Testing High ISO and DOF on X-T1 vs GX7

[jonbrisbincreative]

I rented a Fuji X-T1 for a shoot this weekend (as well as to check it out on the football field). I'll probably write up an article comparing it to my GX7 at some point. I wanted to share this screenshot of a test frame I did to check both the DOF of the APS-C sensor in the X-T1 compared to the mu4/3 sensor but also to check the purported superiority of the X-Trans sensor in high ISO situations.

It is true that--with zero noise reduction--the RAW files from the GX7 show noticeable noise at ISO 3200 whereas the X-T1 RAW files actually need very little noise reduction. They look really clean. However (and this is a big "however") when I applied noise reduction (Lightroom 5), the mu4/3 file was honestly not that much different than the X-Trans file. Here's the 1:1 comparison:

Subscribe to see EXIF info for this image (if available)

Despite the differences in FOV between the effectively 50mm PanaLeica and the effectively 35mm Fuji (both set to f2.0), I don't see a tremendous difference between the two (other than a noticeable difference in color rendition). The GX7 files took noise reduction very well and I can go up to 70-80 on the slider, which takes out enough of the noise to make it comparable to the X-T1.

I also don't see much difference in DOF between the 25mm Leica and the 23mm Fuji. Both cameras were placed on a tripod in exactly the same position relative to the Mamiya RB67 and I focused on the same point (the "M" in Mamiya). Granted the DOF will be different at subjects that are further away, but in my testing so far, I'm not seeing enough of a difference to be a huge factor. I've found the DOF argument so far to be a straw man.

I'm evaluating whether it makes sense to use the X-T1 instead of the GX7 since they both are about the same size and weight and will be used for nearly the same function. I'm not convinced I can part with my 7D yet, so I added an L 200/2.8 prime. The autofocus in the 7D is a beast and the back-and-forth of the X-T1 is probably not going to be ideal when I get it into low light and with fast-moving action. The GX7 with PanaLeica actually does really well in the AF department. It's not as fast as the 7D but it's faster IMO than the X-T1.

So far I like the X-T1's knobs and rings. Some people don't but I find them very convenient and they feel natural. But I also really like the GX7's touchscreen. As great as the knobs on the X-T1 are, the touchscreen is more useful since I can use it to set the autofocus point, manipulate menus, and scroll around a zoomed-in image in playback mode. That in itself is a huge time-saver.

The viewfinder in the X-T1 is everything it's made out to be, however. It blows the pants off the GX7 viewfinder. It's faster to switch when you put your eye up to it and the X-T1's works better all around (I find it easy to fool the GX7's sensor if I put my finger near it but am instead using the touchscreen; that gets old pretty quick).

I think it's actually going to come down to the usability of either with respect to physical controls and menus. The Panasonic has the edge in the menu department by several miles but the Fuji has the manual controls that feel good to use (except for the back arrow pad which really is a pain).

 

[Livnius]

Thanks for the effort Jon, but surely this type of test (the noise aspect aside) is best served when the 2 lenses chosen are of the same FOV.....here, you're putting a 35mm up against a 50mm (effective).

For what it's worth, I think the Fuji is picking up more of the fine dust particles and is showing greater detail with regards to the Mamiya bodies textures. But, I'd be hesitant to try draw too many concrete conclusions based on a single photo.

Would love to see you do a more comprehensive test with multiple photos using the Fujinon XF 35/1.4......a lens significantly closer to the PL25 in focal length.

 

[jonbrisbincreative]

Here's another test of just DOF on the Fuji 18-55mm @ 35mm (effectively 50mm) compared to the GX7 25mm (effectively 50mm):

Subscribe to see EXIF info for this image (if available)

It seems to my eye that the DOF is certainly different between the two (both @ f4) when you look at the barbs on the fence but the OOF areas in the mu4/3 image aren't that much different given I was not in the exact same spot between the two frames (this isn't a scientific test and I don't intend it to be because I frankly don't care that much; I'm interested in the differences in real use). That stands to reason, of course, and I'm not trying to make the case that physics doesn't apply. But I am looking to see whether the OOF areas from the X-T1 + 18-55 combo are "better" than the GX7 + PanaLeica. I know that's subjective and I'm not going to try and define what "better" is for me or anyone else.

