Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Debunking Tony Northrup.
- Thread starter juanm
- Start date
- Sort by reaction score
NO.
This is wrong.
When a m43 sensor based on equivalent technology shoots at F2.8, it is getting 1/2 the light that a FF sensor gets. BUT It only needs half the light because it is half the size. This like how it takes less paint to cover a wall than it takes to redo the whole room. If the paint is the same thickness, it doesn't magically become less opaque because you used less of it. It is covering less area.
There are FF cameras that can shoot crop lenses. It would be easy to prove, but I have a crop sensor already. Hell, Tony Northrup's videos show this already, but everyone clinging to this exposure myth want to make up unsupported crap about how different the sensors are.
Again, 100ISO on a sunny day = F16. It doesn't matter if you are FF, APSC, M43, or medium format film. If the exposure change was true, then the sunny 16 rule would not work.
NO ONE is saying to rename the lens. Tony and I are saying that when the manufacturer gives an equivalent focal length they should also give an equivalent aperture
I've typed example like 10 times already but I will do it again. They should advertise the 30mm f2.8 lens equivalency numbers as "60mm equivalent, f5.6 equivalent"
----------
This is ridiculous...
First of all the MFT sensor is half the size in each direction. That makes it 1/4 the area.
Second of all, no it doesn't need less light because it's smaller. The total image is made up of 1/4 the energy compared to a FF sensor. That's exactly why a MFT camera produces noisier images at the same ISO.
You're talking about using a FF sensor in crop mode to compare, say, a 60mm lens cropped to 90mm equivalent. That's not what we're talking about. We're referring to equivalent FOV lenses for smaller format cameras and how they are advertised. Invariably a manufacturer advertises their MFT lenses using the FF focal length equivalent, without also giving you the equivalent aperture.
You apparently don't think any equivalences should be given. But they are. The problem is ONLY focal length equivalence is given.
Edit: Let me give you yet another example. I have a 60mm f2.8 lens that I use on a D7100. It's equivalent FOV to 90mm on a full frame camera.
So let's compare the 60mm f2.8 lens on APS-C to 90mm f2.8 lens on FF.
What's the difference between the two? Would you argue that they are exactly the same?
They're not. The 60mm at f2.8 on APS-C has an equivalent DOF to 90mm f4.2 on full frame! The total amount of light is also exactly equivalent between the two. If the APS-C and FF are both 24 megapixel, the noisiness of the image will be very similar.
Does this example make more sense?
Last edited:
Tony accuses the manufacturers of not being honest by not making it clear that lenses on DX have different apertures to those the manufacturer states. If that isn't talking about naming or labelling the lens then I don't know what is.
But you are missing the entire point, manufacturers don't need to quote it because there is no real world 'equivalent aperture'. The aperture of the lens is exactly the same and gives the same exposure setting under the same conditions irrespective of sensor/film size or whether you just look through it.
If you want manufacturers to quote some difference based on say DoF then that is just as daft, just as sensor ISO performance depends on the sensor (which the lens maker won't know), DoF depends on subject distance (which the lens maker won't know), hence both of these issues are dealt with by information provided outside the base description of the lens.
In fact try it, I insist - hold up any lens fully open in front of a light source with no body on it. Now please imagine holding your hand up so an image is projected on it. Now a sheet of paper, now a DX body, then FX - if you still imagine the light intensity of the projected image on any of these things is different please explain how. ANY image plane at a fixed distance has an identical light intensity per sq mm falling on it. Unless you interfere with the light coming through the lens this doesn't change.
Not even slightly. Nope, not even a tiny bit.
So the most important aspect of the lens - its ability to give a known, predictable exposure - is accurately handled by the existing aperture measurement calculation, indeed for accurate exposure it MUST NOT be changed.
An FX lens doesn't care what size or technology sensor is in the body it is attached to.
Talk of sensor sensitivity is moot, that is entirely a sensor issue. In the one case where pixel density and technology is the same across both sizes the exposure the body calculates is exactly the same whether FX or DX is selected.
To give any "equivalent aperture" would be misleading and simply wrong. Exposure is determined by light intensity and that is wholly determined by the lens, not the sensor.
