Cropping FF doesn't give sharper images

[Tronhard]

Ted... If you haven't done so, please read the links!!!

[xpatUSA]

Ted... If you haven't done so, please read the links!!!

Ignoring the triple exclamation marks, Trev, will you please clarify "higher signal / noise ratio, which means less better tonal range" namely: what is meant by "less better tonal range"?

Thank you,

[Tronhard]

Tonal range equates to dynamic range in this context - if you will, the different shades of one colour or the number of units of shade between white and black. They human eye can see about twice as many tones as the current crop of sensors, but they are always improving the technology.

So, smaller pixels have a smaller capacity to take in light - that is the signal. Each pixel is prone to interference both from the optics of the pixel lens and from the electronic circuit that serves the light sensor and converts light energy into electrical energy - that's the noise. When the the S/N ratio increased the quality of the resultant digital output is degraded and that reduces the dynamic range.

By having larger pixels there is more light (signal) compared to the noise, enhancing the quality of the output including DR.

There is a bit of a pixel race going on between different camera & sensor makers. It is often seen as better to have more pixels, but accepting the principles outlined above, shoving more pixels within the same area can only make them smaller, resulting in the degradation of the s/n ratio. To offset this designers seek tor reduce the sources of noise.

I just found an item in the CiC tutorial site the I had vaguely hoped was there:
https://www.cambridgeincolour.com/tu...ra-sensors.htm that and the next couple of screens may be illuminating, especially the section on Pixel Size: Noise Levels and Dynamic Range - you should also find that their English is a bit more elegant.

cheers:Trev

[dje]

Of course this assumes that resolution obtained from a specific pixel density is equal across camera models and brands. I am sure as technology advances we may find that this simple equation is not necessarily the case, but is a good basis to work on at present.

Hi Peter, this is a valid point you make.

A classic example would be if one camera had an AA filter and the other didn't. The one without the AA filter would probably be the sharpest in most cases. Also the pixel form factor and micro lens arrangement in each camera could have a bearing on the sharpness comparison.

Dave

[Tronhard]

Absolutely and that would be fine, depending upon what is being photographed. If there are repeated patterns then the will be moire issues. Still, if you are ok with the potential risk, big pixel-count FF sensors such as Nikon and Canon (3DSR) without AA filters, have used produce some incredibly sharp images- they are just MASSIVE files!

Hi Peter, this is a valid point you make.

A classic example would be if one camera had an AA filter and the other didn't. The one without the AA filter would probably be the sharpest in most cases. Also the pixel form factor and micro lens arrangement in each camera could have a bearing on the sharpness comparison.

Dave

[xpatUSA]

Hi Peter, this is a valid point you make.

A classic example would be if one camera had an AA filter and the other didn't. The one without the AA filter would probably be the sharpest in most cases. Also the pixel form factor and micro lens arrangement in each camera could have a bearing on the sharpness comparison.

Dave

Dave, Falk Lumo did an exhaustive analysis of the D800 vs the D800E about five years ago:

http://www.falklumo.com/lumolabs/art...0AAFilter.html

He knows a bit about sharpness too:

http://www.falklumo.com/lumolabs/art...ess/index.html

Both heavy going for "most of us" . . .

[xpatUSA]

So, smaller pixels have a smaller capacity to take in light - that is the signal. Each pixel is prone to interference both from the optics of the pixel lens and from the electronic circuit that serves the light sensor and converts light energy into electrical energy - that's the noise. When the S/N ratio [increases] the quality of the resultant digital output is degraded and that reduces the dynamic range.

Now it is clear, Trev, if I understand correctly: for best quality, I should expose the sensor as little as possible so as to reduce the SNR and thereby upgrade the resultant digital output and increase the DR.

You forgot shot noise but we can reduce that by collecting as few photons as possible which will also help to reduce the SNR.

Since most people have their own definition of DR, it seems, a few might be interested in THE definition of DR: ISO 15739.

Not free, but that standard is described quite well here, while the CiC tutorial barely scratches the surface of the matter and throws in some obfuscation to boot.

http://www.dougkerr.net/Pumpkin/arti...amic_range.pdf

Off I go to set 800 ISO on a shot, so as to get a good DR:



Oh dear . . . what happened?

[Tronhard]

Ted thank you for your patience and doggedness to make sense of all this!
To try to answer your question. A sensor will work at it best when it gets a good dose of light photons - i.e. lots of exposure. You will know that it is easier to get better images on a nice sunny day, or at least one with decent light, and it's harder as the subject is less well lit. Think of it like this... Your cost is the loss of signal within the circuitry which is a relative constant, or at least does not go up at the same rate as the increase in light to the sensor.

The ISO value is essentially a means by which the processors in your camera can "amplify" the signal from the sensor matrix to give you enough light to see and capture what you want. It amplifies the total signal including the noise. That is why you will get noisier images when the light is very low. A parallel experience is If you were to leave your stereo on and turn up the amplifier without any music you would hear more and more hiss - i.e. noise. I had experience with poorly recorded tapes when the music was not recorded at a high enough volume, or was too faint to record properly. To play it back I had to crank up the amplifier on the stereo and get a lot of tape hiss - which is why we call the visual pollution in our images noise. Because large sensors are likely to have less relative noise they can handle high ISO and low light situation better than very small sensors. Note that this varies between camera brands and models.

So with bright days you can get a lot of light to the senor grid, the "amplifier" in your system does not have to be turned up (think ISO can be low) and it's all peachy. You do of course have the other variables available to control your final result- exposure time defined by shutter speed and intensity of exposure, defined by aperture.