It does seem like there's some serious FUD out there around DOF of mu4/3 vs APS-C and the difference isn't that great (though for some people it might be enough, which is perfectly fine; I say to-may-toe you say to-mah-toe we both say the same darn thing). I shot several more test frames of different things but they essentially tell the same story.

I won't try and make any judgements publicly about what I think is more desirable. I'm just interested in cutting through the noise (of the ISO variety as well as the forum post variety) and trying to figure out what the actual differences between the two sensors are versus what some of these chuckleheads on various forums say there are. IMO the differences in operation between the two cameras are every bit as significant as the differences between the sensors. I'm finding the viewfinder of the X-T1 to be just stellar compared to the GX7. In fact, the more I use the X-T1 and GX7 together (taking a shot with one camera then switching to the other and taking the same shot again), the more I find the slow response of the viewfinder in the GX7 when I put my eye to it to be very, very annoying. I also find the contrast of the GX7's VF to be significantly different from the X-T1 and different even from the GX7's LCD.

The autofocus on the X-T1 is a bit frustrating when using C-AF. I can use C-AF on the GX7 without much problem at all and I don't really notice it. But the image is continually going into focus and out of focus by quite a bit on the X-T1. It also seems to "give up" at times, stopping the AF when the subject isn't actually in focus. But on the other hand the X-T1 has a fantastic frame rate in burst mode. Much faster than the GX7's.

 

[jonbrisbincreative]

For what it's worth, I think the Fuji is picking up more of the fine dust particles and is showing greater detail with regards to the Mamiya bodies textures. But, I'd be hesitant to try draw too many concrete conclusions based on a single photo.

I actually agree completely. The Fuji's larger sensor is clearly picking up more detail (in large part, I suspect, due to not having to have noise reduction applied) at 1:1. But when you view both images at "normal" size on a monitor, the difference is (completely my opinion) not nearly as noticeable.

The reason I rented the X-T1 was to see if I'm "missing out" on something. I keep hearing about how much better the X-Trans sensor is than mu4/3 and, while a smaller sensor clearly has more issues with noise at high ISO with no noise reduction, I'm not personally seeing the gargantuan difference that some internet posters say there is. I'm not saying there's no difference, mind you. I do have eyes and can clearly see that there is a difference. But does that difference actually contribute to a tangible improvement at viewing resolutions as well as print? Of that I'm less and less sure.

 

[yakky]

Having owned an X100, XE1 and a G5, EM5 and lots of other stuff, I can say without a doubt that the "amazing detail" Fujis pull in is mostly hype when it comes to M43 vs Fuji. Compare a Fuji to an APSC Nikon with a good lens and things really start to look really bad for the Fuji. Fuji is king of great colors, and to me that holds true in the above example, they also mop the floor with white balance. I love the controls as well. Focusing.. at least on my examples was terribly slow, and I think the reality on the XT1 is setting in as well.

That said, Fuji hype is dying, people are dumping bodies and the sister Fuji forum just doesn't have much traffic on it until a new camera is released, then its back to quiet.

You were smart to rent a body.

 

[brettmaxwell]

The viewfinder switching lag of the GX7 drove me crazy as well, and you can't make it stay on like the EM5, this was the primary reason I sold it!

 

[robbie36]

For what it is worth. If you are going to test iso sensitivity it is very important that you use the same settings at the same time. For instance the GX7 is 1/125 @ F2 and the Fuji is 1/80 @ F2. So inherently the Fuji is getting half a stop (or 50%) more light per sqmm of sensor than the Panasonic. And that is bound to make a significant difference to how the photo will turn out.

