[jsherman999]

Total light in itself is a pretty meaningless measurement, given the images many cameras produce with very small sensors.

Disagree. Total Light describes why for example the older, less efficient sensor in the D700 outperforms the newer, more efficient sensor in the K-5 at the same exposure.

.

[Rondec]

No I don't. I shoot for DOF and SNR. Not some 'exposure'. When I display a picture I don't put a little blurb at the bottom telling people what the ISO was.

So you don't care what the exposure was? That's crazy.

I shoot for a particular exposure -- if I hit my exposure on, there is going to be less visible noise, whether shooting full frame or crop sensor.

[redrockcoulee]

Point your K-r at a scene with shadows and bright light and around a 85mm lens. Now do the same with a 18mm lens. Your DSLR will give you different readings, no? Are you going to carry a big zoom to cover all the equivalent focal lengths you will use on your medium and large format?

Sure, use a DSLR and phone app if that works for you. But if you head over to the Large Format Forum and say, "hey guys, I know most of you use a DSLR to meter your scenes, which one works best for your..." I think you'll get a bunch of WTF responses.

A Pentax digital one-degree spot meter fits in your pocket, is more accurate than using a DSLR, you can meter from a distance and is not fooled by high or low contrast scenes because it uses a human brain to select your middle gray exposure. I strongly encourage using one.

I use a Minolta spotmeter for LF and a Luna Six Pro spotmeter for MF. I simply stated (and it was from the LF forum) that some people do plus stated that a plus of having a small dslr was the ability to carry it along with the larger format. My iPod meter is used for pinhole photography and it does have spot metering, But this is off topic; the point was that a reading at one ISO in one format will give you the same combination f stops and shutter speeds on another format. If at 200 ISO for my Whole Plate camera the exposure required is f11 @ 1/30 then for the same scene it is the same exposure for 200 ISO for my Hasselblad or my K-r or any full frame Nikon I can borrow. If my lens has a larger circle of illumination that the film size that "extra" light gets wasted but the amount of exposure is the same per square unit on any format.

[jsherman999]

jsherman999:

The sad thing is, you didn't even tried to answer my questions, you just copy/pasted from your dear article.

In all honesty I've answered that same question many, many times, even answering it when it came from you before. It really does get tiresome when the person asking the question willfully chooses to not understand the answer, for effect or to continue the argument or something.

[ElJamoquio]

So you don't care what the exposure was? That's crazy.

I shoot for a particular exposure -- if I hit my exposure on, there is going to be less visible noise, whether shooting full frame or crop sensor.

My ISO value matters less to me than the timestamp on the file. You're assuming too much for the rest of your statements.

[normhead]

Disagree. Total Light describes why for example the older, less efficient sensor in the D700 outperforms the newer, more efficient sensor in the K-5 at the same exposure.

.

You've presented no empirical evidence to say that's true...

And not everyone ages with you....




So use equivalence to explain how the K-5 managed to outscore both the 5D and D700 in colour depth and Dynamic Range, in dynamic range by a considerable margin.

[rburgoss]

'Stop' doesn't just refer to exposure, though. The vernacular includes other usage like "One stop DOF," or "one stop noise." Even if you don't like that usage, you should know what it's referring to: the effect on those things that changing f-stop causes.
.

Sorry again: The OP started all this mayhem with the argument about light gathering and about "equivalent f/stop" in terms of exposure. Do not come now arguing its all about SNR, or DoF.

SNR (Signal to Noise Ratio) is something measurable to itself, but not comparable between formats. Simply because SNR numbers change with time as soon as manufacturers come with new sensors. With Northrup's (the OP's and your) argument, then my *istD with APSC sensor size,k should have a much better SNR (less noise) than my new MX-1 compact, with a 1/1.7" sensor. But guess what: It doesn't. The *istD simply sucks in SNR compared to the MX-1. Wanna know the reason? Its 12 years difference in technology development! Not the sensor size!

