Comparing Olympus 4/3lenses to FX "Full Frame" offerings

Started 11 months ago | Discussions thread
bobn2
Forum ProPosts: 32,312
Like?
Re: Wrong
In reply to Tiger1, 11 months ago

Tiger1 wrote:

bobn2 wrote:

That is a circular argument, since ISO is defined by exposure. Same ISO, same exposure but four times as much light on the FF sensor, four times as many photons, twice the SNR. It's that simple.

-- hide signature --

Bob

The surface area is four times bigger Bob so what each unit area "sees" is the same amount of photons regardless the format. Hence the SAME SNR.

You've got yourself confused. Lets go through it one step at a time.

1. For the same f-number the amount of light per unit area projected on the sensor is the same.

2. A FF sensor has four times the area as a FT sensor, so four times the 'unit'.

3. Thus a FF sensor collects four times the light at the same f-number.

4. When you view an image, you will do so at a size independent of the sensor size, the size you want to look at it. Imagine that we view in HD, approximately 2 megapixels. In the viewed image, each output pixel, one 2-millionth of the image represents the photons collected by one 2-millionth of the sensor.

5. Thus each pixel of the output image from FF sensor represents four times the number of photons as that made from a FT sensor, assuming the same exposure (f-number, shutter speed and scene luminance).

6. Hence the FF image will have double the SNR of the FT image.

The fact that smaller sensors tend to have smaller photosites so they can have the same number as larger sensors is what determines noise characteristics.

For what you say to be true, given that most noise in the image is photon shot noise, the smaller pixels would have to have a much lower quantum efficiency (collect a smaller proportion of the photons falling on them) than big pixels. This is not the case, they don't - in fact they typically have higher efficiency.

That is why their SNR is inferior.

Not so, I'm afraid.

Nothing to do with format. I'll even give you an example to prove this. Imagine you were able to take a Nikon Df chip and cut it into a 4/3 sensor and then put it into a 4/3 camera. The resulting image would only be 4MP but its noise and low light performance would be identical to the Df at 100% viewing of their pixels.

No-one in real life is interested in 100% viewing of their pixels. We are interested in what is the effect of a real photograph, and if you thing that a photograph taken from a FT size crop of a Df sensor and the resultant frame viewed same size as one taken from the full frame, would have the same SNR, you are mistaken, in fact it will have half the SNR.

Simple.

But wrong.

-- hide signature --

Bob

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
(unknown member)
(unknown member)
Oh?New
(unknown member)
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow