This is why you might want a big sensor

Enough with the agonizing. It's a great camera. People who don't mind the cost and the weight will have pics they can print on a bus. You like large prints. You know that you can sell it for most of its original price.

Honestly, Lou. The shutterbug philosopher who came up with the idea of 'good enough' did not have Lou Dobson in mind. You know you will not be happy until you at least take one for a spin. So do it already.

Raist3d said:

I think there are some technologies around that certainly push forward but still, today, the sensor size is still a big part of the influence in the expected performance.

Click to expand...

There are all kinds of issues being mixed up in the discussion from various folks. I am simply addressing the original post of this thread, which strongly implies that a 35mm format sensor is the key to being able to pull a usable image out of deep shadows at base ISO. This is false. The latest Sony APS-C sensors come close to the latest high performance 35mm format sensors in this regard.

Things which are true but with little relevance to the original post or the point that I am making:
-Bigger sensors usually have better color sensitivity at a given ISO
-Bigger sensors usually have greater tonal range at a given ISO
-Bigger sensors usually have better overall sensor image quality at a given ISO

-The image quality gulf between smaller and larger sensors tends to get larger at higher ISO values.

Here's another way to state what I am trying to say:

A 4/3 sensor implementing the same technology as the D7000 or NEX-7 sensor would have base ISO shadow recovery approaching (not equaling but close enough for many people) that demonstrated in the example Louis linked.

A 35mm format sensor implementing the same technology as the Panasonic GX1 or Canon 60D would not have base ISO shadow recovery approaching that demonstrated in the example Louis linked.

Bottom line: DR differences amongst current cameras are only weakly dependent on differences in sensor size.

--
http://aminsabet.com

Louis_Dobson said:

And although some people had all sorts of problems with the E3 AF for reasons I never established, mine has always been perfect. Right up there with the D3.

Click to expand...

If your E-3 was able to track moving subjects right up there with your D3 ... then I suspect your D3 had a problem!

Louis_Dobson said:

I suspect the f2 zooms have an AF problem, but I avoided those anyway.

Click to expand...

I don't think it was the lenses. I never touched an f/2 zoom, and AF was all over the place with my two E-3s. My main lenses were 14-54, 50/2, and 50-200 (both versions).

I wonder if you got that idea because you had me mixed up with the other whippet-shooting Julie, who did use a 35-100 on the E-3. She was happy with the S-AF (but now uses a D700 and VRII, for different reasons).

There were other 35-100 users with AF complaints, but I think it was just a case of it being easier to identify misfocus with a faster telephoto lens. Or the problems could have been worse with the 35-100 for some reason, but focus was unreliable with the slower lenses as well.

Julie

Well, exactly my point above! Get the part of the D800 sensor corresponding to m43 and you get a 11MP sensor with the same performance. Less (better quality) pixels is better.

Did I missed something here?

MT said:

I'm sure its been said before but I would not mind that a 1/4 size of that sensor technology in my MFT and take 8mp photos. I would be happy with that = and get far more DR and low noise.

MTMT

Click to expand...

--
Martin Ocando
-------------------------

Amin Sabet said:

Raist3d said:

I think there are some technologies around that certainly push forward but still, today, the sensor size is still a big part of the influence in the expected performance.

Click to expand...

There are all kinds of issues being mixed up in the discussion from various folks. I am simply addressing the original post of this thread, which strongly implies that a 35mm format sensor is the key to being able to pull a usable image out of deep shadows at base ISO. This is false. The latest Sony APS-C sensors come close to the latest high performance 35mm format sensors in this regard.

Click to expand...

Ok. I agree. In fact I would say for all practical purposes it's pretty much the same.

Amin Sabet said:

Things which are true but with little relevance to the original post or the point that I am making:
-Bigger sensors usually have better color sensitivity at a given ISO
-Bigger sensors usually have greater tonal range at a given ISO
-Bigger sensors usually have better overall sensor image quality at a given ISO

-The image quality gulf between smaller and larger sensors tends to get larger at higher ISO values.

Here's another way to state what I am trying to say:

A 4/3 sensor implementing the same technology as the D7000 or NEX-7 sensor would have base ISO shadow recovery approaching (not equaling but close enough for many people) that demonstrated in the example Louis linked.

A 35mm format sensor implementing the same technology as the Panasonic GX1 or Canon 60D would not have base ISO shadow recovery approaching that demonstrated in the example Louis linked.

Bottom line: DR differences amongst current cameras are only weakly dependent on differences in sensor size.

Click to expand...

Gotcha. We agree much more than we disagree.

Amin Sabet said:

--
http://aminsabet.com

Click to expand...

--

Raist3d/Ricardo (Photographer, software dev.)- "You are taking life too seriously if it bugs you in some way that a guy quotes himself in the .sig quote" - Ricardo

Martin Ocando said:

Well, exactly my point above! Get the part of the D800 sensor corresponding to m43 and you get a 11MP sensor with the same performance. Less (better quality) pixels is better.

