[2351HD]

For example, if you compare an A7rII and an A6300 (both new gen) I would fully expect the A7 to be better in every way. No doubts. Graphs or not. There is no denying it. The tech in the sensor has brought it far forward.

Anybody who tries to say M4/3 is not noisy at even base ISO is kidding themselves. Those sensors have grain in blue skies even at base ISO. They just don't perform as well, never have.

This scaling rubbish has to stop. Who scales a 43mp image to 8mp anyway.

[jsherman999]

For example, if you compare an A7rII and an A6300 (both new gen) I would fully expect the A7 to be better in every way. No doubts. Graphs or not. There is no denying it. The tech in the sensor has brought it far forward.

Anybody who tries to say M4/3 is not noisy at even base ISO is kidding themselves. Those sensors have grain in blue skies even at base ISO. They just don't perform as well, never have.

OK, I think they (new m43) generally perform 'well enough', but I assume my aps-c or FF bodies of similar gen should perform better. It's sounding like most people participating agree, and I'm going for a general consensus first.

This scaling rubbish has to stop. Who scales a 43mp image to 8mp anyway.

Well, every image you view or print is 'scaled' one way or another unless you always display or print at native resolution. For example, I think a 4K screen is at about 8MP. A printer driver or display driver does the scaling for you if you don't do it in PP first.

---------- Post added 03-15-16 at 01:12 PM ----------

(thought this black and white image might fit a little with your style, Jay)

Nice image as usual. One should never be under the impression that you need a bigger sensor than X to get pleasing images, as shots like this demonstrate. This is a very specific question for people who may want to know for example "will camera A give me better noise performance, because I'm reaching a limit with my current body and I'm curious about how to know what to do about it. Plus, I don't want to waste money if the perf increase isn't 'enough' for me."

[Rondec]

OK, I think they (new m43) generally perform 'well enough', but I assume my aps-c or FF bodies of similar gen should perform better. It's sounding like most people participating agree, and I'm going for a general consensus first.



Well, every image you view or print is 'scaled' one way or another unless you always display or print at native resolution. For example, I think a 4K screen is at about 8MP. A printer driver or display driver does the scaling for you if you don't do it in PP first.

---------- Post added 03-15-16 at 01:12 PM ----------



Nice image as usual. One should never be under the impression that you need a bigger sensor than X to get pleasing images, as shots like this demonstrate. This is a very specific question for people who may want to know for example "will camera A give me better noise performance, because I'm reaching a limit with my current body and I'm curious about how to know what to do about it. Plus, I don't want to waste money if the perf increase isn't 'enough' for me."

I agree and equivalence does answer the question "Why is my Q camera with an f1.9 lens so small in comparison to a K3 (or K-1 or whatever) with an f2 prime?" While the lenses are fast in terms of absolute aperture, in terms of actual performance, they are unlikely to measure up to a K3 and a DA limited (although more than adequate for many situations).

At the same time, few people actually push the limits of their gear to the point that the only answer is to choose a different sensor size. Of course, there are folks who shoot iso 12,800 all of the time or, are printing at 36 inches on a side, but those are actually the exceptions, not the rule. The biggest reason that someone should get a full frame camera is because they want it and can afford it. Why should someone need another reason? I think a lot of human thought is focused on rationalizations that turn our wants into needs.

The struggle with equivalence is real. One major issue with it is that it requires a similar generation Sony sensor. Pop in a Toshiba or Canon sensor and things don't work as well. Try to compare a K10 to a K5 and everything is totally weird.

It's even worse.

Many beginners interpret the restrictions that "newer" is "better". And they think "better" means "always better".
Newer generation can be worse. And it can be worse in some exposure settings and better in others.

This is why at base ISO old tech APSC K-5 II should expected be more than one stop worse than new tech FF Sony A7RII with regards to dynamic range.
"Equivalency" rule of thumb predicts the Sony will be better by 1,17 stops just for the larger sensor. And usually 4-5 years newer generation sensor mean about 0.3-0.5 stops technical improvement. So you would think the Sony outperforms the Pentax by 1,5-2 stops.
But in reality you have the old Pentax APSC outperform the Sony FF by 0.2 stops:
Sony A7R II vs Pentax K-5 II
So suddenly your equivalency rule of thumb is off reality for about 2 stops here (this is twice the amount people in forums call "a totally different league" ). And this only applies at base ISO. Change ISO, change results.

And yes, since about half the FF cameras out there are Canons, it doesn't work well in about half the real camera comparison cases anyway.

