[clackers]

They might have told you that, but then they where wrong. It's the total area of the sensor that counts, not the size of one pixel.

Bizarre! :-D

That's why Sony produced the A7S - to get better noise performance than the A7R's 36Mp on the same area!

Signal-noise ratio is due to the photon capacity (ergo size) of the pixels.

The origin of the noise is the heat and random electricity from the circuits and rest of the body.

Sensor format size is irrelevant. The best SNR image would come from a single pixel covering the whole frame (high signal), with a huge body allowing lots of heat dissipation (low noise)..

---

Last edited by clackers; 02-06-2015 at 02:12 PM.

[normhead]

F4 on the FF is equivalent to F2.8 on the APS-C So a F4 70-200 (Which Canonikon also make and are like the 50-135 F2.8 on aps-c). But if you want F1.8 prime brightness like on the K-3 You'll need an F2.8. Then again put on a 1.4 prime and you will go blind .

When I use ƒ4 and ƒ2.8 the issue is shutter speed. What gets me to 1/1000 for small birds. IN that regard ƒ2.8 APS-c is exactly the same as 2.8 FF. When I'm locking focus in dark light, -3 EV Aps-c @ ƒ2.8 is exactly the same as ƒ2.8 FF. If I have a -3 K-3 an older -1 EV FF or camera like a D7100, My k-3 will lock focus with 2 EV less light than the -1 EV camera using the same lens and f-stop. There are many more scenarios where the above statement is wrong than there are where it's right. And your eye has an iris in it that adjusts to the level of light your eye is comfortable seeing. SO increasing brightness past an optimal point doesn't make the viewfinder brighter. Your eyes see it exactly the same way. You seem to be totally caught up in imaginary theoretical concepts, the true meaning of which you are blissfully ambivalent.

[derekkite]

I get that, but are you saying you cant have a "luscious" image With APS-c?

I wouldn't describe the K-3 as luscious. In some situations the K-5 had that quality, improved with the K5iis.

You can get nice shots with any body, but the 645z has something special.

[D1N0]

When I use ƒ4 and ƒ2.8 the issue is shutter speed. What gets me to 1/1000 for small birds. IN that regard ƒ2.8 APS-c is exactly the same as 2.8 FF. When I'm locking focus in dark light, -3 EV Aps-c @ ƒ2.8 is exactly the same as ƒ2.8 FF. If I have a -3 K-3 an older -1 EV FF or camera like a D7100, My k-3 will lock focus with 2 EV less light than the -1 EV camera using the same lens and f-stop. There are many more scenarios where the above statement is wrong than there are where it's right. And your eye has an iris in it that adjusts to the level of light your eye is comfortable seeing. SO increasing brightness past an optimal point doesn't make the viewfinder brighter. Your eyes see it exactly the same way. You seem to be totally caught up in imaginary theoretical concepts, the true meaning of which you are blissfully ambivalent.

F 2.8 FF is not exactly the same as F2.8 aps-c because you are factoring out ISO. There will be more noise on aps-c because FF has either bigger pixels -> less noise, or same size or more pixels -> relatively smaller noise. The larger surface will always translate in better IQ.

[clackers]

F 2.8 FF is not exactly the same as F2.8 aps-c because you are factoring out ISO.

D1N0, if the correct exposure on FF for a scene is 1/125s, ISO 100, f2.8, guess what?

The correct exposure on APS-C or m4/3 is 1/125s, ISO 100, f2.8.

Noise is something else, related to pixel size and body design, not sensor size.

[D1N0]

D1N0, if the correct exposure on FF for a scene is 1/125s, ISO 100, f2.8, guess what?

The correct exposure on APS-C or m4/3 is 1/125s, ISO 100, f2.8.

Noise is something else, related to pixel size and body design, not sensor size.

It is not something else. shutter time and aperture are not the only factors. Film or Sensor are also important. At the same correct shutter time you will get better iq with a better sensor. (and bigger is better because of physics). Since a sensor is part of the camera (film wasn't to a certain extent) you have to take it into account. It is not a two factor equation. Shutter speed, aperture, sensor size and pixel pitch are all in there. And then the type of pixel also matters.

[clackers]

sensor size

Prove it.

[robjmitchell]

D1N0
I think you are getting confused by terminology. An individual pixel is a sensor. Sensor as described in a camera is a sensory array.
Performance is determined by signal to noise. To increase signal you need to increase the size of photo receptor. At the same time you can reduce noise through improving electronics.
Doubling the area of your photoreceptor doubles the number of photons that will hit it thus doubling your input signal.

[Ben78]

People will inevitably compare this lens with the current Canikon offerings - It is priced smack bang in with them. Speaking of the Canon which has impeccable optics and built in stabilisation I believe that this Pentax lens will need to be absolutely stellar in performance - considering you are not getting in lens stabilisation.

I personally own a Canon 70-200 non IS f2.8 and it is so ridiculously good I kick myself for hanging on to Pentax for so long.

I hung onto enough Pentax FF lenses that I can easily go there again in the future - and seeing as Canon has decided that the masses want MP over DR in their latest 5D model I guess there will be reason to jump back to Pentax again in the future. What Pentax does with Sony sensors is a special thing.

[Gimbal]

Bizarre! :-D

That's why Sony produced the A7S - to get better noise performance than the A7R's 36Mp on the same area!

Signal-noise ratio is due to the photon capacity (ergo size) of the pixels.

The origin of the noise is the heat and random electricity from the circuits and rest of the body.

