NEX-5N vs 5D mkII: (Part 3, raynox 6X as a tube lens)

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

seta666
Posts: 1063
Joined: Fri Mar 19, 2010 8:50 am
Location: Castellon, Spain

NEX-5N vs 5D mkII: (Part 3, raynox 6X as a tube lens)

Post by seta666 »

Hello,
I had been thinking for a while of buying a second body for macrophotography; I wanted an APS-C body but I think Canon APS-C sensors are a bit behind of what other brands are offering.

When I saw newer Sony NEX bodies had EFSC I knew that one of them would be that camera.

The two normal choices would be NEX-5N and NEX-7; newer NEX-F3 is out of the list because of minimum iso 200. I neither needed some of the features on the NEX-7 and the price difference was quite big.

Because Sony has anounced two new bodies (NEX-5R and NEX-6) there are some pretty good offers for the NEX-5N.
I paid 500€ for mine and sold the kit lens for 140€ two days later; so 360€ for new body with 2 years guarantie.

The NEX-5N offers EFSC, good DR at base ISO, focus peaking and articulated screen; all this things are nice for macro. It lacks cable shutter release and flash hot shoe but IR remote can be used and there are some ISO flash adapters available on ebay.

EDIT: I will divide the test in parts as I plan to update it with my findings

PART 1

This is the first stack I have done with that camera, with mitutoyo 5/0.14 and morfanon tube lens at 4X; I used white diffusor paper and a blue gel to convert the 3000K LED lamps to daylight 5400K (almost). I still have to properly test if there are any benefits on this as there is 1EV light losst (The idea came from a Charles Krebs answer to one of my questions on one of his posts).

85 shots with 0,04mm steps; there is some banding which suggests I should have used smaller steps 0,03mm maybe

Image

Large view
http://farm9.staticflickr.com/8177/7979 ... 22cf_o.jpg

The banding does not let fully apreciate the picture but I would say I am pretty happy with the results; in some aspects image quality is close to that of the EOS 5D mkII; resolution is lower but the difference is not as high as I would have expected. Dinamic range on the NEX-5N is higher, it lets recover detail from shadows which will be very handy for high contrast subjects

I did a small test when I bought the camera, I took a picture at home both with the NEX-5N and EOS-5D mkII; I scaled images to the same size 25mpx and lifted the shadows 4EV. The difference was pretty big on favour of the NEX
Camera A is nex-5N with kit lens at 23mm f5.6 0.4" ISO 100
Camera B is 5D mkII with sigma 24-60 at 35mm f8 0.8" ISO 100
Different aperture setting on 5D mkII to compensate lower DOF, exposure time compensated to match

Image
Large view
http://farm9.staticflickr.com/8180/7887 ... d41b_o.jpg

To me is amazing how some APS-C sensors have reached and surpassed some FF sensors, the results with newer EOS 5D mkIII would be pretty similar. This only shows difference in DR which is an important feature for me. (DR on nikon D7000 or pentax K5 would be even higher but this cameras lack EFSC)

I have read of people complaining of NEX-5N overheating when doing video; this could be a problem also for us as we use liveview continiously. I also read that when the screen is open the overheating is not that bad; I also read of someone using RAM heat spreader with good results (this would look awkward on the field but may be worth testing in the studio)

One thing that I have found on the NEX that I do not like is that the sensor is the stickiest thing I have ever seen; I almost break it because of cleaning the sensor.

Since a few months back I have been using EYELEAD SCK-1 to remove dust from my EOS 5D mkII; after this test shot I found there was a small hair that would not go with the blower so I used the eyelead. Man!!! I was leaving parts of the viscous material from the EYELEAD on the sensor!!! it works perfect on the 5D mkII but on the NEX is like it had glue, I am sure I could lift the camera with the EYELEAD thing.

So what then, I had my sensor with patches of viscous material which I did not know how to remove; I used my favorite sensor cleaning stuff; sensor-film.
http://www.sensor-film.com/cleaning.html
There is two kinds of sensor film, for sensors with fluoride coating and without; I used the one for non-fluoride sensors. It has taken me almost two days to remove it!! I am always able to remove it first time but with this sensor is like glue, I could feel the AA filter bending and everytime I tried to remove it paper strip would break. I contacted sensor-film and they told me this is one of the hard sensors.

I had to paint the paper strap and let it dry so it became hard as the stuff on the sensor, I could finally remove it.

