Low Light: Full frame (ILCE-7) vs APS-C (ILCE-6000)

[benmao]

Advertisement (gone after registration)

Hello,

 

I have an Alpha 6000 (ILCE-6000) and I am very satisfied with that camera.

In recent time I used to make low light fotos several times. Seeing the anouncements for the A7II and the rumours about the A9, I would like to know, how much better a camera with the larger sensor would be regarding the pixel noise.

 

In theory, according to my understanding, the full frame sensor has 2,35times more surface area than the APS-C and thus should catch 2,35times more light. Is it true, that the A7 should therefore at ISO 235 have the same noise like the A6000 at ISO 100? Same number of pixel assumed.

 

I know that there are is also a difference in the depth of the sharpness, but I regard ISO and noise now.

 

Is my assumption correct? Are there any tests regarding this?

 

Ernst

[Advertisement]

Hi,

This might be interesting too: Sony Lenses

[Golem]

Your assumption is accurate when all other parameters

are ignored. IOW, your calculation is correct, but in real

world operations, practice will never follow theory. 

[Vivek]

Yes, correct.  Once you start using a FF, you will not go to your A6000 or any APS-C, however compact it might be.

[benmao]

in real

world operations, practice will never follow theory.

And how much is your real world experience different from this theory? What is your experience?

[BAM]

In theory, according to my understanding, the full frame sensor has 2,35times more surface area than the APS-C and thus should catch 2,35times more light. Is it true, that the A7 should therefore at ISO 235 have the same noise like the A6000 at ISO 100? Same number of pixel assumed.

 

 

 

I do low light photography and went for the Sony A7s because the 12Megapixels in the A7s sensor are spread over a greater area of the full frame sensor giving larger individual pixels, which gives much more light gathering capabilities and in turn a greater ability to shoot in very low light conditions with a better signal to noise ratio at higher ISO settings. Reviews stated at the time that ISO 6,400 on the A7s was the equivalent of IS0 1,600 on other cameras.

 

So far I haven't been disapointed in the image quality at low light. See my web link below

 

I just hope Sony push out an A7sII with 5axis stabilization.

 

Brian

[nippa]

Yes, correct.  Once you start using a FF, you will not go to your A6000 or any APS-C, however compact it might be.

 

There may be an exception to that.

My Fuji X100S's APS-C sized sensor ( and I guess other XTrans sensor cameras from Fuji ) knocks the spots off my A7 at ISO 6400 plus

The pocketable Fuji was always my first choice for low light but I had expected rather better from FF.

[Advertisement]

Hi,

This might be interesting too: Sony Cameras

[Golem]

Advertisement (gone after registration)

 

And how much is your real world experience different from

this theory? What is your experience?

 

And how much is your real world experience different from this

theory? What is your experience?  

   

Wellll ..... If sensor size is the solution, then the only reasonable

way to state that would include "all else being equal". All else is

seldom, perhaps never, equal. Noise-wise, my nex6 and nex7

are far superior to my Canon 5d2 [1/2 frame beats full frame]. 

Obviously "all else" is waaaaay far from equal since we have a

comparison across two different brands [of *approximately* the

same era of development in digital imaging].

  

Another classic example of "all else" not being even close to

equal, is that 35mm film images easily gather finer detail than

4x5" film images. Lens resolution falls as imaging circle grows,

it's just one of those engineering schneckens thaz pretty near

impossible to reverse. Also, 35mm gear is built [of necessity]

to much superior levels of precision [focus, alignment, film

plane, etc].  

   

Back to noise: If the bodies are of similar size, yet two bodies

are one of full frame and one of 1/2 frame, then, especially if

they are of similar construction [frinstintz an A7 and a nex7],

then it's quite likely that the 1/2 frame has better ability to

dissipate sensor heat. Yet, in violation of that concept, I found

the A65 waaaay noisier than the nex6 and nex7 .... but the

much more massive A65 should have better heat sink and

dissipation than the dinky little nex bodies. "All else ... etc" :-) 

   

Don't wanna go on here forever, but "all else being equal" is

never the case outside of controlled laboratories, so in the

real world, results and theory are only tenuously related. 

  

[Sinan]

There may be an exception to that.

My Fuji X100S's APS-C sized sensor ( and I guess other XTrans sensor cameras from Fuji ) knocks the spots off my A7 at ISO 6400 plus

The pocketable Fuji was always my first choice for low light but I had expected rather better from FF.

Try exposing the same scene both with A7 and X100 from the same point of view under same light with same shutter speed, aperture and focal length. You'll notice that A7 produces the same exposure in half of the Fuji's ISO. For instance if Fuji exposed it in 6400 ISO, A7 (or any other camera actually) will produce it in 3200 ISO.

 

That is why Fuji's high ISO values looks less noisy. They are not actually less nosiy they are just having inaccurate/non-standard ISO values.

[billzilla]

There may be an exception to that.

My Fuji X100S's APS-C sized sensor ( and I guess other XTrans sensor cameras from Fuji ) knocks the spots off my A7 at ISO 6400 plus

The pocketable Fuji was always my first choice for low light but I had expected rather better from FF.