 

[tomO2013]

In this example the Fuji is pulling in a lot more detail. However I wouldn't call it conclusive as you are testing different focal lengths and effectively different exposure values as Rob already pointed out. Also consider that DPReview have already mentioned
It would be interesting too to see how both look when pushed through PhotoNinja (there is a trial version available). The Fuji files typically react significantly better to PhotoNinja from a detail and colour handling perspective.

I was at a workshop recently where a photographer was showing before and after shots with the XT1 files on LR and Iridient. The difference was night and day. Did the images look technically 'better' from the alternative Raw processors when dealing with RAF files? Definitely if I am honest, yes, perhaps it was the first time that I could see where Fuji's engineers were going when they claimed that their proprietary design could give small format rivalling results in terms of cleanliness and detail - you may be hard pressed to tell the difference between the A7 and the XT1 assuming you used Iridient/PhotoNinja etc... but this is a real problem for Fuji. Not everybody wants to have to seriously alter their workflow just to leverage that theoretical potential that Fuji have always claimed was locked away in their proprietary sensor design.
I'm not knocking Fuji. I do think the XT1 is an excellent camera and it is nice to have an option that gives a different 'look' OOC to every other bayern sensor camera out there. But the Raw situation is still a pain in the ass and would probably be a serious consideration for me ever moving back to Fuji again.


--Tom.

 

[jonbrisbincreative]

For what it is worth. If you are going to test iso sensitivity it is very important that you use the same settings at the same time. For instance the GX7 is 1/125 @ F2 and the Fuji is 1/80 @ F2. So inherently the Fuji is getting half a stop (or 50%) more light per sqmm of sensor than the Panasonic. And that is bound to make a significant difference to how the photo will turn out.

Fair enough. But it is interesting that the Fuji needed that extra half stop considering all other factors were the same. And in truth, the only thing I cared about being the same was the aperture because the test was "if I set both these cameras to f2 or f4 or whatever, what does the image the camera makes look like based on that?" Not a scientific test, but that's not what interests me. I want to know what the cameras are going to do in a given situation. I want to rely on the camera to do its job in the auto department when I set an aperture and am simply evaluating how well the X-T1 does that versus the GX7.

If the Fuji produces less noise because it overexposes compared to the GX7 then that's an important part of my consideration. I already know ahead of time the sensors perform differently and exhibit different characteristics but what I don't know yet is: when I put the X-T1 in a given situation and try to make an image, what happens? If it produces an image with less high ISO noise, I frankly don't care how it achieves it because what's most important to me usually is aperture; except for shooting sports, then it's shutter speed.

 

[jonbrisbincreative]

It would be interesting too to see how both look when pushed through PhotoNinja (there is a trial version available). The Fuji files typically react significantly better to PhotoNinja from a detail and colour handling perspective.

Here's both RAW files processed through PhotoNinja (trial) to try and get exposure and white balance to match each other (both cameras has wildly different ideas about what temperature the incoming diffused lighting was).

Subscribe to see EXIF info for this image (if available)

As I said before: the X-T1 files were closer to the mark regarding high ISO noise than were the GX7 files. That said, noise reduction has basically wiped out significant differences so that results obtained from the Fuji can be approximated very closely from the GX7.

Even DxO Optics Pro has better noise reduction than Lightroom but it sadly doesn't support the X-T1 so I didn't use it.

PhotoNinja is actually pretty cool software. It's straightforward to use and the results are really quite nice.