I do not know what Sony put on that sensor but I would not recomend a cleaning method other than wet cleaning; compared to sensor-film wet cleaning leaves lots of stuff but we could say is less dangerous. I will probably use both methods. (I never do wet cleaning on my 5D)

I still have to do proper tests against 5D (same subject same framing) but I will try to make something this week.

PART 2


Today I have done a few stacks to see how big is the difference between a modern APS-C and a 3 year old FF camera

First I run a stack with mitutoyo 5/0.14 plus morfanon on both cameras; FOV choosen is 6mm so it accounts to 6X on the 5D mkII and 3.9 on the NEX-5N (the crop factor is not 1.5X but 1.54X)

I have two say that Mitu/morfanon combo does not perform very well at 4X on the corners, maybe I should start looking for another tube lens alternative; I do not know the reason but the morfanon seems not to perform very well under infinity in this case

Because corner quality on the 3.9X shot is not good I run another stack with JML 21/3.5;magnification is slightly higher and I changed the ISO to 400 (I did not realice)

Image

All pictures have been resized to 24mpx (6000x4000) before stacking, as I wanted to see the overall image quality at same printing size. There is not sharpening or noise reduction; only some light levels adjustments to even the images.

5D + mitutoyo at 6X Dmap http://farm9.staticflickr.com/8037/7983 ... f631_o.jpg

NEX-5N + mitutoyo at 3.9X Dmap http://farm9.staticflickr.com/8300/7983 ... 53b6_o.jpg

NEX-5N + JML at 4X Dmap http://farm9.staticflickr.com/8456/7983 ... 430f_o.jpg

5D + mitutoyo at 6X Pmax http://farm9.staticflickr.com/8298/7983 ... b999_o.jpg

NEX-5N + mitutoyo at 3.9X Pmax http://farm9.staticflickr.com/8180/7983 ... b4b1_o.jpg

I did not save the NEX-5N with JML Pmax as it was much more noise because of ISO400


Also here you have the crops of the different pictures; all around center crops but the last one which is left low corner.
Image

Full size http://farm9.staticflickr.com/8041/7983 ... 6fc7_o.jpg

If you compare the center crops difference is almost none, it would be very difficult to tell them appart from each other and do not forget that these are 100% crops at 24mpx. In my opinion is pretty amaizing!!.

The 5D/mitu at 6X is working at an effective aperture of f21, the NEX-5N and mitu at 3.9X at an effective aperture of f14; so even the pixel density on the NEX-5N is higher it gets balanced because of lower effective aperture.

Also one more benefit of lower magnification needed is more depth of field, so less shots are needed (115 vs 75) and also there should be less trouble with halos and transparencies.

Problem is the low performance of morfanon/mitu combo at lower than designed magnification; in my opinion same FOV should give similar corner performance but I see it is not as simple as that.

PART 3
Because the morfanon/mitu had low performance on the NEX at 4X I decided to try another tube lens; I bought a month ago a Raynox CM-3500 set second hand, I gave the 6X (5.9 diopter) a try

First thing this tube lens makes a very small kit, it just fits on the Jinfinance M42 Iris so the distance between the iris and the tube lens is just 2-3mm; I put some teflon tape on the raynos so it makes a tight fit inside the iris. Then I put the tube lens inside a frame made out of M42-52 step ring- 52-52 tube and 52-m42 step ring

Image

And now the results; I used the same subject as yesterday but lighting, background... everything changed. Also I forgot to resample the image before stacking, I did it afterward to the same 24mpx as the other images. I focused the tube lens to infinity which gave me 4.1X with the mitu 5/0.14

NEX-5N RAynox and mitu at 4.1X Dmap http://farm9.staticflickr.com/8304/7985 ... 3100_o.jpg
NEX-5N RAynox and mitu at 4.1X Pmax http://farm9.staticflickr.com/8319/7985 ... 2d96_o.jpg

And the crops
http://farm9.staticflickr.com/8442/7985 ... 1fe3_o.jpg

The performance in the center is very similar to the morfanon, but with better performance on the corners (compared to the NEX/morfanon picture). There are differences in color and levels but has nothing to do with the tube lens.

I think this is a set up I am going to use

Hope it helps ;-)

Regards
Javier
Last edited by seta666 on Fri Sep 14, 2012 9:15 am, edited 17 times in total.

pierre
Posts: 279
Joined: Mon Jan 04, 2010 12:37 pm
Location: France, Var, Toulon

Post by pierre »

Dear Javier,


Many thanks for your great feedback.

