[bxf]

Whoa, BXF ... why do you need the same quantity of light when the frame is half the size?

That's why for exposure you don't have to think about sensor format or length of lens - shutter speed, ISO and lens aperture are the only three camera factors.

When a D800 shoots in cropped DX mode it won't be metering a stop lower or switching on an electronic ND filter!

The 'total light' thing is a furphy.

I'm afraid we may end up chasing our own tails with a question like this

My objective was not to prove that FF is better. Rather, I was (am?) trying to understand why FF consistently outperforms APS-c, even when the FF technology is older, when it comes to high ISO noise. Total Light seems to be the common argument, so I was trying to explain to myself what that could mean. It's not easy to dismiss the Total Light concept, since there have been some articles written on the subject by people who appear to be knowledgeable. DPR has also made a point of making reference to the concept.

BTW, great word "furphy": Australian slang for a rumour, or an erroneous or improbable story, but usually claimed to be absolute fact.

[clackers]

BTW, great word "furphy": Australian slang for a rumour, or an erroneous or improbable story, but usually claimed to be absolute fact.

Consider it a gift to the world, BXF, like penicillin, Google Maps, and Russell Crowe.

FFs beat APS-C in low light performance because the size of the receptors (pixel pitch) is different.

The 24 millionish pixels on an A7 are bigger than the corresponding ones on the K3.

Since the amount of random noise is about the same in each pixel (similar electric currents in the circuitry and heat from internal components), but there are more real photons falling on the bigger one, the SN ratio is higher.

Why not then put bigger but less pixels on a small sensor?

HTC did that, not trying to get into pixel wars with its phone rivals, but it seems people really do want more resolution.

As far as many articles being written about it, long lens 'compression' is another 'furphy'. Stevebrot and others here have proved cropping a wider lens does the same thing, but I'm sure web articles, textbooks and even experts who should know better will keep mentioning it for years to come!

[jsherman999]

Whoa, BXF ... why do you need the same quantity of light when the frame is half the size?

He's attempting to match total light and presenting the parameters needed to do that?

That's why for exposure you don't have to think about sensor format or length of lens - shutter speed, ISO and lens aperture are the only three camera factors.

And it's exactly why you get the same exposure at 24mm f/2.8 on FF and 24mm (equiv) f/2.8 on a point and shoot, and on a phone camera. That's not in question.

What hopefully is becoming obvious by now is that the P&S and phone camera, although they have the same exposure at "24mm" f/2.8, get much, much less Total Light.

The 'total light' thing is a furphy.

You Aussies are so cute.

---------- Post added 10-15-14 at 07:14 PM ----------

.... Rather, I was (am?) trying to understand why FF consistently outperforms APS-c, even when the FF technology is older, when it comes to high ISO noise. Total Light seems to be the common argument.

^^ Bingo, bingo, bingo.

Even when the technology is older and sometimes worse. Even when it has more pixels. Or less.

That crazy Total Light thing is happening!

[clackers]

You Aussies are so cute.

Although, JS, have you ever watched one of our football games?

You Tube

You Tube

[jsherman999]

FFs beat APS-C in low light performance because the size of the receptors (pixel pitch) is different.

How about when it isn't different? The K5/D7000 and D800 have exactly the same pixel size, yet the D800 has about a stop better performance.

"FF is better simply because it has bigger pixels" is a myth that's been debunked for a few years now.

As far as many articles being written about it, long lens 'compression' is another 'furphy'. Stevebrot and others here have proved cropping a wider lens does the same thing, but I'm sure web articles, textbooks and even experts who should know better will keep mentioning it for years to come!

Long lens 'compression' is an illusion. There are no physicists or engineers writing about that. The folks writing about compression (as a reality, not as a sometimes useful illusory effect) are probably the same folks still writing that FF is better because it has bigger pixels.

---------- Post added 10-15-14 at 07:23 PM ----------

Although, JS, have you ever watched one of our football games?


AFL Big Hits, Bumps, Tackles, Punches, Fights and Brawls [HD] - YouTube

Yes. Not as cute.

[clackers]

How about when it isn't different? The K5/D7000 and D800 have exactly the same pixel size, yet the D800 has about a stop better performance.

I'm guessing that others have already commented on your Captain Ahab-like obsession with FF, JS!





But it leads you to howling errors, such as when you said: ""Of course, you could just shoot with one format, too - the D810 - and just stop down 1.2. and then 2 stops to match the K-r and m43 DOF, and about match the noise and DR." and had to correct yourself.


This is another.


