The Olympus thread https://sigforum.com/eve/forums...0601935/m/5040036174 had me talking with a friend who also dabbles in photography. I mentioned what I considered drawbacks in the sensor size and megapixel output, and he was confused. I admit, I may be thinking about this wrong, so please correct me if I am. My point was that the increase in the megapixel count on the micro four thirds sensors, and their size was a hindrance to expansion, given the glass available, and ability for the sensor (per pixel) to gather light. I expanded into the same issue with the iPhone - sensor size and megapixel count.

He came back to this later and said his understanding of light in a camera was different. The way I laid out my comparison was as follows:

Take the total megapixel count and divide by the Sensor size (H*W). I used my Eos R as an example - the sensor is 24*36 mm, or 864 mm squared. My camera is 30.3 megapixels, meaning there are roughly 35,000 pixels per square mm. The fewer megapixels per square mm, the bigger the pixel, the less light needed, and therefor the slower the lens to create a similar image on another camera. When I did the math on Olympus, he thought that because the crop factor was 2.0, it needed 2x the light. The micro four thirds sensor is 17.3 x 13mm, which yields 225 mm squared. My thought was that because a full frame camera is 3.84x larger (864mm sq vs 225mm sq) it would need 3.84x the light assuming the same megapixel camera.

I freely admit, I do not shoot with Olympus, but thought the same math that calculated the light for APS-C to full frame would be linear as you moved down. I learned in a photography class that a Canon M3 with its 23.2 MP sensor with a constant frames image needs 2 stops more light to produce an image similar to a 5D mark 3 (22.3 MP) with the same framed image. This was done in a controlled studio to limit the effects of changing natural light. Obviously, the APS-C sensor camera needed to be moved back to produce the same framed image as a result of the 1.60x crop factor, but using an adaptor, we used the same lens, and effectively duplicated the picture with 2 additional stops.

The question is was this wrong, and purely a coincidence, or does the f-value per the lens determine focused light? I have read that an F/2 lens on a full frame Sony does not produce F/2 light on their A6400 APS-C camera, so the question remains - what is the math to determine light - is it based on sensor size and megapixel count, or is it crop factor and lens based?

Do you also need to account for the ISO setting on the camera. That will set the sensor to it's responsiveness to light coming in.

I'll leave to the experts in photography, math and physics to how determine if an ISO adjustment will equate to an F-stop or not.

I know that doing some astro photography, higher ISO gives you better ability to capture light but at the tradeoff of noise.

quote:

Originally posted by mkueffer:
Do you also need to account for the ISO setting on the camera. That will set the sensor to it's responsiveness to light coming in.

I'll leave to the experts in photography, math and physics to how determine if an ISO adjustment will equate to an F-stop or not.

I know that doing some astro photography, higher ISO gives you better ability to capture light but at the tradeoff of noise.

In the referenced example above, we kept all of the sitting alike (ISO 400, F/2, 1/60 th of a second). We were able to duplicate two separate ways - we moved the F value up F/2.4 on the full frame to match the F/2, 1/60 and ISO 400 on the APS-C, and then went the other way, and produced a picture on the full frame, and adjusted the APS-C down to ISO 100 get match the light while keeping the other setting fixed. We also tried to increase exposure through shutter speed at F/2 and ISO 400 but instead of 1/60, we moved down to 1/15th of second, (2 stops), and recreated the image.

This is what has me confused - I'm not sure if it is linear or luck.

Pixel pitch doesn't affect exposure. Bigger pixels reduces noise at high ISO, but it's the ISO that affects exposure not the pixel pitch.

Lens specifications are relative to the sensor size. To convert specifications for an APS-C lens to it's Full Frame equivalent, multiply by 1.5. Thus, 23mm F1.4 APS-C prime is roughly equivalent to a 35mm F1.8 Full Frame prime.

Accordingly, depth of field and available light through the aperture are affected by this 1.5x factor. That means F2 on Full Frame will get 1.5x more light (1/2 stop more) than F2 on an APS-C. Expressed inversely, F2 on an APS-C will be 1/2 stop less light and 1/2 stop greater depth of field than F2 on a full frame.