If your logic had been followed regarding sensor sensitivity then lenses would have been specified as say "200mm f/2.8 @ ISO100", the assumption being that the aperture necessary would somehow change with different sensitivities of film. That obviously wasn't the case, why would different sensor sensitivity affect the lens now?
Does ANY external lightmeter have ANY aspect of the lens or sensor size as an input? No, it is irrelevant to establishing what shutter speed and aperture to shoot a particular EV with.
When using an external lightmeter does it measure the light intensity? Yes. Does it need waving it across the entire subject to find out some "total light" figure? No.
As I said, its <just> physics...
----------
Nope, none. the total amount of light passed by the lens is the same in both cases. The light intensity captured by the sensor is the same in both cases. The exposure required is the same in both cases. The total light captured by the sensor is different but who cares, the settings required for correct exposure are the same.
The DoF may be different - but given DoF varies with subject distance how is the manufacturer supposed to quote this in a meaningful way? The average user would think they have a fundamentally different lens, exposures calculated by a lightmeter and manually entered would be wrong - that is a way greater error than any DoF effect.
Noisiness of the image is irrelevant, that is entirely a function of the body and sensor they are attached to, not the lens.
If you want to argue that some non-interchangeable lens camera manufacturers artificially quote a tiny lens as having f/2.8 when it patently doesn't (and is achieved by ramping up the gain to move the EV graph), then I would agree, but that is an issue of product description and isn't a fundamental issue with the standalone lens physical properties and unfortunately for interchangeable lenses that is what counts. Just to clarify, that is as if they put in ISO800 film but quoted it as ISO100, but again it is not a lens issue and certainly shouldn't be applied to the interchangeable-lens DSLR market. And Tony uses DSLRs to make his point.
You are simply not understanding Tony's point.
Nobody is saying to change the aperture numbers, or the ISO numbers, or the focal length numbers.
Does a DX lens have equivalent focal length numbers written on it? NO. But the equivalents are given in advertising materials.
That's what we are saying should be done with apertures. Rather than simply say "MFT 30mm f2.8 equivalent to 60mm FF" you say "MFT 30mm f2.8 equivalent to 60mm f5.6 FF".
Let me say this again.
The aperture numbers in camera settings would not change.
The focal length in camera settings would not change.
The ISO in camera settings would not change.
Nobody is saying to change the aperture numbers, or the ISO numbers, or the focal length numbers.
Does a DX lens have equivalent focal length numbers written on it? NO. But the equivalents are given in advertising materials.
That's what we are saying should be done with apertures. Rather than simply say "MFT 30mm f2.8 equivalent to 60mm FF" you say "MFT 30mm f2.8 equivalent to 60mm f5.6 FF".
Let me say this again.
The aperture numbers in camera settings would not change.
The focal length in camera settings would not change.
The ISO in camera settings would not change.
But to state that is incorrect and misleading, I understand his point I just disagree with it and hold it has no basis in physics. I really can't state it any clearer than that. If you hold that it does have a basis in physics please state how - so far there has been nothing proposed that is relevant to how a lens passes light, much stated about sensor noise and pixel density, none of which is remotely a case for re-stating the aperture of a lens as anything different to its physical aperture.
Please don't keep reiterating that I don't understand his point without some evidence that his point is valid...
Last edited:
Yep. TN apparently had a wrong understanding of how a lens works, and yet made a video explaining crop factors -and even insisting on the aperture equivalency. Of course, his videos have quite a production value, good lighting, and he's got (or rents) lots of expensive gear, so novices gobbled his errors like candy. Then when more experienced photographers started calling him out on his errors (poorly understood basic concepts, really) he had no choice than to persist, or he would lose face.
When a manufacturer states that a lens is f/4, it's referring only to the light intensity it's able to get on the focal plane (and it should be in T-stops, actually, but for photography it's close enough), not the DoF, since that would depend on factors external to the lens. Why TN imagined in the first place that it could refer to something else, I can only assume its because of poor knowledge, and the guy did it in good faith.
Last edited:
This is equally true for the light passing through the lens. The actual number of photons needed is reduced because you are illuminating a smaller area. This is why flashlights and speed lights have reflectors. They are trying to up the intensity by reducing the area.