[crussellsprout]

Certainly, it seems difficult to have any reliable general rule in terms of sensor improvements over time. Even within Pentax's own offerings, it took 3 subsequent model generations until they came out with the K-5 whose sensor finally performed better than the K10D sensor -- the K20D and K-7 generally compared about the same or worse than the older K10D (Pentax K7 vs Pentax K20D vs Pentax K10D).

[jsherman999]

...
This is why at base ISO old tech APSC K-5 II should expected be more than one stop worse than new tech FF Sony A7RII with regards to dynamic range.
"Equivalency" rule of thumb predicts the Sony will be better by 1,17 stops just for the larger sensor. And usually 4-5 years newer generation sensor mean about 0.3-0.5 stops technical improvement. So you would think the Sony outperforms the Pentax by 1,5-2 stops.
But in reality you have the old Pentax APSC outperform the Sony FF by 0.2 stops:
Sony A7R II vs Pentax K-5 II

At base ISO, it's 14.1 vs 13.9 Evs. Pretty much tied, and the A7SII pulls away significantly after that.



This is also reflected in the score they publish on the 'front' of the comparison (13.9 vs. 14.1EVs) :



Now, none of this is bad news if you're a K5 landscape shooter who primarily shoots at base ISO - because 'upgrading' to a K7RII would not be much of an 'upgrade' for you. (Congrats, you just saved thousands.)

.

[2351HD]

I just want to get where you are coming from because I think I do. But on the A7rII you will get the excellent DR at 43mp, on the K5 you get it at 16mp. Does this not make the Sony a superior choice as you have a larger image in which to work from for printing etc?

I have printed large images from a 1DX and 645z and the Z files have more detail when printed large. The 1DX up-samples very very well, but the Z has it.

Are most people not printing?

[reh321]

Are most people not printing?

I used film cameras for over forty years, nearly always loading slide film into my camera. During that time I printed a total of two pictures larger than 4" x 6". In 1968, as a college junior, I had a shop make eight 8" x 10" black and white prints from one Instamatic negative; I was running for re-election as a member of the Student Senate, and these pictures of me on the Senate floor went on posters that I knew would be hung 8' off the floor in dining halls. In 1973, as a grad student, six female friends of mine asked me to take a picture of them in "sexy" hitchhiking mode; I had a place that heavily advertised on campus make a poster from the resulting slide and I don't recall anyone admiring the poster from a distance of less than three feet, because if you got any closer than that, you lost the overall effect.

---------- Post added 03-15-16 at 10:24 PM ----------

To the point: I'd guess more noise m43 (but not much, and only visible from a certain ISO upwards), almost indistinguishable result between FF "old" and "new" CMOS sensors.

The difference in generations is vital. I have told the following story several times already, but it bears repeating in this context. Last December, my wife and I attended a Madrigal Dinner at the college where she works. Unfortunately, we were a tad slow in reserving tickets, so our seats were in the second row. Not wanting to lug my K-30 with me, I took my tiny-sensor Q-7 instead. On the way out, I encountered a lady who had been sitting almost directly in front of me {so I was basically shooting around her all evening}, and asked her why she had used flash most of the evening, pointing out that flash washed out the warm "candle light" shades the crew had worked hard to achieve. She showed me the pictures from her five-year-old Nikon DSLR, and it clearly wasn't quite up to the task. When I showed her a picture from my Q-7, she immediately asked me to email some to her {the picture I had randomly chosen happened to be of her daughter's solo}. I would never claim equality between a current Nikon DSLR and a current Pentax Q-7, but those few years made an enormous difference in the camera's ability to deal with relatively dimly-lit scenes.

[jyndi]

... but isn't inherent noise determined mostly by two factors - the age of the sensor and pixel pitch? With processing by the camera to generate the RAW file being the first noise-filter?

That aside:

How many people think the micro four thirds (smaller than aps-c) Olympus OMD E-M5 likely produces as noisy images as the FF Canon 5DIII, or 5DS?

Does the K5 or Nikon D7000 likely produce the same noise in output as the D800?

I do not have a Canon 5DIII or 5DS, but I had a D7000 and I have an oly EPM2, one of the bottom tier m43 cameras that has the 'same' sensor as the top tier Oly m43 cameras. Quite frankly, I've had shots wrecked by noise from both cameras - the problem was always the idiot holding the camera(me), not the camera itself. I had also managed to take some shots with both that I really really like and which are not discernibly noisy when viewed in smallish prints or on screen. I've even got a couple from each which I have printed medium-ish without noise being noticed as an issue. That's the practical point that it really isn't an issue compared to the skill (or lack of it) of the photographer. However, without a doubt, the best shots in terms of noise levels were with my D7000. The two cameras are both nominally 16MP and I think the sensor tech is reasonably close. So the lower pixel density on the D7000 probably made the difference when the other factors (lighting, subject, etc) aligned to support it's somewhat better capabilities.