For M4/3 sensors, the factor is 2x.

Shutter speed and ISO is not affected by lens parameters, so those are not affected by sensor size. ISO 1600 is ISO 1600. The only difference is that ISO1600 will be less noisy with a bigger pixel pitch--butbthe exposure is exactly the same.

Also, you are thinking about light hitting your sensor incorrectly.

A Full Frame sensor doesn't NEED 2x* more light than M4/3 to cover its sensor;

At any given moment, a FF sensor CAPTURES 2x more light than M4/3 in any given instant and aperture opening.

It's like dropping a bucket into a river. A bigger bucket holds more water. Both fill up instantly when immersed.

*It's 2x not 3.8x because the math isn't linear.

You also have to consider the lenses involved. The lens for a full-frame camera will generally let more light through than that of a smaller sensor camera. That is one "advantage" of smaller sensors - the lenses can be smaller and less expensive.

quote:

Originally posted by Aeteocles:
Pixel pitch doesn't affect exposure. Bigger pixels reduces noise at high ISO, but it's the ISO that affects exposure not the pixel pitch.

Lens specifications are relative to the sensor size. To convert specifications for an APS-C lens to it's Full Frame equivalent, multiply by 1.5. Thus, 23mm F1.4 APS-C prime is roughly equivalent to a 35mm F1.8 Full Frame prime.

Accordingly, depth of field and available light through the aperture are affected by this 1.5x factor. That means F2 on Full Frame will get 1.5x more light (1/2 stop more) than F2 on an APS-C. Expressed inversely, F2 on an APS-C will be 1/2 stop less light and 1/2 stop greater depth of field than F2 on a full frame.

For M4/3 sensors, the factor is 2x.

Shutter speed and ISO is not affected by lens parameters, so those are not affected by sensor size. ISO 1600 is ISO 1600. The only difference is that ISO1600 will be less noisy with a bigger pixel pitch--butbthe exposure is exactly the same.

I may be misunderstanding what you have posted, but I think you've got some stuff mixed up here.

Some preliminaries.

APS-C is 2/3 the width and height of a full frame sensor (so the full frame is 1.5x bigger in width and height). The difference in area is 2/3*2/3 = 4/9 or 1.5x*1.5x = 2.25x.

Micro Four Thirds is a different shape, but you can call it about half the width and height (so the full frame is about twice the width and height). The difference in area is 1/2*1/2 = 1/4 or 2x*2x=4x.

We are going to ignore the difference in shape and just pretend the M43 sensor is exactly half the length and width of the full frame sensor.

On to the details.

1. A scene which is properly exposed at f/2.8, 1/500s shutter speed, and ISO 100 is properly exposed at f/2.8, 1/500s shutter speed, and ISO 100. That is true whether the image sensor (or piece of film) is M43, APS-C, 35mm/Full Frame, the size of the head of a pin, or large-format 8"x10" Hasselblad. It is also true whether the lens focal length is 10mm or 500mm or anywhere in between.

2. The depth of field of a scene 10 feet from the camera shot with a 50mm lens at f/2.8 is the depth of field of a scene 10 feet from the camera shot with a 50mm lens at f/2.8. Again, that is true whether the image sensor (or piece of film) is M43, APS-C, 35mm/Full Frame, the size of the head of a pin, or large-format 8"x10" Hasselblad

Now, if you take that shot on a full frame camera and on a micro four thirds camera, they won't look the same. The 50mm lens at f/2.8 projects a certain image onto the plane of the sensor/film. That is true regardless of sensor/film size. But the physically smaller sensor captures a smaller portion of that projected image.

If you take the full-frame image and crop off 1/4 of the image from each side and from the top and bottom, it will look EXACTLY THE SAME (in terms of framing, exposure, and depth of field) as the M43 image.

In fact, full frame cameras often have the ability to capture images using only an APS-C sensor-sized portion from the middle of the full frame sensor (intended for use with APS-C lenses). Ignoring technology differences, with the same lens and settings, a full frame image cropped to APS-C size, an image taken in APS-C mode on a full frame camera, and an image taken on an APS-C camera will be IDENTICAL.

