[redrockcoulee]

What do you think of that meter? I got one in a purchase of some gear but haven't used it. The interface is not the same as the Pentax where you can see all your exposure combinations at a glance as well as count stops easier when you slap on things like a 9 or 10 stop ND filter.

I post on that LF forum. Is your ID the same as here?

It is alright but nothing special. The main reason I use it is I was able to borrow it 6 years ago and not sure if I ever have to return it as the owner bought a better meter for flash.I never used it for ND calculations, that is why I have a iPOd Although the truth is my wife uses it for pinholes when we are out shooting. Yes I am the same handle on LF and APUG and have noticed you posting on LF before. I am not very active on any of them.

[wombat2go]

Here, I gave this example earlier - please read it carefully and consider it::

Two formats, let's assume same sensor efficienncy/generation - same FOV and F-stop and shutter speed (same exposure) :

FF: 70mm f/2.8 = 70 / 2.8 = 25mm physical aperture
m43: 35mm f/2.8 = 35 / 2.8 = 12.5mm physical aperture

So even though they have the same exposure (light density,) the FF image would have more Total Light due to twice the physical aperture (25mm vs 12.5mm) used to get the same FOV, and thus it would have two stops better SNR, better DR - and also two stops less DOF. This is precisely how FF get's it's noise advantage.

I don't think this is correct at all
If you "SNR" means signal to noise ratio, it is intrinsic the each sensor well, not the overall sensor.
It is not related to the array size or the quantity of such wells, or the "Total Light" on the overall sensor.

Signal to noise ratio is related to the integrated charge on the photodiode that is coupled to the cmos gate.

Of course, signal to noise ratio varies with type and size of sensor

If your "Total Light" means the integral over the sensor array of luminous flux per solid angle arriving from the lens,
that is not related to signal to noise ratio in each well either.

I never read about such "Total Light" I don't think it is germane to sensor performance, it is not related to the integrated charge on each well.

The integrated charge on each well is related to the photons arriving at the photodiode which is proportional to the varying luminous flux at each spatial well position, which forms the image.
It can be easily seen that dark areas on a sensor have a different (and worse) signal to noise ratio than the bright areas.

[Dartmoor Dave]

People are still arguing about this? Folks, this stuff was already completely well understood in the late 19th Century.

In direct sunlight, using a film stock or digital sensor with an ISO/ASA rating of 100, a setting of f/16 at 1/125 second will give you exactly the same exposure whether you're using a large format view camera, or a medium format film camera, or a "full frame" DSLR, or an APS-C camera.

A larger film format or digital sensor will give you less film grain or digital noise, and of course the angle of view will change using the same focal length on different formats.

19th Century photographic pioneers such as Julia Margaret Cameron and Mathew Brady would have been able to join in this discussion with a complete understanding of the issues involved. But they would have been too busy making great photographs to bother.

[jsherman999]

When I say technology in the sensor, I mean the capability to put the same number of pixels in a smaller sensor, with similar noise. It's silly to compare a 16mp APS-C with a 36mp FF. But if you have two 24mp sensors, one is APS-C and the other is full frame, then that's where the technology that I'm talking about comes into play.

No. Current example: 24MP D600 vs, 24MP D7100. Same sensor generation, very minor differences in QE - D600 has a stop better SNR performace do to Total Light the sensor is capturing at the same exposure:

[ChristianRock]

No. Current example: 24MP D600 vs, 24MP D7100. Same sensor generation, very minor differences in QE - D600 has a stop better SNR performace do to Total Light the sensor is capturing at the same exposure:

Going back several pages, what I have been saying is that just like the dynamic range gap was shortened, and is now not that different, S/N ratio can also close the gap due to advancements in technology.

Also, you're comparing a Toshiba sensor with a Sony sensor. Not same technology.

[jsherman999]

I don't think this is correct at all
If you "SNR" means signal to noise ratio, it is intrinsic the each sensor well, not the overall sensor.
It is not related to the array size or the quantity of such wells, or the "Total Light" on the overall sensor.

Signal to noise ratio is related to the integrated charge on the photodiode that is coupled to the cmos gate.

Of course, signal to noise ratio varies with type and size of sensor

If your "Total Light" means the integral over the sensor array of luminous flux per solid angle arriving from the lens,
that is not related to signal to noise ratio in each well either.

I never read about such "Total Light" I don't think it is germane to sensor performance, it is not related to the integrated charge on each well.

The integrated charge on each well is related to the photons arriving at the photodiode which is proportional to the varying luminous flux at each spatial well position, which forms the image.
It can be easily seen that dark areas on a sensor have a different (and worse) signal to noise ratio than the bright areas.

I'm using "SNR" loosely, yes, because it's a quick acronym to type and people know what I mean. What I mean is overall image shot noise as determined by the radiant flux ("total light" = less off-putting,) not the per-pixel performance.

.

---------- Post added 10-10-14 at 01:29 PM ----------

The interesting thing to me about that graph is the serious fudging going on in the D800e's stated iso versus measured iso. In some cases, it is off by half a stop as you go up the curve. Makes a difference in comparisons, in my opinion.

How about the K-5 'smoothed' bump starting at ISO 3000?

---------- Post added 10-10-14 at 01:32 PM ----------

Going back several pages, what I have been saying is that just like the dynamic range gap was shortened, and is now not that different, S/N ratio can also close the gap due to advancements in technology.

Also, you're comparing a Toshiba sensor with a Sony sensor. Not same technology.

Then compare a straight Sony 24MP sensor, they're on DXO. And please carefully read this, where I made an effort to explain why Total Light is so important to the equation.

