[LeeRunge]

I recently picked up a K-3 to upgrade from my K-5. I figured I'd post a few comparison RAW's with my other camera's for those who are curious.

I'm happy with the camera, it's very very close to the K-5 for low light but with a big boost in resolution. More importantly though it's much faster to operate and more refined in general, a great upgrade. And for 796 with the grip a steal.

Here's some RAW captures with the K-5, X100s and RX100 for comparison. Note the RX100 and X100s only go to ISO 6400 in RAW.

I used each camera's meter to 0 EV and ISO 6400. Some very different settings where picked. In general the K-3 and X100s are a much brighter exposure and used longer shutter settings while the RX100 and K-5 picked faster shutter and are a bit darker. Had I used a standard shutter they would have all been similar in exposure.

Overall noise for the K-5, K-3 and X100x are very very close. The noise looks worse on the K-3 at full magnification but great at normal crops. It's very much like a 24mp version of the K5. The X100s exposes much brighter with it's meter than any of the others. Had I used -.5 EV or so the shutter speeds would have fallen in line with the K-3 as would the exposure. The little RX100 is a great point and shoot as you can see from it's much smaller sensor.

I should add the JPEG engine on Fuji is far far better than either the K-3 or K-5 , I keep the Pentax's in RAW and the Fuji to Jpeg. Sony's JPEG engine is the worst however turning everything to that smooth plastic look if you get anywhere near a high ISO. All of them produce excellent low ISO shots with the K-3 and X100s being the best. Fuji honestly does have some mojo in the colors of those jpegs and you really need to work the Pentax RAW's to get them where you want it in post.

35mm equivalent or as close as I could get it was used for each. DA* 16-50 for the Pentax's.

K5 1/50 5.6
K-3 1/30 5.6
X100s 1/20 5.6
RX100 1/60 5.6


K-3 X100s



K-3 K-5



K-3 RX100

---

Last edited by LeeRunge; 11-30-2014 at 04:06 PM.

[Weevil]

Surprisingly, the x100s is not bad at all !

[LeeRunge]

Surprisingly, the x100s is not bad at all !

It produces excellent images. The meter goes for brighter settings though, note the 1/20th vs 1/30th for the K-3. The K-5 and RX100 went for 1/50th and 1/60th with darker results at 0EV.

I'm sure someone will chime in with Fuji fakes everything on the X100s but I haven't seem that. They are all very close in exposure if you use the same settings in manual. I'd say Fuji even has a *slight* edge in high ISO over both the K-5 and K-3. In real world shooting they are all about the same once you go through post processing.

Pentax camera's have a far better menu/control setup vs Fuji. Fuji gives me some "wtf" moments when I'm trying to find something.

[ScooterMaxi Jim]

Its an interesting comparison. It would be especially helpful to determine a setting yielding the best exposure (probably the longest one), and a particular white balance (probably incandescent). Obviously, when a sensor gets more light its noise tends to tame down. Even my Q in afternoon daylight looked pretty good one time when I forgot that I had left the ISO at 1250(!).

[LeeRunge]

Its an interesting comparison. It would be especially helpful to determine a setting yielding the best exposure (probably the longest one), and a particular white balance (probably incandescent). Obviously, when a sensor gets more light its noise tends to tame down. Even my Q in afternoon daylight looked pretty good one time when I forgot that I had left the ISO at 1250(!).

Light meter says 4.4 EV/ 52 LUX. 1/20th at 5.6 ISO 6400. Interesting that the Fuji picked the same exposure as the meter and the Pentax/Sony underexposed by a certain degree, that explains the darker exposure on the others.

4 EV is candlelight so not a lot used in those photo's. The K-3 got the most accurate white balance, the light is actually yellowish to the eye. The others made it looks white.

[Fogel70]

Light meter says 4.4 EV/ 52 LUX. 1/20th at 5.6 ISO 6400. Interesting that the Fuji picked the same exposure as the meter and the Pentax/Sony underexposed by a certain degree, that explains the darker exposure on the others.

It also explains why there are more noise in the darker exposed images.

Not surprisingly, the camera that receives most light have best noise performance.
But it's also makes it an unfair comparison, and it become a comparison of light meter more than noise performance,
X100s, recieves 50% more light than K3 and 150% more than K-5. which of course give an advantage to X100s. (unfair advantage if you compare noise)

Try using same exposure settings on all camera and you will get a much more accurate comparison of noise, and much closer in noise performance on all cameras.

[akshay869]

Speaking of noise performance, recently "thecamerastoretv" reviewed the new NX-1and also compared it against K-3 for noise on youtube. Check it out here.

https://youtube.com/watch?v=1flm65f2Gy8#

[LeeRunge]

It also explains why there are more noise in the darker exposed images.