Second of all, no it doesn't need less light because it's smaller. The total image is made up of 1/4 the energy compared to a FF sensor. That's exactly why a MFT camera produces noisier images at the same ISO.
You're talking about using a FF sensor in crop mode to compare, say, a 60mm lens cropped to 90mm equivalent. That's not what we're talking about. We're referring to equivalent FOV lenses for smaller format cameras and how they are advertised. Invariably a manufacturer advertises their MFT lenses using the FF focal length equivalent, without also giving you the equivalent aperture.
You apparently don't think any equivalences should be given. But they are. The problem is ONLY focal length equivalence is given.
Edit: Let me give you yet another example. I have a 60mm f2.8 lens that I use on a D7100. It's equivalent FOV to 90mm on a full frame camera.
So let's compare the 60mm f2.8 lens on APS-C to 90mm f2.8 lens on FF.
What's the difference between the two? Would you argue that they are exactly the same?
but the total amount of light needed is less. If you want to talk equivalent DOF, sure. But the Sunny 16 rule doesn't become a Sunny 11 rule on APSC. Take your D7000 out today and give it a shot (if it is sunny). Got to an area of bright light on full manual 100 ISO, set the aperture to 16 and the shutter to 1/ISO. Snap an image. Use a DX lens. Use an FX lens. Turn the camera sideways and shoot portrait mode. Have someone from an opposing political party push the shutter button. none of it matters.
if they are both 24 megapixels they are not equivalent technology. Now the APSC has smaller pixels.
This.
You can't change two things at once and then claim the end result is due to only one of them.
1. For equivalent pixel size, sensors of all sizes receive the same amount of light per pixel.
2. For equivalent sensor size, smaller pixels receive less light than larger pixels.
3. The only way to determine the actual effect of changing sensor size is to...wait for it...only change the sensor size. The only way we currently have to do that is to use the "crop" mode on FF DSLRs (or, as I've repeatedly pointed out, use different sizes of film).
Again, when I take a photograph on the same film stock with my Leica or on my Hasselblad or on my 4x5 camera, I don't have to compute the equivalent f/stop for each system *with respect to the exposure*. Of course I *do* have to consider that f/5.6 on the Leica is MUCH deeper DoF than f/5.6 on the Hasselblad (let alone the 4x5), if the focal length is held constant. That *is* a consequence of the different film sizes.
When I make 8x10 prints from the three negatives (35mm, 6x6, and 4x5), the 35mm has much more grain. Why? Not because it received less light per unit area (it didn't), but because I have to enlarge it a whole lot more than the others. And when I enlarge, I'm enlarging grain too. So the less I enlarge, the less prominent the grain.
EXACTLY the same process is at work with a digital sensor. The only hiccup with digital is that pixel sizes can vary as well, thus altering the PER PIXEL efficiency of the sensor.
Consider sensor A has 48MP and sensor B (of the same size) has 12. In this case, sensor A has pixels that capture 1/4 the light of sensor B. But the TOTAL light captured over the entire sensor is the same (because there are more pixels). If you then bin the pixels 4:1 (reducing the 48MP image down to 12MP), you effectively increase the efficiency of the 48MP sensor (at the cost of resolution), and thus reduce the noise in the final output. This was demonstrated nicely in a DPReview link posted above re: the A7s and A7r.
An 18MP m4/3 sensor is noisier than an 18MP FF sensor for two reasons: (1) the pixels are smaller, thus placing the output of each pixel closer to the noise floor and (2) the enlargement required to reach a given output size is greater. Only (2) is directly related to the sensor size, however.
Last edited:
The bright side of this argument is all the positive reinforcement that I need a D750 so I can finally prove someone on the internet is just wrong. 
The bright side of this argument is all the positive reinforcement that I need a D750 so I can finally prove someone on the internet is just wrong.
LOL - best reason ever!
This guy is selling a product... any time someone is selling something, especially on technique, skill, and information about a subject you'll have debate of it... This only helps him gain more views on you tube and more money. If you don't like him, don't watch his videos or create threads about him because it only helps him. This guy is like Ken Rockwell 2.0 . Clearly all the negative stuff written about Ken only helped him make more money since his website is still the main source of his income. Now I don't know much of his photography skills, but if you look up his website it's quite shocking how little work he has on it... and I personally don't find it very good overall. So you can use that as a personal indicator of how much he actually knows or doesn't know.