OTOH, f/stops will remain the same with the pass of time. f/2.8 today is the same f2.8 of 75 years ago. Wanna know why? Because its a math formula that doesn't change. It is based on hard facts as focal length and internal lens diameter. It makes no difference if the lens is top quality recently designed all super lo ED IF CD CD AP (pick your choice of letters.... makes no difference) or a cheapo Focal (brand) lens from Kmart from 40 years ago... It makes no difference if the lens is for a 35 mm or a MF or a 1/2 frame or a Pentax auto 110 SLR... f/stop is the same and the measured light DOES NOT CONSIDER NEGATIVE (SENSOR) SIZE, because its irrelevant for EXPOSURE PURPOSES.

If its SNR (noise) you want to discuss, then base your discussion on what's relevant for this particular matter, which is a comparison between sensors and manufacturers at a specific time frame in history, and of course, to make comparisons valid, you will have to compare apples with apples. That is, compare sensors OF THE SAME SIZE where the only difference, would be the sensor brand, thus, its sole performance. Keep the rest under identical conditions. Otherwise any comparison is not valid.

'In learning any second language, you need to start with a translation mechanism - a way to map meanings. Car = Coche. Equivalence simply provides this translation mechanism for learning your photographic second language

And indeed after you learn it it's second nature... until you have to fire up the translation mechanism again to learn a third language (1'' sensor?) .

My God ! Better do your homework again. Your example is as good as a "clay block house during earthquake season..." Let me explain: Car has certain meaning, but such meaning will change according to whoever is asked about CAR, Some may think about their dream car, others about their own car, about their car at the mechanic, about the car they want to buy, etc. Better yet, someone that works in the train business may think of CAR as in "Pullman Car", "Sleeping car" or "Lounge car" in a passenger train...

On the other side, "coche" has far more meanings than you can imagine. It can be a car of course, but as said before, such car could be totally different from one person to another. But coche can also mean "horse carriage", "baby stroller", "supermarket cart" and probably other meanings....

But I have to give it to you, because your mistake in this example lets me stand on what I said before; about photography being like learning a language. If you keep translating... you will never get it right. Just understand it and do not attempt to figure out what's the meaning of anything under the old language. Of course, if you live, play and work with both languages, of course you should do so correctly with both, but not by "translating" between them.

In fact, I believe there's a movie whose title describes exactly what I've been trying to say "Lost in Translation"

[ChristianRock]

Regarding noise - there was a 1.4 stop difference in noise between the 5D Mk II and the 50D, released at about the same time (the 50D was actually released a bit earlier). Then there was a 1.7 difference between the Mk III and the 650D, released exactly at the same time. In Nikon-land, the difference between the D3 and the D300, released at the same time, was 1.8 stops (!!!). Between the D700 and the D90 it was 1.2 stops. It's hard to be conclusive after that because Nikon spreads out their camera releases, but as sensor technology has somewhat stabilized, 1.2 seemed to be the norm (no changes in S/N in either types of cameras recently, so no surprise).

Regarding dynamic range - it's gotten to where it's much closer now between APS-C and 135mm. Certainly less than 1.2 stops. In fact, the D810 and the K-5 IIs have the exact same S/N ratio.

My point is, if these things vary, how can there be equivalence? Sensor size is the only variable that changes by a truly equivalent proportion. Everything else is relative.

Also: I do most of my shooting at ISO 100. I'm already getting great S/N ratio with my APS-C, and mine was released in 2007.

[jsherman999]

That's the part where we will just have to agree to disagree... image noise and DR are determined by the technology in the sensor, not by equivalence.

My birthday was in August, but if you (or anyone) wants to do something for my next brithday that would really please me it's this: Find some way to realize that what I underlined above ^^ is false, or at least so far from the complete story as to be false.

I can provide simple equations and explanations and examples, and I can provide sources, I can maybe even provide information about people who work in the field who can explain maybe better than me. But please consider this:

1) Total Light is determined by physical aperture and sensor size (and distance to subject and subject luminosity)
2) Shot noise (photon noise) is mostly affected by Total Light.
3) Read noise is what's largely determined by sensor efficiency.
4) Total image noise is Shot noise + read noise.