Not surprisingly, the camera that receives most light have best noise performance.
But it's also makes it an unfair comparison, and it become a comparison of light meter more than noise performance,
X100s, recieves 50% more light than K3 and 150% more than K-5. which of course give an advantage to X100s. (unfair advantage if you compare noise)

Try using same exposure settings on all camera and you will get a much more accurate comparison of noise, and much closer in noise performance on all cameras.

I'll post some at what the light meter suggests. I don't think the noise pattern's will change much from using each of them, It's interesting how underexposed the meter's were for the K-5 and RX100 compared to the K-3 and X100s . In RAW Fuji/Pentax are very very close. In JPEG Fuji's engine is just much better at dealing with noise. I ran some other comparison's with the K-5 and Fuji and they are nearly the same (little smaller grain on Fuji) in RAW to iso 6400 (Fuji's max in RAW). I wonder how close we are to the physical limits of sensors and noise these days as the gap between APS-C and Full Frame is very close with some point and shoots making large leaps (RX100's).

---------- Post added 12-01-2014 at 09:51 AM ----------

Speaking of noise performance, recently "thecamerastoretv" reviewed the new NX-1and also compared it against K-3 for noise on youtube. Check it out here.

https://youtube.com/watch?v=1flm65f2Gy8#

That Samsung sensor seems to perform very well in ISO, hopefully the DR is impressive as well. Full Frames really don't have much advantage in ISO like they used too. It's just DOF difference these days.

Shame the AF is not great. I wonder if that's compared to a 7d ii or a K-3. Pretty impressive from Samsung anyways.

[LeeRunge]

updated with manual mode 1/20th 5.6 ISO 6400 RAW. K3 and K5 are using a DA* 16-50 at 22mm (tried for 23 to match the Fuji) RX100 is at 28mm as that's as wide as it goes.

K3 X100s


K3 K5



K3 RX100




I've got to say after comparing these camera's that the X100s produces the best images by a tiny margin in RAW. Fuji's Jpeg's blow away the K3 and K5 by a large margin. K3 and K5 are VERY close in iso performance. Honestly don't buy the K3 for the sensor, buy it for the other upgrades.

As you can see the the X100s isn't faking it's ISO at all. I'm actually a bit let down by the 24mp sensor in the K3 right now. I do however love all the other upgrades and it's worth it for those at the lower price they go for now.

I'm still amazed at how well that RX100 point and shoot does vs APS-C, that's one bad ass little sensor in there.

X100s get's the best sensor and Jpeg engine IMO
RX100's are amazing point and shoots.
K3 has a great WB meter. That light was yellowish and it captured it accurately.
K5 is still a great camera with a sensor that holds up all these years later. The K5ii without filter probably equals the X100s in RAW. (X100s is the same sensor as the XT-1)

[Winder]

That Samsung sensor seems to perform very well in ISO, hopefully the DR is impressive as well. Full Frames really don't have much advantage in ISO like they used too. It's just DOF difference these days.

If the same sensor technology is applied to a larger sensor, you will get better results. Sony and Canon expected to announce a new generation of sensors in 2015 before the show in Feb. It's simply a matter of is the current APS-C performance level good enough for your needs?

[normhead]

On my monitor, I'll take the K-3 image in every instance, but they all suck, what does shooting at 6400 show, except which is the best of the bad? YOU can clearly see looking at the images, that the K-3 and K-5 images have had some noise reduction applied.

A 3200 ISO image before Aperture had the K-3 image profiles...


Notice how the noise compares to your RX100 image. Absolutely no NR applied...

A 400 ISO image after the Aperture profiles...


The high ISO images are just bad images, regardless of the camera used. What I would be interested in seeing would the the result where an image from one camera was acceptable, and one from the other wasn't. Not two totally unacceptable images.

[awaldram]

updated with manual mode 1/20th 5.6 ISO 6400 RAW. K3 and K5 are using a DA* 16-50 at 22mm (tried for 23 to match the Fuji) RX100 is at 28mm as that's as wide as it goes.

K3 X100s


K3 K5



K3 RX100




I've got to say after comparing these camera's that the X100s produces the best images by a tiny margin in RAW. Fuji's Jpeg's blow away the K3 and K5 by a large margin. K3 and K5 are VERY close in iso performance. Honestly don't buy the K3 for the sensor, buy it for the other upgrades.

As you can see the the X100s isn't faking it's ISO at all. I'm actually a bit let down by the 24mp sensor in the K3 right now. I do however love all the other upgrades and it's worth it for those at the lower price they go for now.

I'm still amazed at how well that RX100 point and shoot does vs APS-C, that's one bad ass little sensor in there.

