Need some help for NX200 test

[verybiglebowski]

I started to compare Samsung NX200 with Sony NEX-5n and Canon EOS 550D.

Originally I planned to do only field tests, but because of the bad weather here, I started with some studio shots.

Lack of Adobe RAW converter for NX200, forced me to use their bundled Samsung converter.

However, files that are coming from it, without sharpening are noticeably softer then those from Nex-5n and EOS 550D, developed straight in LR. (also without sharpening).

I published some results on my blog: http://verybiglobo.blogspot.com/ , and wrote about problem with conversion, but maybe someone here can have an idea why it is so.
Possible conclusions that I was able to come are:

1. Maybe, Sony and Canon add some sort of sharpening when compressing their RAW's in camera. Samsung - due to the huge RAW file size, seems to deliver pure RAW.

2. I used center focus area for AF and also MF. In MF, Sony peak assistance is really very helpful. On Canon and Samsung I have to believe my eyes and LCD's. Due to the smaller resolution of the Samsung LCD, I could misjudged sharpness, even in the highest magnification.

In AF, center square area has a different size among cameras. The problem however is, that none part of the image is sharper then Sony or Canon.

3. I tried those tests at 3 FL - 18,35 and 55mm. I published only 18mm for ISO Chart and 55mm for ISO performance. Results are even worse at 35mm for Samsung. Could it be my lens sample?
4. Any other Idea what I possibly did wrong?

I have also jpegs, the difference is obviously smaller, but I am not 100% sure about in camera jpeg engines, how similar they are. Canon has the softest JPEG, (sharpness and contrast set to +1, same as with Samsung).

Here is an example of ISO test at 800. Maybe the difference wont be that visible at small size, but at crops on my blog it is quite noticeable. I retouched those pictures here in LR in effort to equalize WB, and contrast, but the sharpening and noise reduction are still set to 0.

Sony





Samsung





Canon





--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[random78]

You are doing these tests with wide-open aperture on a kit lens at the extreme ends of the focal length range - so many factors are wrong here!. You are are primarily testing the kit lens quality rather than the camera performance

1. At the very least stop down the lenses a couple of stop,s and if possible also get hold of better lenses for each camera

2. Also use the magnified view with manual focus to exactly set the focus plane. It should be doable. And the added depth of field from smaller aperture should reduce the impact of focusing errors as well.

3. Dont use a target which is too close - many lenses don't perform that well close to minimum focus distance. Also at larger focus distances you will get more depth of field which will again help with avoiding out-of-focus blur.

4. Make sure the samsung RAW converter is not applying a lot or NR by default. Samsung's raw converter will probably be applying an NR similar to the NR in JPEGs which is relatively strong.

Hope this helps!

 

[Ariston]

1.> set the lens focal length at 28mm.
2.> set aperture at f8
3.> turn-off OIS
4.> timed shutter and set on a tripod

5.> use manual focus and set magnification at 5x or 8x. whichever suits focus the best.
6.> set shutter speed above 200 if possible.
7.> shooting distance set at 8-10 feet.
8> use sufficient lighting.

as to why the results are like that, here are two reasons why.

1.> the image suffer from blur ( it was discussed that at such slow shutter speed, the camera suffers from shutter induced vibration). you need to shoot at faster shutterspeed.

2.> SilkyPix is poor when it comes to resolving detail and sharpness. I noticed this when I was assessing the NX100 RAW files both from SilkyPix and LR. LR renders the image better even if both image softwares have identical parameters (NR OFF, SHARPNESS set to O), etc.
better wait for LR/ACR support.

 

[verybiglebowski]

1.> set the lens focal length at 28mm.

I might try, but I have shots at 35mm and they are even worse for Samsung.

2.> set aperture at f8

I'd love to, but will need much more light.

3.> turn-off OIS

Sure, checked twice.

4.> timed shutter and set on a tripod

10s Self Timer. Tripod is for large format cameras

5.> use manual focus and set magnification at 5x or 8x. whichever suits focus the best.