3. Because of this "crop factor," if you want an image to LOOK the same between different sensor sizes, things change. You have two options.

3.1. If you want to use a 50mm lens on both cameras, you have to move farther away with the M43 camera to frame the same image. There will be some perspective differences which may or may not be significant, but the framing will be about the same between a 50mm shot at 10 feet on the full frame camera and a 50mm shot at 20 feet on the M43 camera. Because you're twice as far away with the M43 camera, if both shots are at f/2.8, the depth of field will be about twice as large in the M43 shot (size of depth of field increases with distance from the camera).

3.2. If you want to take the shot from the same place, you have to change the lens. If you want to match a 50mm, 10 foot full frame shot on a M43 camera, you use a 25mm lens. If the full frame shot is properly exposed at 1/500s and f/2.8, the M43 shot will be, too. However, because the M43 shot is using a 25mm lens at f/2.8, the depth of field will be about twice the full frame shot with a 50mm lens at f/2.8 (size of depth of field increases with decreasing focal length).

3. (continued) This is why people talk about "effective focal length" (like saying a 25mm lens on a M43 camera is the same as a 50mm lens on a full frame camera) even though the focal length doesn't change. It's in the context of getting a shot with the same framing as a shot on a full frame camera.

This is also why people talk about smaller sensors having larger depth of field, even though depth of field doesn't depend on sensor size. It's in the context of getting a shot with the same framing as a shot on a full frame camera.

4. A 50mm lens at f/2.8, regardless of sensor/film size, projects a certain quantity of light onto the plane of the sensor/film PER AREA of sensor/film.

Given the same focal length and f-stop, a M43 sensor with 1/4 the area of a full frame sensor gets 1/4 the light of a full frame sensor. They get THE SAME amount of light per square millimeter of sensor.

This DOES NOT affect exposure. AT ALL. The same light intensity hitting the sensor means the same shutter speed and ISO results in proper exposure.

What it does affect is noise. A certain amount of electrical noise in the sensor is unavoidable. A larger sensor means more light, which means more signal, which means the the noise is relatively smaller.

Given absolutely identical sensor technology, a M43 sensor gives up two stops of low-light performance (in the sense of, "how high an ISO can I use before the noise becomes unacceptable?") compared to a full-frame sensor because is gets 1/4 the light.

Note that pixel pitch actually isn't especially important to this part of the discussion, even though it might seem to be at first glance. Larger pixels means more light hitting each pixel, which means a higher signal-to-noise ratio, right?

Well, yes. But.

Let's say you have a 20 megapixel M43 camera and an 80 megapixel full frame camera, so the size of the sensor pixels are identical (and let's further assume the sensor technology is identical). Then you take two pictures of the same scene, using the same settings on each camera.

The per-pixel noise should be the same, right? Yes, it absolutely is.

However, the full-frame shot has four times as many pixels. You can resample the full-frame image down to a 20 megapixel image (or use noise reduction algorithms that result in a similar loss in sharpness without technically reducing the resolution of the image) and reduce the noise by a factor of about 4 (two stops). This doesn't work perfectly but it's pretty close.

quote:

Originally posted by Aeteocles:
Also, you are thinking about light hitting your sensor incorrectly.

A Full Frame sensor doesn't NEED 2x* more light than M4/3 to cover its sensor;

At any given moment, a FF sensor CAPTURES 2x more light than M4/3 in any given instant and aperture opening.

It's like dropping a bucket into a river. A bigger bucket holds more water. Both fill up instantly when immersed.

*It's 2x not 3.8x because the math isn't linear.

Okay, if that is the case, why is my friend's Sony A7rII not perform as well at weddings with its 42 MP sensor than his A7II with the 24 MP sensor when using the same lens? He has mentioned numerous times he needs 2 to 3 more ISO steps to generate the equal lighting that the 24MP camera gets.

What am I missing?

I think you mis-understood my post, I understand a smaller sensor need more light. My understanding was pixel size impacted ability to gather light, but you say this is not the case?

quote:

Originally posted by 280nosler:

quote:

Originally posted by Aeteocles:
Also, you are thinking about light hitting your sensor incorrectly.