Did I missed something here?

Click to expand...

Yes. It's early here, so excuse mental confusion, but:

Let's say you have a choice of four small pixels or one big one covering the same area.

Ad up the photons from the four small pixels, you get the same result as the one large pixel.

So at worst four small pixels will give you exactly the same result as one large pixel if you simply block downsample.

But in practice you can do better: look for odd man out pixels and ignore them, then average the others out.

And actually, fancy NR does rather better than that....

DR is the gap between saturation and the noise floor. So, if you have 11MP you get one result, have 44MP and you can can, at worst, sample and get the same (effectively 11MP) result, in practice with NR you will get a far better result. And when noise is not a problem you can use all 44MP for resolution.

More pixels is better in every way - for noise, or for res, or for a combo of the two.

Martin Ocando said:

MT said:

I'm sure its been said before but I would not mind that a 1/4 size of that sensor technology in my MFT and take 8mp photos. I would be happy with that = and get far more DR and low noise.

MTMT

Click to expand...

--
Martin Ocando
-------------------------

Click to expand...

--
http://www.flickr.com/photos/acam
http://thegentlemansnapper.blogspot.com

windsprite said:

Louis_Dobson said:

And although some people had all sorts of problems with the E3 AF for reasons I never established, mine has always been perfect. Right up there with the D3.

Click to expand...

If your E-3 was able to track moving subjects right up there with your D3 ... then I suspect your D3 had a problem!

Click to expand...

I had two D3s (one had crud on the sensor and went back). Both behaved the same. I found the C-AF with the 12-60 on the E3 fine. And the E3 S-AF fine and blisteringly quick. When I borrow it back off Eva, I still do. The trouble with the D3 AF is that all the points are crammed in the centre, so in practice it is a single point AF system, which I find a nuisance. Not a huge nuisance, but it was one of the things I liked least about the camera. The GH2, E-P1 and E3 all let me put the focus point wherever I like.

windsprite said:

windsprite said:

I suspect the f2 zooms have an AF problem, but I avoided those anyway.

Click to expand...

I don't think it was the lenses. I never touched an f/2 zoom, and AF was all over the place with my two E-3s. My main lenses were 14-54, 50/2, and 50-200 (both versions).

Click to expand...

Whereas the AF has been spot on with mind for four years. A mystery. Too late to solve it now though...

windsprite said:

I wonder if you got that idea because you had me mixed up with the other whippet-shooting Julie, who did use a 35-100 on the E-3. She was happy with the S-AF (but now uses a D700 and VRII, for different reasons).

There were other 35-100 users with AF complaints, but I think it was just a case of it being easier to identify misfocus with a faster telephoto lens. Or the problems could have been worse with the 35-100 for some reason, but focus was unreliable with the slower lenses as well.

Click to expand...

Lots of people have had E3 AF problems. I have no idea why they get them and I don't. But I think the 14-35 and possibly the 35-100 have AF problems of their own anyway . Nobody seemed to be able to get realiabble results with the 14-35 - dunno if it was ever fixed, I always regarded the F2 zooms as a stupid idea anyway. If you are putting up with the inherent disadvantages of a smaller sensor, why fit a truck on the front end?

Ironically FT could come into its own soon. The D800 shows that there is really no IQ purpose in going beyond 24MP because of diffraction, and we are rapidly reaching the point where you could have a 24MP FT sensor clean to ISO1600, which is as far as most rational people go. So then it is down to the sharpness of the glass, and the reality is those f2 zooms were way sharper than the iffy huge Canon and Nikon FF lenses (the vignetting and corner performance on my 70-200VR was diabolical). However, all the endless whining about noise killed FT, more's the pity. It was an excellent system.

windsprite said:

Julie

Click to expand...

--
http://www.flickr.com/photos/acam
http://thegentlemansnapper.blogspot.com

Amin Sabet said:

Raist3d said:

I think there are some technologies around that certainly push forward but still, today, the sensor size is still a big part of the influence in the expected performance.

Click to expand...

There are all kinds of issues being mixed up in the discussion from various folks. I am simply addressing the original post of this thread, which strongly implies that a 35mm format sensor is the key to being able to pull a usable image out of deep shadows at base ISO. This is false. The latest Sony APS-C sensors come close to the latest high performance 35mm format sensors in this regard.

Things which are true but with little relevance to the original post or the point that I am making:
-Bigger sensors usually have better color sensitivity at a given ISO
-Bigger sensors usually have greater tonal range at a given ISO
-Bigger sensors usually have better overall sensor image quality at a given ISO

Click to expand...

We are only talking base ISO here, where colour sensitivity, tonal range and DR are all way beyond what can be printed regardless of (sensible) sensor size or pixel pitch.

Amin Sabet said:

-The image quality gulf between smaller and larger sensors tends to get larger at higher ISO values.