Sensor format size is irrelevant. The best SNR image would come from a single pixel covering the whole frame (high signal), with a huge body allowing lots of heat dissipation (low noise)..

"Sensor format size is irrelevant."
Now THAT is bizarre.

Ok pixel size does matter, but very very little compared to sensor size if one look at the resulting picture.

Pixel size matters if one compares one small pixel with one large pixel. But that is the wrong way to look at it since we usually use the entire sensor with all pixels together to make a picture. We should compare the noise in the end result, the whole picture,

Imagine two cameras, one with a crop sensor and one with a full frame sensor. Give them both the same pixel size (and a sensor from the same generation of technology). Both will have an equal noise level at each pixel, but the full frame sensor will of course have a lot more pixels since the sensor is larger.

Now compare two pictures at 100% zoom from each camera and they will be equally noisy. But then look at the entire pictures, either at a screen or printed to a given paper size and hey presto, the picture from the full frame camera will appear to have less noise.

So it's a matter of how you look at it. If you compare pictures, sensor size rules when it comes to noise level. Pixel size doesn't matter, it might only give you more detail if you choose to zoom into pixel level, but then you also get the noise. If you don't like the noise at pixel level you can always down sample to a lower resolution with less noise.

read all about it at DXO mark
Contrary to conventional wisdom, higher resolution actually compensates for noise - DxOMark

[totsmuyco]

Thanks for the explanation Gimbal. Simple and direct.

[Nicolas06]

Prove it.

Simple:

D610 and D810 have same sensor size and not the same pixel density: their score are similar in DxO and both are FF.
K5II and D810 have the same pixel density and different sensor size. The high iso score of D810 is 2.4X the high iso score of K5II while the sensor surface of D810 is 2.3X the surface of K5II. Do you really think that just it is just by random coincidence that a 2.3X sensor size give rougly a 2.3X high iso performance?

645Z pixel density match a 30MP FF... Why would any one buy a 645Z then if any FF except D810 and A7R would have better performance due to lower pixel density? Do you really think that a D610 would get better results than 645Z due to lower pixel density ?

Pixel density has a small influence... After all A7S manage to get 1/3 stop better high iso performance than D810 with 3 time less pixels and also having a new design to gather more light for the photosite. Not even all of this 1/3EV gain comes from the reduced pixel density. But for A7S at least it come with a cost: dynamic range and color deph that is no better than a K5-II meaning that at low iso you get more resolution and higher quality picture than A7S.

With the same sensor technology, the sensor size is much more important than the pixel density, particulary for high iso performance.

I would agree that pixel density does still count. But that when we speak of very small sensors that has a much higher pixel density. The pixels are so small that there no possibility to correctly register the color deph and dynamic range of a scene. This show in the picture theses camera take.

As for what 645Z (or before 645D) offer, to me the difference comes more on what a bigger sensor permit on lenses design and the rendering you'll get out of it than anything related to pixel density.

[Rondec]

Prove it.

Everything assumes similar technology. But most full frame sensors have 1 stop better SNR at all isos than the equivalent APS-C cameras, which equates to 1 stop better control of noise. Dynamic range is a little more iffy, because cameras like the K5 actually do better at low iso than Canon full frames do at low iso, but once you get above 800 iso, there is a clear full frame advantage in dynamic range as well.

DXO Mark has a screen button which shows measure per pixel SNR, dynamic range etc and this is usually pretty much the same between full frame and APS-C . The D800 and K5 curves lie right over top of each other. However, that isn't particularly useful. The Print button shows you an 8 megapixel normalized image and that is where full frames jump in front of crop frame cameras.

[Nicolas06]

D1N0, if the correct exposure on FF for a scene is 1/125s, ISO 100, f2.8, guess what?

The correct exposure on APS-C or m4/3 is 1/125s, ISO 100, f2.8.

Noise is something else, related to pixel size and body design, not sensor size.

So you get the same exposure on FF and APSC or MF or m4/3 (or a compact) with 1/125s, iso 100, f/2.8. Yes

But there an infinite number of combination with different apperture iso, shutter speed that would do the same a few example

- f/1.4, 1/2000, iso 400
- f/5.6 1/125s, iso 400
- f/16, 30s, iso 50 + ND64

I expect that one doesn't choose theses values by random but for a reason.

This suddenly mean that all theses setting are important. Apperture is for deph of field control, shutter speed is for movement control... And sensitivity is totaly abstract number that affect the noise level.

So I would hope somebody to choose the apperture and shutter speed for good reason and let the iso adapt, couting the noise is not too high.

The effect of apperture on a scene depend a lot of the focus distance and sensor size. I would expect somebody to choose different appertures on m4/3 and MF for getting the same results.

Some details like the bigger sensor perform noticably better at high iso and that lenses tend to perform noticably better stopped down a bit and that bigger sensor can keep acceptable sharpness with lenses more stopped down means:
- For long exposure bigger sensor can get less light while keeping diffraction low and get longer exposure with same ND filter.
- For shallow deph of field bigger sensor can use smaller apperture and thus higher picture quality for same deph of field.
- When shooting in low light bigger sensor can keep lower noise level at the expense of deph of field or get the same deph of field and same noise.

Of course all of this come with the issue of having quite bigger camera/lenses, higher price and difficultlies to achieve wide deph of field in some extreme conditions.

For me APSC is a great compromize. FF could with small primes. m4/3 I never tried, but why not? For now I'am invested in Pentax APSC and I'am happy with it.

[filoxophy]

Please, for the love of all that is beautiful, can we not start this here?