Last edited:
Exactly, you have to take into account the size of the system when figuring out DOF. That's why we need equivalent aperture along with equivalent focal length, so that a buyer understands that an f2.8 lens on MFT gives the same DOF as an f5.6 lens on FF, for any given field of view.
As it is, many if not most MFT users believe f2.8 on MFT is equivalent to f2.8 on FF. It's not.
Right, you can "bin" pixels from a higher resolution sensor. That's why we need to look at total light on the sensor, not per pixel light. UNLESS we are comparing two different sized sensors with the same resolution. Then we CAN simply compare pixel sizes directly.
But you keep referring to enlargement and I'm not sure why. The noise is due to smaller light gathering area and that's it.
For film, yes the **grain** becomes bigger if you take a smaller format film and enlarge it. But with digital this isn't what is happening. The equivalent to grain is pixel size, and if resolution is the same, DPI of the final image is the same. What changes is noise, because as you said, the smaller sensor has worse signal to noise
----------
The thing is, Tony Northrup isn't wrong. It's just that a lot of people don't understand what he's saying because they don't have the technical knowledge. Here's an article about equivalent aperture with a LONG discussion thread. If you read through it, it will probably make sense http://www.dpreview.com/articles/2666934640/what-is-equivalence-and-why-should-i-care
From that linked article "As such, you can say that a 50mm f/2 for Micro Four Thirds is equivalent to a 100mm f/4 Full Frame lens in terms of both field-of-view and depth-of-field."
Note, no mention at all of exposure.
No argument with that at all. The problem comes with Tony using DSLR and interchangeable lenses to demonstrate. As soon as you can move a lens between formats you cannot quote an "equivalence-for-depth-of-field" aperture 'cos you don't know what format body you will be attaching it to.
Do what you like with a fixed, non-changeable lens camera, any of the described factors are open to be described as anything, you have no idea what actual numbers are, and it largely doesn't matter to those users.
Once you have an interchangeable lens then the most important aspect of its aperture is for exposure, not DoF, as that has to be correct and match with the expectations of different bodies.
DoF tables have always been separate information, they should remain so. TBH honest changing the "equivalent" aperture even for DoF is IMHO a waste of time, anyone who doesn't understand the implications won't care, those that care will likely have enough knowledge to cope and adjust based on the DoF they see in their images.
It's not a waste of time, it's useful for directly and fairly comparing different formats. Many knowledgeable photographers fall for the lie that the little compact camera they're buying is good in low light because the lens if f1.7 or whatever.
BTW we actually could design cameras so that the equivalent aperture is used for settings and ISO is adjusted accordingly.
For example a MFT camera can have its aperture numbers doubled and ISO numbers doubled too. It would help any photographer who has experience with FF, because the settings for a given DOF would be identical
Those compact cameras are good in low light because of aperture 1.7. Certainly they are better than they would be at f3.4, right?
Now that tiny little sensor still has all the limitations of a tiny little sensor. This includes the inability to isolate a subject with focus. It includes having tiny little pixels that cannot gather light as readily as a full frame sensor with the same number of much larger pixels.
However, it does not alter the exposure reading. light coming through the lens and focused on an M43 sensor is EXACTLY THE SAME as the light coming through the lens and focused onto an M43 sized area of a full frame sensor. The light that the FF sensor collects on the outer reaches is light that wouldn't hit the M43 sensor anyway.
Maybe what you need to argue is how the camera manufacturers are assigning ISO values to their sensors. Who says a sensor is at ISO100, 400, or 3200? How is this measured and are camera makers fudging it a bit to make the specs look good? Because, the laws of physics for light are not bending for anyone.
Remember that the focal length of the lens isn't changing when you put it on a M43 camera. The field of view is all that changes.
http://en.wikipedia.org/wiki/Crop_factor (wikipedia link for easy reading)
http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/index.html (more technical version for advanced readers)
So, the focal length does not change. The Diameter does not change. Therefore, the APERTURE DOES NOT CHANGE. End of story.
The only equivalence is depth of field. People who care about this should already know it. No need to relabel every lens for it.