A Full Frame sensor doesn't NEED 2x* more light than M4/3 to cover its sensor;

At any given moment, a FF sensor CAPTURES 2x more light than M4/3 in any given instant and aperture opening.

It's like dropping a bucket into a river. A bigger bucket holds more water. Both fill up instantly when immersed.

*It's 2x not 3.8x because the math isn't linear.

Okay, if that is the case, why is my friend's Sony A7rII not perform as well at weddings with its 42 MP sensor than his A7II with the 24 MP sensor when using the same lens? He has mentioned numerous times he needs 2 to 3 more ISO steps to generate the equal lighting that the 24MP camera gets.

What am I missing?

I think you mis-understood my post, I understand a smaller sensor need more light. My understanding was pixel size impacted ability to gather light, but you say this is not the case?

As Maladat explained better than me (in my defense, I'm typing from a phone), same aperture+iso+shutter speed on the same size sensor gets the exact same amount of light.

The A7RIII may have more NOISE at high ISO because of its high resolution (and smaller pixel pitch). Thus, in order to get a "clean" imgae, your friend may decide to drop the ISO down two stops and compensate with a faster aperture or slower shutter speed.

However, given the same aperture+iso+shutter, both cameras will produce the same EXPOSURE.

Note, that EXPOSURE is not the same as resolution (ability to blow the picture up and get fine details), noise, or dynamic range. Exposure is a expression of how much light is being captured. It does not speak to the quality of that light.

quote:

Originally posted by 280nosler:
Okay, if that is the case, why is my friend's Sony A7rII not perform as well at weddings with its 42 MP sensor than his A7II with the 24 MP sensor when using the same lens? He has mentioned numerous times he needs 2 to 3 more ISO steps to generate the equal lighting that the 24MP camera gets.

What am I missing?

Do you mean he needs 2 or 3 stops higher ISO to get the same exposure? That with both cameras set to f/2.8 and 1/500s, the A7R II needs to have the ISO set two or three stops higher or the pictures are super dark?

If that's the case, there's a discrepancy somewhere. Exposure does not depend on the sensor size or pixel pitch.

It may be down to what you might call calibration errors.

If he's using different lenses, the f-numbers are based on the actual size of the aperture (which is what determines depth of field), but some light is lost to internal reflection off of lens element surfaces and such, so the amount of light that actually makes it through the lens is typically somewhat less than the f-number would indicate. Some lenses are a stop or two dimmer than you would expect. When people get really anal about this, they talk about T-numbers (T for "transmission") to refer to the actual light transmission of the lens. Cinema lenses (which are STUPID expensive) are often marked in T-stops rather than F-stops.

Some sensors vary by a stop or so from the ideal of what ISO is supposed to mean.

If he's shooting JPEG rather than RAW, the image processor in the A7R II may be different and may be producing darker images than the one in the A7 II.

Or there might be a problem somewhere.

If you mean he needs to be shooting 2 or 3 stops lower ISO on the A7R II to get the same amount of noise as the A7 II... I would not expect that.

The difference in pixel area is squareroot(24/42) = 0.75. The A7R II is more recent than the A7 II, right? Assuming similar sensor tech, the per-pixel noise should be less than a stop worse based on the smaller pixels (and, as discussed in my dissertation above, on a per-photo-area basis, the noise should be about the same). Sony advertised the A7R II as using some uncommon sensor technology that was supposed to reduce noise.

Maladat,

Your explanation is good, and may very well be more correct than my understanding. I have one discrepancy in my understanding that needs your help in resolving:

My Fuji (APS-C) 23mm F1.4 has a shallower depth of field than a 35mm Full Frame lens at F1.4. EDIT: ugh, I meant wider depth of field.

My understanding is that my aperture is affected by the same 1.5x "crop factor". That, in order to get the same look (bokeh), I have to "upconvert" my F number by 1/2 stop or 1.5.

Further, my little cell phone camera has a fairly large F number (F1.8) but is incapable of throwing bokeh without the help of software.