Do you have sense some lag shutter during you test?
Regards

Pierre

seta666
Posts: 1063
Joined: Fri Mar 19, 2010 8:50 am
Location: Castellon, Spain

Post by seta666 »

I would say shutter lag is 0 (sony claims 0,02"); the thing is that first curtain is so silent (cero noise) that you may feel there is a lag, because only second curtain sounds.
If you have long exposure (0.5" or 1") it may feel as a lag, only because there is no noise (on the 5D mkII you may hear a "zig"noise, here EFSC is completely silent).

Regards
Javier

pierre
Posts: 279
Joined: Mon Jan 04, 2010 12:37 pm
Location: France, Var, Toulon

Post by pierre »

Thanks for your answer Javier.

So I would consider there is no lag since the 2nd shutter does not impact the picture.
Regards

Pierre

jotafoto
Posts: 134
Joined: Fri Feb 26, 2010 8:20 am
Location: Sevilla-España

Post by jotafoto »

I see better even in the center the 5D MII. I do not understand why you have not tried the JML with Canon.
The difference is small, but I do look better the canon.
regards

rjlittlefield
Site Admin
Posts: 22605
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: NEX-5N vs 5D mkII, updated with new images

Post by rjlittlefield »

seta666 wrote:Also one more benefit of lower magnification needed is more depth of field, so less shots are needed (115 vs 75)
How did you determine these numbers?

I ask because it's not obvious to me there should be any difference. For example if you consider microscopyu's formula HERE, the first term depends only on NA and in the second term the changes to M and e will balance out if the pixel count is the same.

--Rik

seta666
Posts: 1063
Joined: Fri Mar 19, 2010 8:50 am
Location: Castellon, Spain

Post by seta666 »

jotafoto wrote:I see better even in the center the 5D MII. I do not understand why you have not tried the JML with Canon.
The difference is small, but I do look better the canon.
regards
Of course the 5D mkII picture is better, but not that much in my opinion. I am not sure if the slight resolution is higher because of being full frame or because its higher resolution of 21mpx. Do not forget we are comparing 21mpx FF vs 16mpx APS-C. In my opinion they are on par.

Because at 6X JML and mitu on the 5D are about the same; I would say the mitu is slightly better. I just wante to see if the low corner performance had only to do with the lens combo or could be because of sensor design.It is known that the NEX-7 has had some problems (color shifts, low resolution) with corner performance with some lenses (Some rangefinder wide angles specialy).
rjlittlefield wrote: How did you determine these numbers?

I ask because it's not obvious to me there should be any difference. For example if you consider microscopyu's formula HERE, the first term depends only on NA and in the second term the changes to M and e will balance out if the pixel count is the same.

--Rik
Well, the truth is that for low magnification I use a DOF table based on lefkowitz formula; I just treat the mitutoyo 5/0.14 as a f3 lens. (I use this DOF table in the field with the 5D +MP-E)

At 6X a f2.8 lens would have a DOF of 0.03mm (I used 0.02mm steps for the 6X shot); at 4X DOF would be 0.04mm but because I should use a lower CoC of 0.020 for the APS-C camera I used 0.027mm steps. Maybe they both could work OK with a same step size of 0.03mm ;-)

I should make another table with this 0.020 CoC for the NEX but for now I just guess ;-) In the first picture of the tenebridae you can see this guessing did not work.

Also normally I preffer to take some more pictures than to be short; I only try to get the right size step when doing deep stacks with 200+ images

Regards
Javier
Last edited by seta666 on Thu Sep 13, 2012 3:08 pm, edited 1 time in total.


rjlittlefield
Site Admin
Posts: 22605
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: NEX-5N vs 5D mkII, updated with new images

Post by rjlittlefield »

seta666 wrote:
rjlittlefield wrote: How did you determine these numbers?

I ask because it's not obvious to me there should be any difference. For example if you consider microscopyu's formula HERE, the first term depends only on NA and in the second term the changes to M and e will balance out if the pixel count is the same.

--Rik
Well, the truth is that for low magnification I use a DOF table based on lefkowitz formula; I just treat the mitutoyo 5/0.14 as a f3 lens. (I use this DOF table in the field with the 5D +MP-E)
OK, thanks for the explanation. So most of the calculated difference is not real, but instead it's due to various approximations.

