MMy threads