Yes we all get it. Focal length does not change. But we refer to equivalent focal length because it's useful information to have when comparing different systems and compact cameras etc. There is a reason compact cameras aren't advertised as "5mm-100mm lens". Instead they say "25mm-500mm equivalent lens" because consumers understand the field of view they're getting with equvalent numbers. Again, we all understand that the focal length is what it is.
In the same way that equivalent focal length is useful, equivalent aperture is useful. It tells you plainly and simply what kind of DOF you can expect, and what performance in low light you can expect
Is there anyone experienced in Cine in here to clear it up? I would imagine someone who frequently uses the same lenses on different film stock and size or different systems (RED, Canon, Black Magic, Alexa...) would have a pretty good grasp on this. I've read through this entire thread, and frankly I'm just confused now... I don't buy Northrup's argument either, but have only my very tiny understanding to go by.
Has anyone out there used the same lens on 16mm and 35mm film from the same stock? Does that affect the exposure? Is that even a good comparison to sensor sizes?
Has anyone out there used the same lens on 16mm and 35mm film from the same stock? Does that affect the exposure? Is that even a good comparison to sensor sizes?
If everyone all gets it, why do people still argue about exposure?
And equivalent DoF is enough f a niche that people who care about it should just be able to do some quick and dirty math in their head. Do you really need to match 1.53 instead of 1.5 when you do not have an infinitely variable aperture anyway?
It doesn't affect the exposure. That's not what this is about!
The exposure at a given ISO is always the same, because film is chemically created to have a certain ISO that gives the right exposure with a certain light intensity at the sensor.
And it's not about using the same lens. It's about comparing equivalent lenses. Like I was saying earlier, a 30mm f2.8 lens on MFT is equivalent in field of view and DOF to a 60mm f5.6 lens on FF.
When you take a picture with those two systems with the same f2.8, they will have the same field of view but different DOF and the MFT one will be noisier. Why? Because the sensor is smaller and therefore collecting less light even though lux at the sensor is the same. If you adjust the FF lens to 5.6 (cutting lux to 1/4) and raise ISO accordingly, the noise level will be very similar, assuming same level of technology. And the DOF will also be the same! Using the same *equivalent* settings, you get the same image. That's what makes it useful.
Last edited:
What doesn't get enough attention in this discussion is that the Image Quality of an ISO 400 Shot with FX is equivalent to the Image quality of an ISO 100 shot with MFT (regarding noise).
So TN is saying, that you need a 2 stops faster lens, when shooting with MFT to achieve equivalence in Noise levels, because you have to use 1/4th of the FX ISO.
That's what changes exposure and that's why f/2.8 on MFT is equivalent to f/5.6 on FX.
In the 1st video he is talking so much about DOF and not at all about exposure, so it seems like he is getting things wrong. He later explains in 2 followup videos, that he was talking about differences in Noise with same ISO.
The quadratic Noise to Crop-factor relationship (1/CF²) is only an approximation. If you want to have the correct values, look up the accurate ISO measures on DXOMark.
With larger sensors until about 1" (CF 2.7), 1/CF² is a very good approximation, so actually you should multiply f-numbers with crop-factor for an approximation of equivalence. (see attached spreadsheet)
So the video is right after all:
25mm f/2.8 on MFT is equivalent to 50mm f/5.6 on FX
So TN is saying, that you need a 2 stops faster lens, when shooting with MFT to achieve equivalence in Noise levels, because you have to use 1/4th of the FX ISO.
That's what changes exposure and that's why f/2.8 on MFT is equivalent to f/5.6 on FX.
In the 1st video he is talking so much about DOF and not at all about exposure, so it seems like he is getting things wrong. He later explains in 2 followup videos, that he was talking about differences in Noise with same ISO.
The quadratic Noise to Crop-factor relationship (1/CF²) is only an approximation. If you want to have the correct values, look up the accurate ISO measures on DXOMark.
With larger sensors until about 1" (CF 2.7), 1/CF² is a very good approximation, so actually you should multiply f-numbers with crop-factor for an approximation of equivalence. (see attached spreadsheet)
So the video is right after all:
25mm f/2.8 on MFT is equivalent to 50mm f/5.6 on FX