 

Fuji X-Trans loses detail in RAW due to demosaicing conversion problems. JPEGs are also smoothed out, so what may look like better noise control at higher ISO is just a big fat loss of detail. So yes, I have no doubt it 'knocks the spots off' - literally, it's killing pixel level detail. There's absolutely no comparison to the A7/r/s cameras. None.

 

Fuji also cheats their ISO numbers and does not match up with measured values on other cameras. ISO-3200 on X-Trans is more like ISO-1600 on other cameras. If you don't believe it, look at DPR's comparisons. ISO-3200 on A7 vs. X-Pro1 should be exactly the same shutter at equal aperture, but the Fuji needs 1/1000 to get the same exposure as the Sony at 1/1600.

[Vivek]

There may be an exception to that.

My Fuji X100S's APS-C sized sensor ( and I guess other XTrans sensor cameras from Fuji ) knocks the spots off my A7 at ISO 6400 plus

The pocketable Fuji was always my first choice for low light but I had expected rather better from FF.

 

Couldn't help it at all.

[nippa]

Oh dear , perhaps OT but all I suggested was that XTrans APS-C sensors (  being without an AA filter ) compares to FF in low light.

 

Then the Fuji  myths appear again ; "sensor read ISO" is never straightforward and in one sense the statement about ISO relativity we know to be true but irrelevant as the onboard processing alters the effect .

Taking the same scene with both cameras and similar apertures X Trans sensor does very well so that there is no practical difference.

Equally it is not because the noise processing is aggressive although if you relied on jpegs or Lightroom 5 you might think so.

Using the correct raw converter like PhotoNinja/Iridient there are no de mosaicing issues and the X Trans sensor remains my choice sensor for low light use.

Of course I'm comparing to my plain old A7 not an A7S.

 

I'm not a Fuji fanboy and currently use Leica , Sony ,Olympus , Canon and Sigma cameras.

Now there's another APS-C sensor that can compete with my A7 but not in low light ; the Sigma DP3M

[Advertisement]

Hi,

This might be interesting too: Sony Lenses

[motamanx]

I am looking forward to a Sony FF a6000, this time without all those color "aids" inserted into the menu.  And, as I am fantasizing my dream camera: a fast, fixed focus 35mm lens.

Sounds as if I am talking RX1R?  I tried it, and wasn't impressed.  Too many frivolous bells and whistles for pro use.  

[billzilla]

Oh dear , perhaps OT but all I suggested was that XTrans APS-C sensors (  being without an AA filter ) compares to FF in low light.

 

Then the Fuji  myths appear again ; "sensor read ISO" is never straightforward and in one sense the statement about ISO relativity we know to be true but irrelevant as the onboard processing alters the effect .

Taking the same scene with both cameras and similar apertures X Trans sensor does very well so that there is no practical difference.

Equally it is not because the noise processing is aggressive although if you relied on jpegs or Lightroom 5 you might think so.

Using the correct raw converter like PhotoNinja/Iridient there are no de mosaicing issues and the X Trans sensor remains my choice sensor for low light use.

Of course I'm comparing to my plain old A7 not an A7S.

 

I'm not a Fuji fanboy and currently use Leica , Sony ,Olympus , Canon and Sigma cameras.

Now there's another APS-C sensor that can compete with my A7 but not in low light ; the Sigma DP3M

 

I've shot Fuji X100 and Sigma DP, Olympus and Canon myself as well.

 

DP3M compares at base ISO because it's a 3X Foveon sensor. You're getting comparable or better pixel level detail and great green color sensitivity (less interpolation), but in low light, that camera is useless for anything but tripod work. It's also slow and needs the Sigma X3F converter to DNG. Not optimal.

 

The XTrans is not a 'spot' on the the A7. If you have EXIF unretouched RAW images proving your 'knocks the spots off' claim, please post them. Otherwise, even with nonstandard RAW conversion the demosaicing is still an issue and APS-C cannot compete (ironically, greens are poor, the reverse of your DP3M). Fuji images are smooth, but smoothED and lack pixel detail. Even if per pixel noise was exactly the same with identical exposure (it is not), there's the matter of resolution equalization. Comparing 100% crops of 16MP vs. 24MP is apples/oranges.

 

You may prefer your X100S for various reasons, but its base IQ is not equal or better than FF.

[IronFilm]

I've been shooting with m4/3 mostly up until now, and the formula to get excellent low light is quite simple:

1) get a fast lens, such as the Nikon 50mm f/1.2

2) use a focal reducer, such as the RJ Lens Turbo which is what I use:
http://www.personal-view.com/talks/discussion/9086/rj-lens-turbo-m43-adapters/p1

I've just recently joined E mount, and I'll be doing the same process with my new A5100. And with a focal reducer on the A5100, not only will it DOUBLE the amount of light, but also I'll get a FF FoV too!

[jdronan@gmail.com]

Brian, went to your web site and you do beautiful low light work! My primary interest is low light, Astrophotography an d light painting. Until recently used the A77, just purchased the A77 ii. Lot of noise in the A77, but manageable. Want to rent the a7s to compare, b ut I am afraid I will not want to return it!