I did it. Once you are in MF and touch the ring, picture magnifies. I selected maximum magnification.

6.> set shutter speed above 200 if possible.

Light, light...

7.> shooting distance set at 8-10 feet.

I tried to stay close to 20x FL.

8> use sufficient lighting.

I know, I know... I had two lamps each with 4 day light bulbs of aprox. 350W. I will try to use instead 3 halogens -f 800W each. That's all I have. I would normally use the studio strobes, but no way with Samsung or Casio.

as to why the results are like that, here are two reasons why.

1.> the image suffer from blur ( it was discussed that at such slow shutter speed, the camera suffers from shutter induced vibration). you need to shoot at faster shutterspeed.

YOU NAILED IT! I think this is the most probable reason. I forgot about that. Unfortunately, if this is true for all ISO settings, then it sounds like a serious flaw, as it is impossible to use camera for longer exposures, which might disqualify landscape and architecture shooters.

2.> SilkyPix is poor when it comes to resolving detail and sharpness. I noticed this when I was assessing the NX100 RAW files both from SilkyPix and LR. LR renders the image better even if both image softwares have identical parameters (NR OFF, SHARPNESS set to O), etc.

Yes I checked that twice, trying to equalize post processing as much as possible. But thank you noticing Silky Pix. I never had Samsung before, so it is good to know.

better wait for LR/ACR support.

I agree, I will repeat processing.

Thanks for your help. Much appreciated.

--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[verybiglebowski]

You are doing these tests with wide-open aperture on a kit lens at the extreme ends of the focal length range - so many factors are wrong here!. You are are primarily testing the kit lens quality rather than the camera performance

Yes, lenses are what I am usually testing. I did the test at 35mm too, but the results are even worse for Samsung, so I didn't want to publish them, until I figure out, the problem.

1. At the very least stop down the lenses a couple of stop,s and if possible also get hold of better lenses for each camera

That is a light problem. I can't use strobes with them and at 100 ISO you need really lot of light for proper exposure. At least all cameras and lenses are tested at same conditions.

2. Also use the magnified view with manual focus to exactly set the focus plane. It should be doable. And the added depth of field from smaller aperture should reduce the impact of focusing errors as well.

I did use maximum magnification for Samsung and Canon. (for ISO chart) However, tolerances are very small, so it might be my fault. Nevertheless, results with Canon are better, despite its kit lens horrible MF ring.

3. Dont use a target which is too close - many lenses don't perform that well close to minimum focus distance. Also at larger focus distances you will get more depth of field which will again help with avoiding out-of-focus blur.

I tried to stay at aprox. 20x FL. 18mm is extreme, I was planning to use 35mm, but was interested in the distortion. However 35 looks really bad, so I would like to repeat it.

4. Make sure the samsung RAW converter is not applying a lot or NR by default. Samsung's raw converter will probably be applying an NR similar to the NR in JPEGs which is relatively strong.

As far as I was able to understand their RAW converter, I switched all corrections off.

Hope this helps!

Thanks a lot for your time.

--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[Adventsam]

c/w the others? same ap, same iso, same fl, not same same or even similar shutter but same illumination?

I started to compare Samsung NX200 with Sony NEX-5n and Canon EOS 550D.

Originally I planned to do only field tests, but because of the bad weather here, I started with some studio shots.

Lack of Adobe RAW converter for NX200, forced me to use their bundled Samsung converter.

However, files that are coming from it, without sharpening are noticeably softer then those from Nex-5n and EOS 550D, developed straight in LR. (also without sharpening).

I published some results on my blog: http://verybiglobo.blogspot.com/ , and wrote about problem with conversion, but maybe someone here can have an idea why it is so.
Possible conclusions that I was able to come are:

1. Maybe, Sony and Canon add some sort of sharpening when compressing their RAW's in camera. Samsung - due to the huge RAW file size, seems to deliver pure RAW.