Again, if you're fixated by the light part of the signal-noise ratio, but not the equally important noise component, you're going to struggle to explain this:





The K5 you use as your example beats the Alpha 55 at low light performance with the same sensor, a 1.6 stop advantage if the quoter of this source is to be believed. (BTW, we're told as gospel that full frame means more dynamic range, but look at the K3 versus 5D MkII, so again, there are obviously more important factors).


It may be hard for someone posting again and again about sensor formats to notice, but there are a lot of other components in a camera.


The K3 beats the NEX-7 with the same sensor by a similar margin (I just checked at DXO - 1216 to 1016), presumably because the NEX-7 is so small. The components are closer together, and the heat sink will be much smaller.


The designers of the Sony A7S went for a smaller number of bigger pixels versus what's in the A7 and A7R. They fear the noise that 4k processing might bring to their cameras, so they're behind Panasonic. The A7S has to offload the job to an external recorder by HDMI.


Samsung's NX uses BSI to cut down on noise between pixels, the optical equivalent of crosstalk.


IIRC, Pentax have said that an issue with FF is that the distinctive SR mechanism would need to be larger i.e more heat.


The 645Z's ISO improvements over the D come from CMOS technology, not a change of frame size.

So, JS, you say one thing, the designers of cameras say another.


Who should we trust?

[ElJamoquio]

"Of course, you could just shoot with one format, too - the D810 - and just stop down 1.2. and then 2 stops to match the K-r and m43 DOF, and about match the noise and DR."

Yay! A good statement.

The K3 beats the NEX-7 with the same sensor by a similar margin (I just checked at DXO - 1216 to 1016), presumably because the NEX-7 is so small. The components are closer together, and the heat sink will be much smaller.

The K-5 is much smaller than the D7000 and the K-5 is better IIRC.

So, JS, you say one thing, the designers of cameras say another.

False choice on two levels.

[bxf]

Whoa, BXF ... why do you need the same quantity of light when the frame is half the size?
....

Would the following test provide any useful information? Shoot a scene of borderline lighting (i.e. dark enough to produce some noise) at a correct exposure, and also at +1 f-stop. Let's assume that the scene is such that the modest 1 stop overexposure will not cause any significant image deterioration.

If we were to reduce the exposure of the overexposed shot in PP, what differences in noise would we see between the two shots?

[clackers]

Yay! A good statement.

???

The FF shot ends up incorrectly two stops underexposed compared to the m43!

The exposure argument in that thread and the (Tony Northrup????) Youtube clip that inspired it that comes from is ... I'll find another word for furphy, I'm tired of it too ... 'bunkum'.

[clackers]

The K-5 is much smaller than the D7000 and the K-5 is better IIRC.

Hence, the sensor size and 'Total Light' cannot be the reason for the difference, EJ ... your example, remember!

[clackers]

Would the following test provide any useful information? Shoot a scene of borderline lighting (i.e. dark enough to produce some noise) at a correct exposure, and also at +1 f-stop.

No, BXF, if you get a K-3 and do this, there will be less speckling in the overexposed image - you've added more information than noise.

But it's true of any camera, any sensor, because you don't expose differently for DX vs FX modes, in Nikon speak.

Do you own a light meter, like a Sekonic?

Look at its controls. To correctly expose you don't consider sensor size.

[Fogel70]

Would the following test provide any useful information? Shoot a scene of borderline lighting (i.e. dark enough to produce some noise) at a correct exposure, and also at +1 f-stop. Let's assume that the scene is such that the modest 1 stop overexposure will not cause any significant image deterioration.

If we were to reduce the exposure of the overexposed shot in PP, what differences in noise would we see between the two shots?

There would be about the same difference in noise as if you set ISO to one stop lower instead of overexposing 1 stop. But as the images get different processing there might be some differences.

[Dartmoor Dave]

So, to summarise the "Total Light" hypothesis:

"Total Light" is the only factor responsible for noise in a digital camera.

All the other factors that can generate noise in the camera's electronics are irrelevant, because they contradict "Total Light".

Therefore, "Total Light" is the only factor responsible for noise in a digital camera.

[clackers]

So, to summarise the "Total Light" hypothesis:

I'd hope that none of the people in this thread believe in the Hard Version of that hypothesis, Dave. I don't think they do.

Some of them really seem to have an axe to grind about FF, though.

Myself, I may buy a K-1 or A7 one day, but it won't be for Bad Science reasons.

[Kunzite]

So, to summarise the "Total Light" hypothesis:

"Total Light" is the only factor responsible for noise in a digital camera.

All the other factors that can generate noise in the camera's electronics are irrelevant, because they contradict "Total Light".

Therefore, "Total Light" is the only factor responsible for noise in a digital camera.

And with a 1pt font, light gray on white: "only if it's the same technology"