[Rondec]

I'm using "SNR" loosely, yes, because it's a quick acronym to type and people know what I mean. What I mean is overall image shot noise as determined by the radiant flux ("total light" = less off-putting,) not the per-pixel performance.

.

---------- Post added 10-10-14 at 01:29 PM ----------



How about the K-5 'smoothed' bump starting at ISO 3000?

---------- Post added 10-10-14 at 01:32 PM ----------



Then compare a straight Sony 24MP sensor, they're on DXO. And please carefully read this, where I made an effort to explain why Total Light is so important to the equation.

For what it's worth, the K3 doesn't have the "smoothed" section and SNR is the same between it and the K5/K5 II.

[jsherman999]

For what it's worth, the K3 doesn't have the "smoothed" section and SNR is the same between it and the K5/K5 II.

Newer sensor = less need to 'smooth'

---------- Post added 10-10-14 at 02:09 PM ----------

People are still arguing about this? Folks, this stuff was already completely well understood in the late 19th Century.

In direct sunlight, using a film stock or digital sensor with an ISO/ASA rating of 100, a setting of f/16 at 1/125 second will give you exactly the same exposure whether you're using a large format view camera, or a medium format film camera, or a "full frame" DSLR, or an APS-C camera.

Where in this thread do you see the definition of Exposure questioned?

A larger film format or digital sensor will give you less film grain or digital noise,

Follow through with the thought - why would that ^^ be?


.

[dtmateojr]

After reading some of the responses here it seems like I'm not the only one who thinks that equivalence is nonsense afterall
The truth prevails

[Kunzite]

Because with the larger film format you'd have more of the same emulsion (the claim doesn't necessarily stand otherwise). So you'd have more of the same "grain per square mm" surface. It's obvious.
This is a nice, consistent way of thinking which can cope with everything you'd throw at it; because you know both factors (surface and "grain per square mm"), instead of combining them or worse, adding the lens into the equation.

[dtmateojr]

Newer sensor = less need to 'smooth'

---------- Post added 10-10-14 at 02:09 PM ----------





Follow through with the thought - why would that ^^ be?


.

You don't question exposure but you keep pushing total light as if it is independent from total sensor area. You keep claiming that as long as total captured light is the same then noise will be the same.

Photography has always been total light over total area. Smaller sensors need lesser light. Exposure is maintained. Noise is maintained. It's very basic.

Film grain does not change from one format to another. It's the same emulsion. The only time it will "change" is when you follow the prescriptions of equivalence-fu and that's by underexposing the larger format proportional to crop factor and then pushing in development. This would actually result in the larger format looking much much worse than the properly exposed smaller format.

---------- Post added 10-11-14 at 06:37 ----------

Because with the larger film format you'd have more of the same emulsion (the claim doesn't necessarily stand otherwise). So you'd have more of the same "grain per square mm" surface. It's obvious.
This is a nice, consistent way of thinking which can cope with everything you'd throw at it; because you know both factors (surface and "grain per square mm"), instead of combining them or worse, adding the lens into the equation.

But if you don't have the same emulsion, all bets are off.

Grain actually remains the same. What sherman wants you to say is that if you print at the same size then larger format results in lesser "apparent" grain. Note though that this is print grain NOT film grain. So sherman is still very wrong.

[wombat2go]

I'm using "SNR" loosely, yes, because it's a quick acronym to type and people know what I mean. What I mean is overall image shot noise as determined by the radiant flux ("total light" = less off-putting,) not the per-pixel performance.

Does that mean your model is a of "Uniform Gray Wall" that completely fills the 2 sensors you are comparing?
When calculating snr to get your assertions, did did you assume both sensors had identical wells with the same noise floor?
Did you consider CosFourth law for brightness distribution on the 2 different sized planes?
Do you have any references to this method that you can put up?
Thanks

[Dartmoor Dave]

Where in this thread do you see the definition of Exposure questioned?

It was questioned in the video linked in the original post, which contained the asinine assertion that f-stop values are more about depth of field than exposure.

Follow through with the thought - why would that ^^ be?

It's because of many factors including pixel size and density, capacitance, analog-domain amplification, and A/D conversion. Nothing at all to do with this specious notion of Total Light that you keep capitalising as if that makes it real.

[Kunzite]

Size is a factor, an important one. But to see it's not the only significant factor, it suffices to look at same format, different sensor cameras. Let's take the A7 family - all MY 2013-2014:
Sony A7S versus Sony A7 versus Sony A7R - Side by side camera comparison - DxOMark

The A7S is soundly beating the other two in "Low-light ISO" with a score of 3702 (it also has the lowest DR and color depth, and of course the lowest resolution - so don't hurry to declare it an overall winner).
On the second place, showing that more pixels are not necessarily worse, it's the A7R.

In the end, we can't avoid measuring the actual results.

[dtmateojr]

It is very easy to debunk equivalence with very simple math:

Assume that noise is only affected by total light. Now current sensors have a dynamic range of about 13 stops max. For the sake of example let's use 15 stops of DR.

It follows that if you have a very large sensor such that the smaller sensor being compared to has a crop factor of 15, equivalence is telling us that if we underexpose the larger sensor and bump the ISO by 15 stops then they would result in the same image. Common sense will tell you that this simply won't work. The extreme underexposure will result in the larger sensor not recording any data at all.

We don't even need a crop factor of 15. Even a crop factor of 8 will debunk equivalence because you do not have to cover the upper half of the entire DR (that's why some people insist on "expose to the right").

Corollary to that, if you can find a larger sensor of crop factor N where N is an insanely big number, equivalence will result in that sensor not capturing a single photon of light as a result of underexposure.

QED.