There is a recurring myth that sensor size has some big effect on DOF. In truth, it does not. What matters is the angle of the entrance cone. Given the same subject width, the same entrance cone, and the same pixel count, all sensor sizes will have the same DOF and the same diffraction blur.

With a finite setup, the larger sensor will require more magnification, which will move the lens slightly closer to the subject. In that case, the entrance cone will get a little wider, producing slightly less DOF and slightly less diffraction blur to go along with it. This effect does not occur in an infinite setup where different magnifications are produced by changing the focal length of the tube lens. In the infinite setup with different tube lenses, sensor size doesn't matter at all.

It sounds to me like you're using the Mitutoyo+morfanon as a finite combination, changing magnification by changing the extension. In this case the larger sensor will have slightly less DOF. But offhand I don't think it's nearly as much as 115 vs 75, and in any case the difference just says that if you're willing to sacrifice resolution by using a slightly smaller effective aperture, then you can get by with fewer steps.

--Rik

seta666
Posts: 1063
Joined: Fri Mar 19, 2010 8:50 am
Location: Castellon, Spain

Post by seta666 »

I guess other way we could look at this would be something like this:
We have two images with the same FOV; one is made out of 16mpx (NEX-5N) and the other of 21mpx (5D mkII). The one with higher pixel count will have less DOF. Does this make sense?

I take more shots than needed on low magnification shots (bellow 10X) for few reasons:

- The total amount of images is not a big deal, normally is well under 100
- When doing very low magnification shots (something that happens often in the field) having more images makes stereos look good
- I also feel that the higher the overlap the higher the final image quality; I guess you normally can do with a 10-20% overlap but a 30-40% overlap uses only the sweet spot from each individual image.
-Also when working with live subjects sometimes you may find there is movement in one or more shots. With high overlaps normally does not matter if you have to delete one image

I think that now I know the way my EOS 5D mkII behaves, I do not normally calculate the DOF needed and sometimes I just "guess". I However I still have to get to know the way the NEX sensor behaves.

Regards
Javier

rjlittlefield
Site Admin
Posts: 22605
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

seta666 wrote:We have two images with the same FOV; one is made out of 16mpx (NEX-5N) and the other of 21mpx (5D mkII). The one with higher pixel count will have less DOF. Does this make sense?
Sure. Again, this is saying that by sacrificing a little resolution you can get more DOF. You would have to shrink the aperture a little bit too so as to optimize that tradeoff, but just changing the pixel count has some effect.

--Rik

Pau
Site Admin
Posts: 5742
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

Javier, nice test. I think this will be useful for other people. Can the Sony NEX work tethered to the computer like the Canon EOS?

About your camera bag picture, whithout questioning that the Sony may have better DR, the left images show different in my monitor, being the B (canon) one more underexposed, so the +4EV test may not be acurate
seta666 wrote:We have two images with the same FOV; one is made out of 16mpx (NEX-5N) and the other of 21mpx (5D mkII). The one with higher pixel count will have less DOF. Does this make sense?
I think it only makes sense if viewed at 100% and if the lens has enough resolution. If you view the full images at the same size without oversampling the smaller one it wouldn't apply. (please correct me if I'm wrong)
Pau

seta666
Posts: 1063
Joined: Fri Mar 19, 2010 8:50 am
Location: Castellon, Spain

Post by seta666 »

Pau wrote:Javier, nice test. I think this will be useful for other people. Can the Sony NEX work tethered to the computer like the Canon EOS?

About your camera bag picture, whithout questioning that the Sony may have better DR, the left images show different in my monitor, being the B (canon) one more underexposed, so the +4EV test may not be acurate
Well; I do not use the NEX on the computer and the truth is that I have no idea about that, my computer and macro set ups are separated so I did not bother trying it.

About the exposure on the bags; it should be the same. They were done with 1 minute difference in the same room, maybe the diference is because the way their ISO are calibrated.
Anyway, the exposure difference is not that big to make such a big difference in shadow noise.

I also did the same with more cameras downloading the RAWS from Dpreview; all RAW were resampled to 26mpx
100% crops
A Pentax K5 B Canon 60D C Nikon D800 D Canon 5D mkIII E Canon 5D mkII F Sony NEX-5N

http://farm9.staticflickr.com/8038/7905 ... edef_o.jpg

Regards
Javier

Post Reply Previous topicNext topic