 

[tomO2013]

The results from the second shot look much better on both cameras - showing that a good RAW processor can make as big an impact as a sensor!
The colour handling and AWB subjectively looks much better on the fuji overall and there is no doubt a lot more detail retained in the patterns in the blacks. The Mamiya writing is also a lot clearer and sharper on the Fuji. It's hard to know without knowing what NR setting that you were using in PN, but generally speaking I find it's noise reduction very good. There is a nice feature in it too where you can include some residual noise, and generally speaking for darker patterns with fine detail such as your mamiya camera, that is a nice balance between NR and also maintaining detail.
As for the overexposing/underexposing issue.... at lot has been written on the DPReview forums about this. The best thing of the lot if you are testing this would be to get out your hand held light meter, take a reading and set both cameras manually for the sake of fun and giggles to the same exposure and see how they both take the shot. I did this with the XE1 and EM1 and didn't really see much difference, the XE1 may have been fractionally brighter. But there are those on DPReview who claim that it over exposes by a half stop. I'm not sure I ever saw that myself but YMMV.
BTW if you were to run those same shots through the bundled Silkypix you would be horrified by what it does to both the GX7 and XT1 at high ISO!

Did you use the same NR settings in PN on both images ? You would probably get away with significantly less NR on the Fuji. Similarly on the GX7 I'd bump up the residual noise slider to get a more realistic looking image.
I've seen examples comparing A7's , EM1's and XT1's at ASA3200 and with good NR at ISO3200, the differences are a LOT more subtle than the physical size of the sensor would have you believe. At ISO6400 and above, the XT1 and A7 would tend to be in a different level, but that is why on the m43 format we have IBIS
If you want to pull more XT1 details ... http://www.marksoon.com/blog/2014/fuji-x-trans-raw-processing take a look at that.

 

[Deleted member 20897]

Fair enough. But it is interesting that the Fuji needed that extra half stop considering all other factors were the same. And in truth, the only thing I cared about being the same was the aperture because the test was "if I set both these cameras to f2 or f4 or whatever, what does the image the camera makes look like based on that?" Not a scientific test, but that's not what interests me. I want to know what the cameras are going to do in a given situation. I want to rely on the camera to do its job in the auto department when I set an aperture and am simply evaluating how well the X-T1 does that versus the GX7.

If the Fuji produces less noise because it overexposes compared to the GX7 then that's an important part of my consideration. I already know ahead of time the sensors perform differently and exhibit different characteristics but what I don't know yet is: when I put the X-T1 in a given situation and try to make an image, what happens? If it produces an image with less high ISO noise, I frankly don't care how it achieves it because what's most important to me usually is aperture; except for shooting sports, then it's shutter speed.

It is a well known and documented fact that the higher the ISO on the Fuji cameras, the "less accurate" to true ISO it is. I believe DXO Mark has documented this pretty well.
Some people cry fowl and curse Fuji from the rooftops.

I'm from the other camp and I don't care how you get great image quality from your cameras...just that the image quality is something that I can work with and create my finished product from.

 

[LovinTheEP2]

The Fuji has far greater resolution the panasonic and handles noise way better. You even said as such that you had to apply noise reduction to the m43 to sort of match the Fuji and in doing so, muddled the IQ as a direct result.

A longer longer lens with the same f number will always give a nicer bokeh so by needing to use a shorter length lens on a m43 to get sort dof equivalence always means a nicer bokeh on aps camera even though it's a smallish difference between m43and aps, both pail in comparison to a ff. crop sensors will by physics always be at a disadvantage in terms of bokeh but also have a direct advantage in faster shutter speed at a focal length depth of field equivalent so you can get ISO down 2 stops negating some of the noise advantage of ff.

Ie 50mm f4 dof and angle of view approx equiv on m43 is 25mm f2.... Nearly equivalent image but on m43 you have 2x light advantage with respect to exposure setting so... So if the 59mm was shot at ISO 1600 you can get equiv image at ISO 400 on m43..which by most test charts is pretty close noise wise and may even have a bit more dynamic range. Same applies to m43 and Fuji but only 1 stop advantage to m43s at equivalent exposure. So if aps sensors has 1 stop better noise vs m43, in real life that is zero if on m43s you can shoot 1 stop faster due to smaller sensor.. Fuji though is way more then 1 stop noise advantage....