[Cheatah]

Many people complaining about noise of their cameras, but NO ONE ever makes such large prints, so that noise would disturbing the image. People are just playing games on their computer and monitors! These are just worth- and useless stories!

 

Folks, get out and take nice pictures! 

[AlexanderWaetzel]

@Cheatah good point! All you guys in here are mostly doing is math, what does it have to do with photography?

Noise is definitly a part of photography so deal with it! Im owning the a6000 and the a7 aswell and im not bothered with the noise perfomance at all!

If there is grain (and it is there, starting to get noticable at about ISO 2000 on both cameras) it looks alright and doesnt hurt the picture at all!

I mean sure if youre doing studio work or stuff like that you dont want noise, but then youre also shooting at ISOs where no modern camera will

let you down in noise perfomance. And if youre shooting in low-light a picture with no noise at all will look super artificial.

So in my opinion we photographers should stop talking about numbers and start talking about pictures! #

Please guys no offense, and if youre definitly need low-noise pics at high ISOs (above 6400) there are specialised cams out there (a7a, Nikon D4(s))

but in my opinion most ppl do not shoot above that professionally so that noise might be an issue!

[sharpshooterlove]

Everyone already knows that a better photographer will take a better photograph when using the same camera as a worse skilled photographer. Photography takes creativity and skill, yes, yes we know this. That doesn't mean that a racing car driver shouldn't be excited over new racing cars, on a racing car rumour site, does it? 

 

A digital camera is essentially a computer with a sensor to turn what you see into 1's and 0's. If you want to discuss what camera is better than another camera, you have to look at it objectively. It's no good concerning yourself with the art of photography to the point you forget what a modern camera is. If you want to accept a modern camera is perfect and the perfect tool for the job, that's fine but it's not good for the industry. If the sales keep rising and people are happy, companies don't have a reason to innovate. Isn't that the very reason a lot of people are switching from nikon to sony? 

 

 

 

NO ONE ever makes such large prints, so that noise would disturbing the image.

 

Wrong.

 

Fortunately artists don't complain when the latest wacom tablet has more levels of sensitivity or is more accurate. For some reason, photography is the only art I've known where people complain about objective thought. It's ironic really, because digital cameras are far closer to a modern day computer than the digital paint brush is 

[Advertisement]

Hi,

This might be interesting too: Sony Cameras

[Guest PenGun]

Many people complaining about noise of their cameras, but NO ONE ever makes such large prints, so that noise would disturbing the image. People are just playing games on their computer and monitors! These are just worth- and useless stories!

 

Folks, get out and take nice pictures! 

 Oh, so my 16x20 prints are made by no one. Cool, I was wondering how that worked.

 

 I bought an a7R because it has enough pixels to meet my printer's need for 360 DPI. I take my pictures from RAW just so I can produce the smoothest posible image. My 30" 2560x1600 IPS screen is nice but my 16x20s are sumptuous.

[Golem]

Many people complaining about noise of their cameras, but NO ONE ever makes such large prints, so that noise would disturbing the image. People are just playing games on their computer and monitors! These are just worth- and useless stories!

 

Folks, get out and take nice pictures! 

    

Apparently your concept of lowlight and high ISO noise is waaaay different that

some other folks, like frintintz me. It takes no pixel peeping whatsoever to see

the noise that bugs me. Heck, I can see it on my rear LCD monitor when I do a

quick check on my shooting situation. 

  

I'm not even super IQ fanatic. I do not require a facimile of reality in place of a

photograph. If my photo images look more like photo images rather than looking

like "reality under glass", I'm quite OK with that. But there are some types of high

ISO noise that actually conflict with the photo image ... they are like an overlay

of meaningless distracting crud, not like elements of the image that just happen

to carry some medium-related texture ... that latter "effect", I definitely prefer. 

[fotographer]

Fuji X-Trans loses detail in RAW due to demosaicing conversion problems. JPEGs are also smoothed out, so what may look like better noise control at higher ISO is just a big fat loss of detail. So yes, I have no doubt it 'knocks the spots off' - literally, it's killing pixel level detail. There's absolutely no comparison to the A7/r/s cameras. None.

 

Fuji also cheats their ISO numbers and does not match up with measured values on other cameras. ISO-3200 on X-Trans is more like ISO-1600 on other cameras. If you don't believe it, look at DPR's comparisons. ISO-3200 on A7 vs. X-Pro1 should be exactly the same shutter at equal aperture, but the Fuji needs 1/1000 to get the same exposure as the Sony at 1/1600.

I experienced this feature too, and my conclusion was, that the X-trans-Sensor needs more light to make the same photo.

[Cheatah]

 Oh, so my 16x20 prints are made by no one. Cool, I was wondering how that worked.

 

 I bought an a7R because it has enough pixels to meet my printer's need for 360 DPI. I take my pictures from RAW just so I can produce the smoothest posible image. My 30" 2560x1600 IPS screen is nice but my 16x20s are sumptuous.

Sorry, but 16 x 20" prints are just peanuts!