So yeah, all things being equal, a smaller sensor does give more noise. That is a reasonable expectation from a technical point of view. Why is that something to argue about?

(Well, aside from that common impulse to argue that our personal choices are the best choices period instead of just best for us. I mean, I decided the extra IQ I could sometimes extract from the D7000 to be insufficiently worthwhile to keep it when m43 came so close and was so. much. smaller. That's the compact vs. DSLR argument really. I do plan to get a K1 as I think that's enough of a performance leap to be worthwhile... for me.)

I do think too that the oly m43 cameras demonstrate just how much the processing tech can affect noise too. The OMD cameras - EM1/EM5/EM10 technically have the same sensor as my little EPM2. But they do produce cleaner images - because Olympus has refined their ability to extract data from that sensor read-out.

[derekkite]

I just want to get where you are coming from because I think I do. But on the A7rII you will get the excellent DR at 43mp, on the K5 you get it at 16mp. Does this not make the Sony a superior choice as you have a larger image in which to work from for printing etc?

I have printed large images from a 1DX and 645z and the Z files have more detail when printed large. The 1DX up-samples very very well, but the Z has it.

Are most people not printing?

I print up a calendar, this is my fourth year. 8x11. It is amazing how good the shots need to be to turn out nicely. I haven't checked but low iso or almost no crop slightly higher iso work ok. I did a large 24x36 shot of grizzly family swimming across the river and it turned out pretty good. It doesn't take many prints to see how much you need to improve your technique.



This was the first time I shot with the Sigma 500 f4.5 and K3. It was an amazing day, light was good and the subjects were unbelievable. We had a variety of equipment, one guy a 6D full frame Canon and 300 f4, a D7100 with 300 2.8. Most of the technique was not shaking with excitement, and we all got memorable shots.

[jsherman999]

I just want to get where you are coming from because I think I do. But on the A7rII you will get the excellent DR at 43mp, on the K5 you get it at 16mp. Does this not make the Sony a superior choice as you have a larger image in which to work from for printing etc?

It's tangential to where I'm going with this but yes, you can print or display larger with more MP to work from, but more MP also gives you some advantages even if you print at the same size - more data = more processing capability. Sometimes this just means re-sampling (printer driver or display driver scaling, downsizing) gives you a more crisp, 'sharper' looking image, sometimes it means you have more leverage when you do the PP yourself - I've found 36MP allows me to both sharpen and NR at the same time more than I could at 12MP before I notice artifacts.

Then there's cropping - 43MP means you can crop it quite a bit, display the crop fairly large, and still get very nice results from that. Wedding photogs loved the D700 for its low-light capability and AF, but they liked the 5DII more for it's cropability. I don't crop a heck of a lot, but I do do it, ymmv.

I have printed large images from a 1DX and 645z and the Z files have more detail when printed large. The 1DX up-samples very very well, but the Z has it.

Doesn't surprise me one bit, although it is lens-speed dependent too.

Are most people not printing?

I would say fewer and fewer are. I'm doing it less... not a good thing as far as I'm concerned though, I should do it more.

---------- Post added 03-16-16 at 12:16 PM ----------

... but isn't inherent noise determined mostly by two factors - the age of the sensor and pixel pitch? With processing by the camera to generate the RAW file being the first noise-filter?

Those factors, but also sensor area. For example two sensors cut from the same wafer, but one cut to aps-c and one to FF dimensions - the larger sensor will collect more total light at the same FOV and exposure, and it's resulting image will have about 1.3 stops less noise as a result.

...I do not have a Canon 5DIII or 5DS, but I had a D7000 and I have an oly EPM2, ...However, without a doubt, the best shots in terms of noise levels were with my D7000. The two cameras are both nominally 16MP and I think the sensor tech is reasonably close. So the lower pixel density on the D7000 probably made the difference ...

No - some or most of the difference there (assuming similar sensor gens, I'm not really familiar with the EPM2) may have been due to the D7000 having larger sensor area. Can't know what lenses you were using on each and what f-stops so I can't say for sure, but if you were shooting at the same FOV and exposure, the D7000's larger sensor kicks in.

So yeah, all things being equal, a smaller sensor does give more noise. That is a reasonable expectation from a technical point of view. Why is that something to argue about?

It shouldn't be.

---------- Post added 03-16-16 at 12:33 PM ----------

I used film cameras for over forty years.... In 1973, as a grad student, six female friends of mine asked me to take a picture of them in "sexy" hitchhiking mode;.

If you were shooting an aps-c camera at the time it would have only been 4.6 females.