Here's another way to state what I am trying to say:

A 4/3 sensor implementing the same technology as the D7000 or NEX-7 sensor would have base ISO shadow recovery approaching (not equaling but close enough for many people) that demonstrated in the example Louis linked.

Click to expand...

No it wouldn't. It would be a theoretical two stops off the same technology implemented in an FF camera. I picked the original shot simply for drama - I don't think most people realised it could be done. Completely black to nearly normal picture - that's not something you normally see demonstrated.

Amin Sabet said:

A 35mm format sensor implementing the same technology as the Panasonic GX1 or Canon 60D would not have base ISO shadow recovery approaching that demonstrated in the example Louis linked.

Click to expand...

No, the GX1 is, for my purposes, a hopeless sensor. Hence my concern that the OM-D may be lumbered with it and I shall have to move system AGAIN.

Amin Sabet said:

Bottom line: DR differences amongst current cameras are only weakly dependent on differences in sensor size.

Click to expand...

If you exclude the two back sliders, Canon and Panny at the low end, I disagree.

Amin Sabet said:

--
http://aminsabet.com

Click to expand...

--
http://www.flickr.com/photos/acam
http://thegentlemansnapper.blogspot.com

Louis_Dobson said:

We are only talking base ISO here

Click to expand...

Hence my clarification re: people discussing different things. You and I were talking base ISO.

Louis_Dobson said:

Louis_Dobson said:

A 4/3 sensor implementing the same technology as the D7000 or NEX-7 sensor would have base ISO shadow recovery approaching (not equaling but close enough for many people) that demonstrated in the example Louis linked.

Click to expand...

No it wouldn't. It would be a theoretical two stops off the same technology implemented in an FF camera.

Click to expand...

That is not correct. Two stops difference applies to shot noise. The noise dominating in shadows is read noise, which is not dependent on sensor size.

Louis_Dobson said:

I picked the original shot simply for drama - I don't think most people realised it could be done. Completely black to nearly normal picture - that's not something you normally see demonstrated.

Click to expand...

I've seen similar examples from the K5 and D7000.

--
http://aminsabet.com

Amin Sabet said:

Louis_Dobson said:

We are only talking base ISO here

Click to expand...

Hence my clarification re: people discussing different things. You and I were talking base ISO.

Amin Sabet said:

Amin Sabet said:

A 4/3 sensor implementing the same technology as the D7000 or NEX-7 sensor would have base ISO shadow recovery approaching (not equaling but close enough for many people) that demonstrated in the example Louis linked.

Click to expand...

No it wouldn't. It would be a theoretical two stops off the same technology implemented in an FF camera.

Click to expand...

That is not correct. Two stops difference applies to shot noise. The noise dominating in shadows is read noise, which is not dependent on sensor size.

Click to expand...

Yes, sorry, rushing out messages without thinking. However, assuming the same technology there is still a direct relationship. And you can see that fact reflected in the DxO numbers. As for your take that you can't assume the same technology - if high DR is a priority you are not going to pick a camera with an old tech sensor....

I do so hope Panny have not lumbered Oly with one!

Amin Sabet said:

Amin Sabet said:

I picked the original shot simply for drama - I don't think most people realised it could be done. Completely black to nearly normal picture - that's not something you normally see demonstrated.

Click to expand...

I've seen similar examples from the K5 and D7000.

--
http://aminsabet.com

Click to expand...

--
http://www.flickr.com/photos/acam
http://thegentlemansnapper.blogspot.com

Louis_Dobson said:

Amin Sabet said:

Louis_Dobson said:

We are only talking base ISO here

Click to expand...

Hence my clarification re: people discussing different things. You and I were talking base ISO.

Louis_Dobson said:

Louis_Dobson said:

A 4/3 sensor implementing the same technology as the D7000 or NEX-7 sensor would have base ISO shadow recovery approaching (not equaling but close enough for many people) that demonstrated in the example Louis linked.

Click to expand...

No it wouldn't. It would be a theoretical two stops off the same technology implemented in an FF camera.

Click to expand...

That is not correct. Two stops difference applies to shot noise. The noise dominating in shadows is read noise, which is not dependent on sensor size.

Click to expand...

Yes, sorry, rushing out messages without thinking. However, assuming the same technology there is still a direct relationship. And you can see that fact reflected in the DxO numbers. As for your take that you can't assume the same technology - if high DR is a priority you are not going to pick a camera with an old tech sensor....

Click to expand...

It depends on how we define "same technology". If same technology means that both saturation capacity and read noise scales with the pixel area, then a 16mp mFT sensor and a 16mp FF sensor will have (pretty much) the same DR on DxO. But, like I said above, then DxO's DR figures can be a bit misleading. If a 16mp mFT sensor and a 16mp FF sensor both have for example 12 stops of DR on DxO at iso100, then the FF sensor will in practice handle a 3 stops shadow lift better than the mFT sensor, because the FF sensor has less shot/photon noise, and that'll give it an SNR advantage if looking at the shadows e.g. 10 or 8 stops below full saturation.