This message has been edited. Last edited by: Aeteocles,

quote:

Originally posted by Aeteocles:

As Maladat explained better than me (in my defense, I'm typing from a phone), same aperture+iso+shutter speed on the same size sensor gets the exact same amount of light.

The A7RIII may have more NOISE at high ISO because of its high resolution (and smaller pixel pitch). Thus, in order to get a "clean" imgae, your friend may decide to drop the ISO down two stops and compensate with a faster aperture or slower shutter speed.

That is the point I am hung up on - he is a professional wedding photographer and has shown me pictures of each camera and you can see the settings in Lightroom. The equivalent Sony G-master 85mm f/1.4 needs the iso at 1600 or 3200 on the A7rII (42MP) in order to produce a like exposed image on the A7II (24MP) with an ISO of 400.

In this we are comparing apples to apples, where the only thing changing is the megapixel count, and it seems as if megapixels per square MM has en effect on ability to convert light. Would you agree?

quote:

Originally posted by maladat:

Do you mean he needs 2 or 3 stops higher ISO to get the same exposure? That with both cameras set to f/2.8 and 1/500s, the A7R II needs to have the ISO set two or three stops higher or the pictures are super dark?

If that's the case, there's a discrepancy somewhere. Exposure does not depend on the sensor size or pixel pitch.

It may be down to what you might call calibration errors.

If he's using different lenses, the f-numbers are based on the actual size of the aperture (which is what determines depth of field), but some light is lost to internal reflection off of lens element surfaces and such, so the amount of light that actually makes it through the lens is typically somewhat less than the f-number would indicate. Some lenses are a stop or two dimmer than you would expect. When people get really anal about this, they talk about T-numbers (T for "transmission") to refer to the actual light transmission of the lens. Cinema lenses (which are STUPID expensive) are often marked in T-stops rather than F-stops.

Some sensors vary by a stop or so from the ideal of what ISO is supposed to mean.

If he's shooting JPEG rather than RAW, the image processor in the A7R II may be different and may be producing darker images than the one in the A7 II.

Or there might be a problem somewhere.

If you mean he needs to be shooting 2 or 3 stops lower ISO on the A7R II to get the same amount of noise as the A7 II... I would not expect that.

The difference in pixel area is squareroot(24/42) = 0.75. The A7R II is more recent than the A7 II, right? Assuming similar sensor tech, the per-pixel noise should be less than a stop worse based on the smaller pixels (and, as discussed in my dissertation above, on a per-photo-area basis, the noise should be about the same). Sony advertised the A7R II as using some uncommon sensor technology that was supposed to reduce noise.

He only shoots RAW, he only shoots with Sony A7rII, A7II. His wife does the video, and his lenses are:
Sony 85MM f/1.4 GM
Sony 70-200 f/2.8 GM
Sigma 35MM f/1.2 ART

As for when they came out, my guess was sometime around 2015, maybe? I just looked it up, and the A7rII has a low pass filter.

quote:

Originally posted by 280nosler:

quote:

Originally posted by Aeteocles:

As Maladat explained better than me (in my defense, I'm typing from a phone), same aperture+iso+shutter speed on the same size sensor gets the exact same amount of light.

The A7RIII may have more NOISE at high ISO because of its high resolution (and smaller pixel pitch). Thus, in order to get a "clean" imgae, your friend may decide to drop the ISO down two stops and compensate with a faster aperture or slower shutter speed.

That is the point I am hung up on - he is a professional wedding photographer and has shown me pictures of each camera and you can see the settings in Lightroom. The equivalent Sony G-master 85mm f/1.4 needs the iso at 1600 or 3200 on the A7rII (42MP) in order to produce a like exposed image on the A7II (24MP) with an ISO of 400.

In this we are comparing apples to apples, where the only thing changing is the megapixel count, and it seems as if megapixels per square MM has en effect on ability to convert light. Would you agree?

Well, yeah, I can't tell you why there's a discrepancy. The theoretical is that the pixel pitch shouldn't affect exposure, and surely not to that extent.

quote:

Originally posted by Aeteocles:
Maladat,

Your explanation is good, and may very well be more correct than my understanding. I have one discrepancy in my understanding that needs your help in resolving:

My Fuji (APS-C) 23mm F1.4 has a shallower depth of field than a 35mm Full Frame lens at F1.4.