2. I used center focus area for AF and also MF. In MF, Sony peak assistance is really very helpful. On Canon and Samsung I have to believe my eyes and LCD's. Due to the smaller resolution of the Samsung LCD, I could misjudged sharpness, even in the highest magnification.

In AF, center square area has a different size among cameras. The problem however is, that none part of the image is sharper then Sony or Canon.

3. I tried those tests at 3 FL - 18,35 and 55mm. I published only 18mm for ISO Chart and 55mm for ISO performance. Results are even worse at 35mm for Samsung. Could it be my lens sample?
4. Any other Idea what I possibly did wrong?

I have also jpegs, the difference is obviously smaller, but I am not 100% sure about in camera jpeg engines, how similar they are. Canon has the softest JPEG, (sharpness and contrast set to +1, same as with Samsung).

Here is an example of ISO test at 800. Maybe the difference wont be that visible at small size, but at crops on my blog it is quite noticeable. I retouched those pictures here in LR in effort to equalize WB, and contrast, but the sharpening and noise reduction are still set to 0.

Sony





Samsung





Canon





--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[paulhome]

while you're doing you tests, can you observe shutter speeds. the 3 shots above all look correctly exposed at iso800 but 1/40 on the sony vs 1/15 samsung is slightly worrying.

 

[Ariston]

YOU NAILED IT! I think this is the most probable reason. I forgot about that. Unfortunately, if this is true for all ISO settings, then it sounds like a serious flaw, as it is impossible to use camera for longer exposures, which might disqualify landscape and architecture shooters.

it's not the same for all ISO setting since you will increase shutterspeed as well which negates shutter vibration. it does not affect longer exposures since shutter only moves slightly when opening and closes slightly after a long period. maybe you should do longer exposure test with ND filter?

 

[Ariston]

while you're doing you tests, can you observe shutter speeds. the 3 shots above all look correctly exposed at iso800 but 1/40 on the sony vs 1/15 samsung is slightly worrying.

maybe he could set exposure to manual, rather than aperture priority. then adjust shutter speed.

 

[hpchan]

Samsung NX200 sample shown the best DR of all three pictures.

PP before shooting: Canon win deeper colour, on user setting NX200 try saturation +2.

Canon 55mm have great reflection on iris plate; the best sharp pic would be obtained on F1.4 of F1.2 lenses. F5.6 is not the best choice. If the test lenses is F1.4 try F2.8 for best colour. NO sharpening, or additional + on contrast. When the wide range of brightness and shadows obtain, I think NX200 is the winner. Just NX experience apply to NX200.
--
hpchan

 

[Identity]

My only advice would be to include more pages from that Fine Art Photo book so that we can closely observe sharpness and detail

 

[random78]

From your responses it seems that you are not using a tripod and doing handheld tests. And combining with the fact that you are doing 55mm shots (ie 80mm equivalent) at 1/15s, no wonder you get all types of blur (and you are not even using OSS its seems).

This has got nothing to do with the the shutter-induced blur that dpreview talked about. You are basically using a far too slow shutter speeds to handhold.

I am sorry but your testing methodology is sooo utterly flawed that there is not much use discussing this further until you fix your methodology. You have to do at least the following to make this testing close to being anything serious:

1. Use a tripod to eliminate hand shake
2. And now since you have a tripod so set all lenses to something like f8.

 

[verybiglebowski]

From your responses it seems that you are not using a tripod and doing handheld tests. And combining with the fact that you are doing 55mm shots (ie 80mm equivalent) at 1/15s, no wonder you get all types of blur (and you are not even using OSS its seems).

I am using a tripod. Heavy studio tripod designed for a large format cameras. I wrote that in my responses, and will post the picture of the setup, to stop that kind of doubts.

This has got nothing to do with the the shutter-induced blur that dpreview talked about. You are basically using a far too slow shutter speeds to handhold.

Actually it is probably not due to shutter-induced blur, because, that may happen at slower shutter speeds I assume. But with a higher ISO, shutter speed is fast enough to eliminate any vibrations, especially on such a tripod.