 

[robbie36]

Fair enough. But it is interesting that the Fuji needed that extra half stop considering all other factors were the same. And in truth, the only thing I cared about being the same was the aperture because the test was "if I set both these cameras to f2 or f4 or whatever, what does the image the camera makes look like based on that?" Not a scientific test, but that's not what interests me. I want to know what the cameras are going to do in a given situation. I want to rely on the camera to do its job in the auto department when I set an aperture and am simply evaluating how well the X-T1 does that versus the GX7.

If the Fuji produces less noise because it overexposes compared to the GX7 then that's an important part of my consideration. I already know ahead of time the sensors perform differently and exhibit different characteristics but what I don't know yet is: when I put the X-T1 in a given situation and try to make an image, what happens? If it produces an image with less high ISO noise, I frankly don't care how it achieves it because what's most important to me usually is aperture; except for shooting sports, then it's shutter speed.

Well this depends. Some people believe (and are probably right) that a lot of Fujis 'good high iso' performance comes from how they are defining their ISO. In other words iso 3200 on a Fuji is actually iso 1600 on other cameras. I mean if in your example the Panasonic is 1/125 @f4 iso 3200 and the Fuji is 1/80 @F4 iso 3200, it follows you could have got exactly the same exposure from the GX&7 from settings of 1/60 F4 iso 1600. Then, of course, you would be comparing the GX7's iso 1600 to Fujis iso 3200 and it would be a fair comparison.

 

[sge998]

Having used plenty of m43 and Fuji cameras myself, I can conclude Fuji Xs' high ISOs are definitely inflated. I've done side by side comparisons using the GX7 against the Fuji X cameras in both controlled and outdoor situations. The results are that Fuji Xs often shoot at a lower shutter speed compared to the GX7 to compensate for their inflated ISOs (Not in low ISOs <640 though). The perceived sharper image from the Fuji Xs might be the result of a nonexistent anti-alising filter. High ISOs 2500+ on Fuji Xs often results in a pastel like color if you try to lift up too much shadow in post. The Fuji Xs have really good auto white balance though.

 

[Livnius]

The Fuji ISO inflation phenomena has been extensively tested by a number of people on the sister sites and it is universally agreed that there is in fact a degree of ISO inflation going on....although to what degree nobody can provide an absolute figure with absolute certainty, most however agree that it is there or there abouts a 1/3 of a stop. Taking that into account, the noise performance gap between m43 and FujiX does narrow, but it's still there.

 

[Fri13]

Noise.... The thing so many have not yet learn to love and praise....

There are bad noise = RGB color splatters
There are good noise = Film grain look.

Noise (good) can actually bring up details what otherwise would get hidden. It brings "the look" to the photography. And it can be removed from gradient areas and kept in detailed areas to boost them.

I didn't find Fuji X-T1 impressive one by quality. But I had it so short time that any skills learning to process it's photos I didn't learn. More important for studio is then lighting and post process where we can apply dozens of different denoise algorithms specifically to each part of photograph and then blend them together with layers. Result is _always_ better than one slider in Lightroom or manufacturer own software when using higher ISO. But task is half longer and twice as demanding.

What was more alarming for me was Fuji blowing highlights so easily unrecoverable state and dumping shadows as well. But that was before the "ultimate denoise" patch, so it can have been improved.

Subscribe to see EXIF info for this image (if available)

Subscribe to see EXIF info for this image (if available)

Photos from stevehuffphoto.com

 

[battleaxe]

Maybe it's because I own a older m4/3 camera(GF3) and don't care too much if there some ISO value inflation, but I really enjoy my X-E1. I sold my Sony SLT camera for it, and don't regret the decision(so far), though I do miss the lens options for wildlife(but I am back to MF like I was with my G2). I can easily say the white balance is way better than anything mf GF3 off or my Sony did. I have also noticed that some shot are like overexposed by 1/4-1/3 of a stop. However, this has a feature that my Sony and Panasonic lack, which is the ability to convert raw in camera, and it even gives some control over exposer when converting in camera.

For me it was not about the pictures(though Fuji colors don't hurt), but about bonding and having fun with the camera, which I am.