My understanding is that my aperture is affected by the same 1.5x "crop factor". That, in order to get the same look (bokeh), I have to "upconvert" my F number by 1/2 stop or 1.5.

Further, my little cell phone camera has a fairly large F number (F1.8) but is incapable of throwing bokeh without the help of software.

Your experience with your Fuji is not what I would expect. A shorter focal length lens at the same distance (to match the framing of the 35mm lens on the full frame camera) should have a larger depth-of-field.

That's what you observe with your cell phone camera. Because of the tiny sensor, to get framing about the same, they use absurdly short focal length lenses. The iPhone X uses a 4.25mm wide angle lens and a 6mm telephoto lens ( https://lumoid.com/guides/why-...%2028mm%20and%2033mm. ). Because the focal length is so small, when you frame an image the same way you would on a full frame camera, the depth of field is HUGE.

Incidentally, this is also why inexpensive camcorders seem to do a better job of video autofocus than expensive DSLRs or mirrorless cameras - they aren't actually better at autofocus, they just have tiny sensors so the depth of field is huge.

This is also why a lot of portrait photographers use medium-format or full-frame cameras and refused to switch to digital back when all the DSLRs were APS-C. For portrait photography you want razor-thin DOF. Probably the iconic portrait photography lens on a full-frame or 35mm camera is an 85mm f/1.4. To match the DOF with the same framing on an APS-C camera, you'd need a 130mm f/1.4 lens, which would be huge and tremendously expensive.

There is an excellent discussion of this topic with very well done demonstration photos here, showing what happens when you use a smaller sensor and crop to get the same framing, move back to get the same framing, and change focal length to get the same framing as a larger sensor:

https://fstoppers.com/educatio...s-depth-field-312599

quote:

Originally posted by 280nosler:
That is the point I am hung up on - he is a professional wedding photographer and has shown me pictures of each camera and you can see the settings in Lightroom. The equivalent Sony G-master 85mm f/1.4 needs the iso at 1600 or 3200 on the A7rII (42MP) in order to produce a like exposed image on the A7II (24MP) with an ISO of 400.

In this we are comparing apples to apples, where the only thing changing is the megapixel count, and it seems as if megapixels per square MM has en effect on ability to convert light. Would you agree?

That's really bizarre. The same scene with the same lens, the same f-stop, and the same shutter speed requires 2-3 more stops of ISO on the A7R II?

That shouldn't be the case.

If you haven't specifically confirmed everything is the same except the ISO, I would suspect a metering setting somewhere is trading ISO for aperture or shutter speed differently between the two cameras (or maybe that he's using a flash which the two cameras handle differently), but I can't imagine a professional wedding photographer not noticing that.

If it IS the case that everything is the same except the ISO (and isn't an issue with how the RAW files are post-processed), then it is either the result of some quirk specific to the type of equipment he is using or there is a problem (or possibly some screwed up setting somewhere) with his equipment.

The behavior you describe is ABSOLUTELY not something that you would expect to happen.

Thank you both for the detailed responses.

I'll follow up with him when he his next wedding and watch his workflow in Lightroom.

I still can't process that a smaller pixel can receive light as easily as a bigger pixel, but I will try to test that out myself.

Maladat,

The Fuji 56mm 1.2 (APS-C) is the Full Frame equivalent of 85mm F1.4. I think you got the conversion backwards. (Recall, APS-C shooters have a telephoto advantage).

Because the APS-C sensor is smaller, and the focal length is correspondingly shorter (56 vs 85), the aperture circle is going to be smaller and the depth of field is wider at any particular aperture as compared to Full Frame.

Thus, in order to get the same razor thin depth of field, my APS-C lenses have to be 1/2 stop faster than it's FF equivalent.