X100s get's the best sensor and Jpeg engine IMO
RX100's are amazing point and shoots.
K3 has a great WB meter. That light was yellowish and it captured it accurately.
K5 is still a great camera with a sensor that holds up all these years later. The K5ii without filter probably equals the X100s in RAW. (X100s is the same sensor as the XT-1)

You may wish to standardize on a pixel density else your favoring lower megapixel images

Either interpolate to 24 MP (not fair on lower MP cameras)
or
Bit bin to 12 mp

If all things are equal a 12mp camera with the same size sensor as a 24mp camera will show less noise but also less detail (less crop-ability)

You comparison shows this but as we already know that will be true doesn't show whether the higher resolution cameras are less noisy (@pixel density then the x100, you comparing something the x100 cannot do (16 and 24 mp images).

[LeeRunge]

Downsizing the 24mp to 16mp doesn't seem to make much difference. K3/5 and X100s are very close anyway in RAW. Fuji just has a small advantage unless your using Jpeg's.

One thing I'm noticing is you really need to hold this camera steady to get sharp images. This will probably get worse when the even higher (50mp) resolution camera's come out.

---------- Post added 12-03-2014 at 07:09 AM ----------

On my monitor, I'll take the K-3 image in every instance, but they all suck, what does shooting at 6400 show, except which is the best of the bad? YOU can clearly see looking at the images, that the K-3 and K-5 images have had some noise reduction applied.

A 3200 ISO image before Aperture had the K-3 image profiles...


Notice how the noise compares to your RX100 image. Absolutely no NR applied...

A 400 ISO image after the Aperture profiles...


The high ISO images are just bad images, regardless of the camera used. What I would be interested in seeing would the the result where an image from one camera was acceptable, and one from the other wasn't. Not two totally unacceptable images.

Norm I see what your saying. ISO 6400 (or really anything over 1600) should be avoided if possible. I tend to do a lot of shots in low light or in the dark hence my comparison. Really it's just a look at the outer limits of performance. DR is a much more important trait in a sensor and the K3 and K5 do very well in that regard, better than the X100s. Fuji in many ways saves itself with it's Jpeg production, its really good.

At low ISO levels the 24mp K3 should be better than the other options in RAW. Even the RX100 performs 90-95% of what the K5 does at low ISO (without the depth of field and less DR). Honestly I've seen comparisons of the K3 and Nikon D610 which look nearly the same at low ISO. For landscapes it really shouldn't matter if your using APS-C or FF. Birding seems to favor APS-C for the telephoto crop factor because that 400mm L glass is kinda pricey!

[normhead]

It sounds like you're the kind go guy who would like an D4s or the cheap guy's version, a Sony A7s... one being $6.5k and one being $2.5k. There are times in birding when it would be great to have 6400 ISO performance.. to achieve high shutter speed for freezing action... but these cameras are pricey. Also the 645z should be awesome, if you can find a long enough lens. But from what I understand from the wedding guys, there are more than a few wedding guys carrying a couple D4s cameras for everything but the formal images, where they may go to something more high res. There are options out there, they just aren't cheap. And for birders, the ability to crop is very helpful. They really don't have the perfect camera for us yet.

I probably would have bought an A7s, except, that for landscape I would want the A7r, and I'm waiting to see what Pentax comes out with as an FF, but, so I just live with the compromises my K-3 presents, although in a perfect world I'd own a D4s and a 645z. IN actual fact, I'd never carry them. And the odds of Pentax coming out with something high ISO are pretty slim. It's definitely an expensive specialty item in other brands.

[eddie1960]

updated with manual mode 1/20th 5.6 ISO 6400 RAW. K3 and K5 are using a DA* 16-50 at 22mm (tried for 23 to match the Fuji) RX100 is at 28mm as that's as wide as it goes.

K3 X100s


K3 K5



K3 RX100




I've got to say after comparing these camera's that the X100s produces the best images by a tiny margin in RAW. Fuji's Jpeg's blow away the K3 and K5 by a large margin. K3 and K5 are VERY close in iso performance. Honestly don't buy the K3 for the sensor, buy it for the other upgrades.

As you can see the the X100s isn't faking it's ISO at all. I'm actually a bit let down by the 24mp sensor in the K3 right now. I do however love all the other upgrades and it's worth it for those at the lower price they go for now.

I'm still amazed at how well that RX100 point and shoot does vs APS-C, that's one bad ass little sensor in there.

X100s get's the best sensor and Jpeg engine IMO
RX100's are amazing point and shoots.
K3 has a great WB meter. That light was yellowish and it captured it accurately.
K5 is still a great camera with a sensor that holds up all these years later. The K5ii without filter probably equals the X100s in RAW. (X100s is the same sensor as the XT-1)

as noted The XT1 Fujis will have similar performance - so will the XE2 I have both and they are really close to one another due to firmware updates. (I've been blown away by how flexible mine are in low light - of course 1.2 and 1.4 lenses don't hurt )
I've thought about the RX100iii as a Point shoot but lean towards the X100T or the new Pana LX100 - both a little more money but with better sensors