(badumbum)

[jyndi]

Those factors, but also sensor area. For example two sensors cut from the same wafer, but one cut to aps-c and one to FF dimensions - the larger sensor will collect more total light at the same FOV and exposure, and it's resulting image will have about 1.3 stops less noise as a result.

That... does not make sense. What are you meaning here by exposure? In this context, I take it as the light volume registered per square mm of sensor - it's independent of total sensor area. But assuming the same light volume per square mm and the same iso setting, then the noise level per pixel will be the same. Each square mm of sensor doesn't know if it's part of a large sensor or a little sensor - it generates the same sensor noise regardless. (Although, if they're cut from the same wafer, the larger sensor has -more pixels-. So if you downsize the result to the same dimensions as the smaller sensor, sure, the larger sensor has less noise. You're losing noise with the downsampling (along with resolution). But if you're not downsizing, then the noise in each 100% crop is inherently the same, because it comes from the sensor.)

But... that doesn't mean you can get an identical image out of them, though, because in order to get the same FoV and ISO, aperture and consequently DoF must vary. If you want to hold the DoF and FoV constant, then ISO must vary (due to, as you say, different light volume) and on the 'cut from the same wafer' pair of sensors, that means they will have a different level of noise. But that is a result of a choice to vary ISO in order to duplicate FoV -and- DoF across the two sensors, it isn't that a smaller sensor cut from the same wafer suddenly has different noise characteristics by being cut smaller.

(Please tell me if that was coherent!)

Which comes down to... if you are trying to capture an image strictly defined by it's FoV and DoF, then the smaller sensor will need a higher iso setting. Higher iso = higher noise, if the sensors are 'same wafer' identical tech. Hence, larger has less noise. How much difference than makes in the the real world is mostly 'not much' and occasionally 'a lot'. Which side of that fence an individual's photography falls is made up of way more factors than the nice little exposure triangle has.

No - some or most of the difference there (assuming similar sensor gens, I'm not really familiar with the EPM2) may have been due to the D7000 having larger sensor area. Can't know what lenses you were using on each and what f-stops so I can't say for sure, but if you were shooting at the same FOV and exposure, the D7000's larger sensor kicks in.

This one is a case where the sensors are not from the same wafer. They both have the same 16 megapixels, but with the D7000 being larger, the pixels are larger. The difference isn't from the size of the sensor, but in the size of the pixels on the sensor.

[clackers]

Those factors, but also sensor area. For example two sensors cut from the same wafer, but one cut to aps-c and one to FF dimensions - the larger sensor will collect more total light at the same FOV and exposure, and it's resulting image will have about 1.3 stops less noise as a result.

Is there anyone still taking you seriously?


How cunning to misrepresent @Havelock, showing him the postprocessed Print tab of his comparison, instead of the camera's RAW file readings in the Screen tab. Let's redress that:






You can see the APS-C RAW files the K-5II puts on your SD card are just as clean as the newer tech FF's RAW files.


You can then remove noise when you do lossy postprocessing, but as was pointed out to you in the other thread it's got nothing to do with 'sensor area' or 'total light', and *only* to do with pixel numbers.


If anyone wants clarification, I invite them to PM me only ... I promise this will be my only post here.


Why? I think it was pretty poor form that with the other thread closed down, JSherman immediately set this one up to continue his earbashing. Sounds like subverting the moderators' action to me. :thumbdown:

[Fogel70]

Is there anyone still taking you seriously?

Yes. Most of us realize that you can quantify noise in in more than one way.
If you FI would compare noise per area size, I bet most smartphones have better noise ratio than most APS-C or FF cameras.

Instead of looking at graphs try examine images instead.
If you compare an image from D810 with K5 that capture the same scene using same exposure settings, how do you explain the difference you see?
Try checking images of high ISO for easily seeing a difference.

Try FI using imaging resource comparometer.
Imaging Resource "Comparometer" â„¢ Digital Camera Image Comparison Page

Dxomark:
Screen graphs = pixel comparison (pixel noise)
Print graphs = image comparison (image noise)

I believe most photographers are more interested in how much noise they perceive in the images they capture, vs how much measurable noise there is in each pixel.

[2351HD]

Ok so it's settled then. For all those who are happy to post the DXO graphs, you can all keep your K5ii bodies and all the rest of us will use K-1s and Sony A7rII bodies and we will never have to speak of this again.

We can all do without hearing how well your crop body stacks up to the industry best FF sensors. I am not sure what's trying to be proved, but when people want to spend money on new gear, why on earth we need to hear about how good 5 year old tech is, well it's beyond me.

I am happy to adopt the newest gear and parting with my hard-earned is an absolute pleasure, makes me feel great!