I think we're on the same page, but the typing is hard

(Edit: in one of my posts, I said my 23mm F1.4 had a shallower depth of field than the Full Frame equivalent 35mm F1.4. Thats wrong, I meant to say not as shallow, ie a wider depth of field. )

quote:

Originally posted by 280nosler:
I still can't process that a smaller pixel can receive light as easily as a bigger pixel, but I will try to test that out myself.

Well, in a sense, it can't, which is why the per-pixel noise is worse in a sensor with smaller pixels.

I think the issue you're having is a misunderstanding of what ISO really means.

These numbers are COMPLETELY MADE UP, but they should get the point across.

Think of ISO 100 as meaning "when white light of intensity 1000 lights per square inch hits the sensor, the picture will come out medium gray."

Then ISO 200 means "when white light of intensity 500 lights per square inch hits the sensor, the picture will come out medium gray."

Then ISO 400 means "when white light of intensity 250 lights per square inch hits the sensor, the picture will come out medium gray."

And so on.

The thing is, that's a specific amount of light PER AREA, not PER PIXEL.

If your image sensor has 10 pixels per square inch, then ISO 100 means a pixel outputs medium gray when it gets hit by 100 lights. ISO 400 means it outputs medium gray when it gets hit by 25 lights.

If your image sensor has 100 pixels per square inch, then ISO 100 means a pixel outputs medium gray when it gets hit by 10 lights. ISO 400 means it outputs medium gray when it gets hit by 2.5 lights. There's a lot less light hitting each pixel with this sensor, but what ISO 100 MEANS is that when a certain intensity of light hits it, it's supposed to output a certain brightness. The smaller pixels get less light, but when the camera converts the sensor output to an image file, it ends up being the same brightness.

The actual amount doesn't really matter, you can just scale it. In a simplified way, you can think of ISO 100 as recording the direct output of the pixel, and ISO 400 as putting the output of the pixel through an amplifier that multiplies it by 4 before you record it.

That works great great, but there are two problems.

One is noise: you multiply the noise by 4, also. The more you crank the ISO to boost low sensor output, the more you boost the noise.

The other is dynamic range. Basically, there's a smallest signal the sensor can actually put out. Last time I looked, for high-end modern full frame sensors at their native ISO (effectively, recording direct sensor output, usually 100), that's about 14 stops. That means that at the native ISO, the range between the lightest thing in the picture and the darkest is at most 14 stops. Anything outside that range gets lost (e.g., blown highlights). When you start cranking up the ISO, the smallest signal the sensor can put out doesn't get any smaller, but you lose some of the top end (you're making everything brighter... so stuff that was close to as bright as the sensor could see before is now too bright for the sensor to see).

So if you go to ISO 6400 (six stops from ISO 100) you would lose 6 stops of dynamic range. Instead of having 14 stops, you only have 8. You lose more to blown highlights and black shadows and can't accurately record scenes with as much contrast as at lower ISO.

Shooting ISO 6400 in a digital camera is essentially the same thing as taking an ISO 100 RAW picture and pushing the exposure six stops in Lightroom.

quote:

Originally posted by Aeteocles:
Maladat,

The Fuji 56mm 1.2 (APS-C) is the Full Frame equivalent of 85mm F1.4. I think you got the conversion backwards. (Recall, APS-C shooters have a telephoto advantage).

Because the APS-C sensor is smaller, and the focal length is correspondingly shorter (56 vs 85), the aperture circle is going to be smaller and the depth of field is wider at any particular aperture as compared to Full Frame.

Thus, in order to get the same razor thin depth of field, my APS-C lenses have to be 1/2 stop faster than it's FF equivalent.

I think we're on the same page, but the typing is hard

(Edit: in one of my posts, I said my 23mm F1.4 had a shallower depth of field than the Full Frame equivalent 35mm F1.4. Thats wrong, I meant to say not as shallow, ie a wider depth of field. )

Yes, we're on the same page, I was actually wondering if you'd just misspoken (mistyped?).

In terms of keeping the framing the same, smaller sensor = larger depth of field.

There are some ABSURD lenses designed for thin-DOF applications like portrait photography on less-than-full-frame cameras, e.g., Voigtlander makes 42.5mm f/0.95 and 60mm f/0.95 lenses for micro four thirds.