But sensors of the same generation that show the same read noise - cut from the same exact wafer - will show less image noise if they're larger and the image was taken with the same FOV, distance and exposure. This is because of the Total Light, which is easily described by equivalence.

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.

[Kunzite]

In all honesty I've answered that same question many, many times, even answering it when it came from you before. It really does get tiresome when the person asking the question willfully chooses to not understand the answer, for effect or to continue the argument or something.

Nope, you didn't - not even once. You do not even accept the questions, that's why you'd copy paste over and over again - to switch the discussion back to the limits of your dear article.

All the 3 points from my question are things that could break "equivalence" as a valid tool for making informed choices. Because a tool which points you to an incorrect answer is worse than useless.

[normhead]

My birthday was in August, but if you (or anyone) wants to do something for my next brithday that would really please me it's this: Find some way to realize that what I underlined above ^^ is false, or at least so far from the complete story as to be false.


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]

Maybe you missed this. here, I'll repost it for you...

Disagree. Total Light describes why for example the older, less efficient sensor in the D700 outperforms the newer, more efficient sensor in the K-5 at the same exposure.

.

You've presented no empirical evidence to say that's true...

And not everyone ages with you....




So use equivalence to explain how the K-5 managed to outscore both the 5D and D700 in colour depth and Dynamic Range, in dynamic range by a considerable margin.

I'm not giving up on this, it goes to the heart of this argument.

[tuco]

I use a Minolta spotmeter for LF ...

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?

[jsherman999]

So use equivalence to explain how the K-5 managed to outscore both the 5D and D700 in colour depth and Dynamic Range, in dynamic range by a considerable margin.

The progress in sensor efficiency explains it. I would hope some progress was made in the years between those sensors were developed! (take a look at the noise/snr score difference between the D700 & K5 there.) Also, the DR advantage that DXO reports is only at base ISO, look at the chart that includes the ISO range:



The point I was refuting was that "Total Light doesn't matter", and that's demonstrably false when you see how a less efficient, older sensor still outperforms a newer one in SNR - all due to the difference in total light. (and yes, the D700 sensor is less efficient in QE than the K5 as measured by sensorgen and other sources.)

The Canon sensor is less efficient too - but so much less efficient that the effects of read noise take over and really affect the SNR score, drop it down into aps-c territory. Canon really is still playing catch-up to Sony, but not as much as they were back then.

Ask yourself this, Norm - if Total Light doesn't matter, how could the older, less efficient 5D possibly match a newer sensor (K-5) in SNR - and how could the older, less efficient D700 still beat it?



.

[ChristianRock]

My birthday was in August, but if you (or anyone) wants to do something for my next brithday that would really please me it's this: Find some way to realize that what I underlined above ^^ is false, or at least so far from the complete story as to be false.

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.

[Rondec]

No I don't. I shoot for DOF and SNR. Not some 'exposure'. When I display a picture I don't put a little blurb at the bottom telling people what the ISO was.

The progress in sensor efficiency explains it. I would hope some progress was made in the years between those sensors were developed! (take a look at the noise/snr score difference between the D700 & K5 there.) Also, the DR advantage that DXO reports is only at base ISO, look at the chart that includes the ISO range:



The point I was refuting was that "Total Light doesn't matter", and that's demonstrably false when you see how a less efficient, older sensor still outperforms a newer one in SNR - all due to the difference in total light. (and yes, the D700 sensor is less efficient in QE than the K5 as measured by sensorgen and other sources.)

The Canon sensor is less efficient too - but so much less efficient that the effects of read noise take over and really affect the SNR score, drop it down into aps-c territory. Canon really is still playing catch-up to Sony, but not as much as they were back then.

Ask yourself this, Norm - if Total Light doesn't matter, how could the older, less efficient 5D possibly match a newer sensor (K-5) in SNR - and how could the older, less efficient D700 still beat it?



.

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.