I am sorry but your testing methodology is sooo utterly flawed that there is not much use discussing this further until you fix your methodology. You have to do at least the following to make this testing close to being anything serious:

1. Use a tripod to eliminate hand shake

I am using it

2. And now since you have a tripod so set all lenses to something like f8.

Problem with aperture is a light problem. I repeated tests today and I am working on files now. I switched-on everything I had and get to f7.1 at 28mm (aprox) and 1/40s at 100 ISO.

Anyway, whatever aperture I used first time, it is used same on all cameras. Shutter speed varies, but that is due to certainly different metering from Samsung. Repeating tests today, I used manual exposure, but at the same settings, Samsung was at 0 EV in most cases, while Canon and Sony at +0.7...

Once again, I don't believe, that Samsung kit lens is significantly worse then Canon or Sony, at least reviews doesn't suggest so, and even in the first test, all cameras were used under same condition, with very different results.

--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[verybiglebowski]

My only advice would be to include more pages from that Fine Art Photo book so that we can closely observe sharpness and detail

For such an inspiring advice, I will reward you soon Be patient
--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[verybiglebowski]

Once again thank you for a great advice. It really make most sense. I took this picture from Fine Art Photography just for you, to let you enjoy creative lighting and composition. I even took a time to cover those disruptive parts of the images, that might disturb your artistic photo soul.
Enjoy.





--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[random78]

I am sorry if I sounded offensive but I got the impression from your posts that you are not using a tripod.

I dont understand why you have a light problem if you are using a tripod. You can easily extend the exposure a bit longer.

And yes you were using all the kit lenses at similar settings but still there can be significant variation in terms of performance so stopping down is essential to reduce such variations.

Hopefully after the steps you have taken, you will get more consistent results. Good luck!

 

[verybiglebowski]

No offense taken, I understand that most people are first looking at usual mistakes. I am testing, (mostly lenses) for over an year now, but I always note, that my tests are not scientific. I am financing my tests myself and could only work with what I can afford. I am also usually using powerful strobes for testing, but this was not possible with Samsung and Casio, so I had to dig my old continuous lights.
Here is how my set look up. I love my tripod BTW





Longer exposure can be compromised by shutter-induced blur, if that problem exists, and many here advised to go for higher shutter speeds.

I will process new files and will let you all know, if the results are any better.
Thanks.
--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[Ariston]

Longer exposure can be compromised by shutter-induced blur, if that problem exists, and many here advised to go for higher shutter speeds.

dpr did a longer exposure test with their ISO 100-800 samples to compensate for shutter induced blur. they even put ND filter to compensate for overexposure. the results was the images were sharper now and focused as opposed to having to shoot at 1/13 or 1/15 which were giving inconsistent results in the ISO levels (some where sharper, some were soft).

what it basically means is that such problem could exist at 1-3 sec exposure delay to about 1/15 or 1/30 (rough estimates). although such blur is only noticeable at pixel level or unless you blow up large prints.

since you are shooting at a studio, why don't you do long exposure with the NX200 and see if the results are better? let's say 5-8 sec exposure delay? or better yet, refer to dpr's test exif on the NX200.

 

[verybiglebowski]

As I wrote I am not expert on continous light, but I managed to get exposure at 1/40s for f7.1 and ISO 100, measured with Sekonic light meter. While Samsung shows correct exposure, other two cameras shows EV +0.7.This is the reason for difference in shutter speed in previous effort.
I am afraid that longer exposures involves even more unreliable result.
We will see, once I manage to process those small Samsung RAWs.
Thanks.
--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/

 

[verybiglebowski]

I repeated tests, talking many of your advises into consideration, but the result is still disappointing.

This time it is manual exposure, f7.1, FL 27(28) mm, MF and AF, taken 3 shots of each mode and scene and selected sharpest one.

I think, there must be something about processing RAWs in camera during compression in Canon and Sony.
The only other explanation is, that my Samsung is faulty :-(

100% Crops:













--
http://www.flickr.com/photos/viktor_viktor/
http://verybiglobo.blogspot.com/