I said before that if I compare my tiny SX60HS camera with a FF one it actually behaves quite well when it gets lots of light, but as the light degrades the quality of the image degrades much faster than would the bigger sensor and pixel matrix of the FF. Also, as I mentioned, if you pixel peep the dynamic range of the smaller sensor is not as good as the FF sensor. That is one of the big efforts going on my camera manufacturers - to make reduce the noise factor.

Now how does that translate to taking the photo? Well in the way is the light meter and the set of algorithms that your camera manufacturer creates to give you what they consider the best from their configuration, then you have other variables like lens performance. So there is quite a bit going on between the sensor and your decision on how much exposure to given an image.

Looking at the file format that you use: If you shoot in RAW the general consensus is that you "expose to the right", i.e give as much exposure as possible without blowing your highlights. That is because your pixel data will contain information from the largest number of cells, including ones that might otherwise register as black. Note that fully while cells are blown out and essentially contain no data, so don't go overboard. The resultant image file will contain ALL the data that was captured in a relatively unprocessed or compressed format (hence RAW). You can then do a lot of manipulation over the image in post production.

When one shoots JPG it is arguably better to expose as per the meter's suggestion or even slightly tot he left to enhance the colour saturation. That is because the camera's algorithms will toss out quite a bit of the data as they process the image for compression - JPGS are much smaller but they have been processed and compressed by the camera, reducing your ability to do as much post production, although you can do some. Because this process irrevocably loses some data, it is called a Lossy compression.

Personally I shoot, capturing both RAW and JPG as it suits my type of photography and I have two versions of the same image.

Does that make any sense?

[xpatUSA]

Ted thank you for your patience and doggedness to make sense of all this!
To try to answer your question. A sensor will work at it best when it gets a good dose of light photons - i.e. lots of exposure. You will know that it is easier to get better images on a nice sunny day, or at least one with decent light, and it's harder as the subject is less well lit. Think of it like this... Your cost is the loss of signal within the circuitry which is a relative constant, or at least does not go up at the same rate as the increase in light to the sensor.

The ISO value is essentially a means by which the processors in your camera can "amplify" the signal from the sensor matrix to give you enough light to see and capture what you want. It amplifies the total signal including the noise. That is why you will get noisier images when the light is very low. A parallel experience is If you were to leave your stereo on and turn up the amplifier without any music you would hear more and more hiss - i.e. noise. I had experience with poorly recorded tapes when the music was not recorded at a high enough volume, or was too faint to record properly. To play it back I had to crank up the amplifier on the stereo and get a lot of tape hiss - which is why we call the visual pollution in our images noise. Because large sensors are likely to have less relative noise they can handle high ISO and low light situation better than very small sensors. Note that this varies between camera brands and models.

So with bright days you can get a lot of light to the senor grid, the "amplifier" in your system does not have to be turned up (think ISO can be low) and it's all peachy. You do of course have the other variables available to control your final result- exposure time defined by shutter speed and intensity of exposure, defined by aperture.

I said before that if I compare my tiny SX60HS camera with a FF one it actually behaves quite well when it gets lots of light, but as the light degrades the quality of the image degrades much faster than would the bigger sensor and pixel matrix of the FF. Also, as I mentioned, if you pixel peep the dynamic range of the smaller sensor is not as good as the FF sensor. That is one of the big efforts going on my camera manufacturers - to make reduce the noise factor.

Now how does that translate to taking the photo? Well in the way is the light meter and the set of algorithms that your camera manufacturer creates to give you what they consider the best from their configuration, then you have other variables like lens performance. So there is quite a bit going on between the sensor and your decision on how much exposure to given an image.

Looking at the file format that you use: If you shoot in RAW the general consensus is that you "expose to the right", i.e give as much exposure as possible without blowing your highlights. That is because your pixel data will contain information from the largest number of cells, including ones that might otherwise register as black. Note that fully while cells are blown out and essentially contain no data, so don't go overboard. The resultant image file will contain ALL the data that was captured in a relatively unprocessed or compressed format (hence RAW). You can then do a lot of manipulation over the image in post production.

When one shoots JPG it is arguably better to expose as per the meter's suggestion or even slightly tot he left to enhance the colour saturation. That is because the camera's algorithms will toss out quite a bit of the data as they process the image for compression - JPGS are much smaller but they have been processed and compressed by the camera, reducing your ability to do as much post production, although you can do some. Because this process irrevocably loses some data, it is called a Lossy compression.

Personally I shoot, capturing both RAW and JPG as it suits my type of photography and I have two versions of the same image.

Does that make any sense?

I'm a bit surprised by this response, Trev. Some time back you said:

"When the the S/N ratio increased the quality of the resultant digital output is degraded and that reduces the dynamic range."

That statement is just plain incorrect.

I responded somewhat sarcastically with some ridiculous statements and a silly example, but your response reads as if you took it literally and have assumed that I know next to nothing about the subject at hand.

The last thing I need today is a big lecture on how sensors work and how to take pictures, so perhaps we should just drop it?

[Tronhard]

I obviously mistyped that and apologize for doing so - I should have said either the noise/signal ratio increased or signal/noise ratio decreased.

Had I understood that you thought there was a reversal in my statement when you polity questioned that typo then you would have got the same response and apology as I just gave. You asked me to clarify that statement so I responded to clarify - I focused on your emphasized text and didn't twig on the typo itself.

On my last response I was responding having taken sleeping pills and never even noticed it was you, thus I treated your response as if you meant it, if you HAD meant it and I treated it as sarcastic that would have been even worse I suspect.