Max ISO (Page 16): It might seem that way but that's not the whole story. A camera's base ISO is the ISO setting at which the entire physical capacity (of the analog side of the...

selmslie Loc: Fernandina Beach, FL, USA

anotherview wrote:

I understand the ISO value as fixed according to the engineering of the camera sensor. ...

It might seem that way but that's not the whole story.

A camera's base ISO is the ISO setting at which the entire physical capacity (of the analog side of the process 100% of it) can used to reliably record an image.

If you shoot at 2x the base ISO you will only be able to use 50% of the analog signal, at 4x the base ISO only 25% of the analog signal, etc. To get the same brightness in the image that you recorded at base ISO you will need cut the exposure by 1/2 stop, 1/4 stop, etc.

So what you are giving up by using a higher ISO is the full dynamic range of the analog side of the analog/digital process.

The digital side is limited by the range of luminance values available. For a 12-bit file it's 0-4095 and for 4-bit it's 0-16383. A 16-bit raw file will have a digital range of 0-65535.

You can only use 100% of both ranges at base ISO.

selmslie Loc: Fernandina Beach, FL, USA

Ysarex wrote:

You've said otherwise in this thread. ...

P.S. Do you really think I don't understand this stuff?

So you don't even understand the instructions.

Refer only the the evidence in this post and the following one.

It is obvious you don't understand this stuff. You already proved that here.

If you ignore the RawDigger evidence you will put yourself at a disadvantage.

You have wasted your first two hours. Class ends at 4PM EST (3 PM CST). If you can't complete the exam by then you will fail the class.

So far everyone else has turned in their papers with the right answer and gone home.

Ysarex Loc: St. Louis

selmslie wrote:

So you don't even understand the instructions. b... (show quote)

I take your choice not to respond to the content of my last post as acknowledgement that you were wrong and leave it at that.

selmslie Loc: Fernandina Beach, FL, USA

Ysarex wrote:

I take your choice not to respond to the content of my last post as acknowledgement that you were wrong and leave it at that.

Then you would be wrong.

The content of your last post looked like your usual rant and it was not worth my time to read it.

You obviously don't know enough to pass the test or you are too embarrassed to admit that I am right.

Are you giving up?

TriX Loc: Raleigh, NC

anotherview wrote:

I understand the ISO value as fixed according to the engineering of the camera sensor.

This sensor captures photons that translate to an analog signal. The strength of this signal depends on the amount of photons landing on the sensor.

The in-camera processor converts this analog signal to a digital signal. This signal remains subject to amplification (gain). At this point, the signal functions as part of the exposure triangle.

This is my 2 cents.

Your 2 cents are a good distillation. Let me just add that the amplification (analog) can occur between the sensor and the digitizer, or digitally, after the digitizer (or both).

anotherview Loc: California

Thank you. Then I sense that some of the controversy here arises from when signal gain happens, before or after digital processing. The effect on signal noise may differ accordingly.

BTW: I'm technically inclined and received Navy training as an electronics technician. So I understand the controversy over ISO on a conceptual level. Some of the other material in this thread would require me to school myself further.

TriX wrote:

Your 2 cents are a good distillation. Let me just add that the amplification (analog) can occur between the sensor and the digitizer, or digitally, after the digitizer (or both).

Bbarn Loc: Ohio

Ysarex wrote:

GAIN DOES NOT ADD DATA.

True, but all the digital data isn't necessarily useful. If 60% of the data is 0 because the chosen ISO is too low for the SS and aperture used, you end up with a large amount of pure black. That happens when exposure is very low, the analog sensor's output is correspondingly very low, and all those very low analog values are subsequently converted to a digital 0 in the file.

OTOH, if additional gain is applied prior to A/D, some of that low analog data is converted to digital values above 0. Then the digital file contains more values above 0, contributing to a more accurate capture of the image.

If you skip the analog gain stage and do all the gain digitally (such as by moving a slider), all of those 0 values remain 0, and the low level data that might have been captured using an analog gain stage are lost and cannot be recovered.

anotherview Loc: California

From what I understand, the ISO value of the camera sensor has been engineered. It does not change. Its analog signal output strength may vary, of course.

Users who refer to ISO typically mean the digital signal that the camera has processed. This signal is subject to amplification (gain).

Bbarn wrote:

True, but all the digital data isn't necessarily u... (show quote)

Ysarex Loc: St. Louis

Bbarn wrote:

True, but all the digital data isn't necessarily u... (show quote)

Absolutely. Analog gain is almost always preferable to digital scaling. The only reason, and it's a pretty obscure reason, to chose digital scaling in post would be a situation where shooting circumstance came together with a benefit to retaining DR that the analog gain would otherwise clip.

selmslie Loc: Fernandina Beach, FL, USA

Ysarex wrote:

... The only reason, and it's a pretty obscure reason, to chose digital scaling in post would be a situation where shooting circumstance came together with a benefit to retaining DR that the analog gain would otherwise clip.

It's so obscure it's unimaginable.

Under what circumstances could you have clipping in the analog gain that could be recovered in on the computer? When you use one of the low ISO settings (below base ISO)?

If either one is clipped, it's gone. There is nothing that can be recovered.

Or are you referring to the JPEG SOOC rather than the raw file?

Maybe you need to post an example.

By the way, there is only an hour left for you exam. Are you conceding failure?

That's four questions. They only need four succinct answers.

Ysarex Loc: St. Louis

selmslie wrote:

It's so obscure it's unimaginable.

Under what circumstances could you have clipping in the analog gain that could be recovered in on the computer? When you use one of the low ISO settings (below base ISO)?

If either one is clipped, it's gone. There is nothing that can be recovered.

Or are you referring to the JPEG SOOC rather than the raw file?

Maybe you need to post an example.

https://www.uglyhedgehog.com/t-686341-2.html#12016716

selmslie wrote:

By the way, there is only an hour left for you exam. Are you conceding failure?

https://www.uglyhedgehog.com/t-687844-15.html#12062945
https://www.uglyhedgehog.com/t-687844-16.html#12063074

selmslie Loc: Fernandina Beach, FL, USA

Ysarex wrote:

https://www.uglyhedgehog.com/t-686341-2.html#12016716

There is nothing in that example that shows anything that anybody couldn't do simply by recovering the shadow information. It has nothing to do with ISO invariance.

And your time ran out. You failed to demonstrate that you understand the subject.

So you are a total failure. Take up finger painting.

Ysarex Loc: St. Louis

selmslie wrote:

There is nothing in that example that shows anything that anybody couldn't do simply by recovering the shadow information. It has nothing to do with ISO invariance.

You're wrong about that. Guess you didn't read it or understand.

selmslie wrote:

And your time ran out. You failed to demonstrate that you understand the subject.

https://www.uglyhedgehog.com/t-687844-15.html#12062945
https://www.uglyhedgehog.com/t-687844-16.html#12063074

selmslie Loc: Fernandina Beach, FL, USA

Ysarex wrote:

You're wrong about that. Guess you didn't read it or understand.

I didn't need to read it.

You didn't show any of the images at their full size so we can't see whether there is any noise. The biggest one is 1920x1280

You obviously had something to hide. You can't post an honest example. We all know why.

And you can't hide the fact that you don't know what you are talking about. You are just a poseur.

You have already made enough of a fool of yourself.

TriX Loc: Raleigh, NC

Guys, after 14 pages of argument and clearly no end in sight, maybe time to agree to disagree and let this one go?