PetaPixel

Are Camera Manufacturers Misleading Us by Not Calculating Sensor Size Into Specs?

Tony Northrup, an award-winning author and well-known reviewer of camera gear, recently put out a video that takes an interesting, in-depth look at how mirrorless camera companies might be fudging the specifications of mirrorless cameras to make them seem better than they are.

The video (which is basically an expansion on this shorter, very controversial demonstration) is an extremely long watch, coming in at almost 40 minutes, but it goes into great detail regarding how Northrup believes camera manufacturers are cheating and misleading us. He claims the stats like ISO, focal length and aperture are untrue as presented due to variations in hardware — specifically the sensor size.

MirrorlessLies_1

I won’t go into the nitty-gritty details, as that’s what the video is for, but it’s well worth noting that he specifically calls out Panasonic, Sony, Olympus and Fuji as the main perpetrators, while giving a bit of praise to Canon, Nikon and Sigma for being the honest “Good Guys.”

However, as Mirrorless Rumors admin Ale explained in an update to their original coverage of the video, Northrup may have made a few logical errors in his calculations, rendering the insights irrelevant.

According to MR (and the laws of physics), the aperture of lenses remains constant throughout sensor sizes, not relative, as Northrup asserts. Specifically, they point to this article from Admiring Light, which does a great job of explaining why this is.

MirrorlessLies_2

This has been a heavily debated topic, especially in the mirrorless game where users were initially concerned they weren’t getting what they were promised, but things seemed to have sorted themselves out at least a little bit.

Feel free to give us your input, however, and let us know what you think of the numbers Northrup has come up with.

(via Mirrorless Rumors)


Update: Mr. Northrup was kind enough to reply to some of the more common comments in the comments section, but since his reply has gotten buried among the almost 100 comments this post has as of this update, we’re reprinting it here in full.

The guy in the video here, let me address some common comments:

* I know it’s long, but please do watch the entire video before accusing me of being wrong.

* With that said, I do have a few minor corrections that I made with YouTube annotations (which might not be displayed on all clients): 18mm should have been 16mm on one of the slides (and there was another typo that I forget), and Canon, Nikon, & Fuji have also abused crop factor… but not quite as blatantly. Oh, I use ‘bokeh’ very casually, and I’m aware of the technical definition, but you know what I meant, and that’s what counts. To the best of my knowledge, those are my only mistakes, and I’ve done my best to verify hundreds of comments.

* If you read the Admiring Light article, and watch my entire video, you’ll see that we totally agree with each other… to such an extent that I feel like I have to say I didn’t see the Admiring Light article before making my video. I did my own primary research and came to the same conclusions. Note that the article goes into detail about how sensor technologies differ, and my video is a bit more theoretical, but I specifically say, “given similar sensor technology” to address the concern. If you think this article disagrees with my video in a fundamental way, than you probably misunderstood one of us.

* The Mirrorless Rumors guy says that if you get the same brightness photo using the same settings on two cameras with different sized sensors, than it proves I’m wrong. I think this just proves that he didn’t watch my video, because I actually show the results of that exact test multiple times (and of course the brightness is the same). The fact that those settings produce images with the same brightness is the basis for my entire argument.

* I’m not discussing per-pixel noise, I’m discussing per-image noise. An A7R and A7S both have the same per-image noise, as will film or your eye. Different pixel densities will yield different per-pixel noise, yes, but that’s simply not the topic of the video. At several people’s request, I am planning a video to discuss the merits of pixel density. It’s an interesting subject; it’s just a different subject.

* The “35mm camera makers are cheating us because of something with medium format” argument: Nobody labels non-medium format lenses with medium-format equivalents. My math holds true for medium format systems, which have a crop factor < 1.

I'll check back and I'm happy to answer any questions! Thanks for watching!


 
Get the hottest photo stories delivered to your inbox.
Get a daily digest of the latest headlines:
  • Joey Miller

    Just a preliminary observation: his ISO crop factor thing doesn’t scale to medium format, or at least he hasn’t shown it to be true there. I think he’s confusing sensor size with pixel size.

  • SAR_Admin

    Just do a simple test:

    1) take a picture with you Full Frame camera with the same FF lens (let’s say a 50mm f/1.8). let’s say in automatic it takes it at ISO 100, f/1.8 and 1/250 shutter speed.

    2) Put the same lens on a APS-C camera. Take same picture in automatic and the camera will take it at exactly the same speed, ISO and aperture.

    Voilà. Tony was wrong :)
    Of course the FF image is better because of the larger pixel size (if both sensors have same total pixela mmount). But that’s another story that nothing has to do qith equivalence on lenses.

  • Allen

    Stopped watching when he pronounce ISO as “eye-so” and bokeh and “bouquet.”

  • Joey Miller

    Also, aperture has less to do with depth of field than sensor or film size. Aperture is the amount of light coming in. That’s calculable. Maybe Tony needs to learn about circles of confusion, unless he wants to assert that lens manufacturers have been lying to us since the beginning of lens manufacture. I’m sure MF and LF film shooters aren’t pissed off that f/2.8 looks different on their respective formats than it does on any of these smaller formats.

  • BE426

    He’s an Idiot…. Aperture is determined by the ratio of focal length to lens diameter… So f1 would be a case where the focal length is equal to the diameter of the aperture blades in their widest settings. This is a physical principle of the lenses… by the same argument he’s making you could say that a 90mm f4 Lens on a 4×5 Camera is a 24mm f1.1 uts bnit f2.8 is f2.8, f4 is f4 end of story.

    Where he’s getting confused is in apparent field of view: a 25mm 1.4 M43rds lens gives you a similar perspective and DOF as a 50mm lens at 2.8… which is the same as if you did a 50mm crop on a 25mm lens.

  • Zos Xavius

    He conveniently ignores pixel pitch which has a much greater effect on noise than sensor size. He also cannot seem to wrap his head around the fact that f2.8 is f2.8 regardless of sensor format. The only thing that is affected is dof and not exposure settings. I don’t think he really understands fstop, and the idea of throwing away ISO as a standard is really retarded. Yeah manufacturers (looking right at you fuji) lie and cheat, but it still is the closest thing we have to an equivalent between cameras. I like how he keeps claiming that film somehow works differently than digital. They are both light sensitive mediums. No difference really other than how they arrive to similar conclusions. ASA400 film should be the same as a digital sensor at ISO400. Of course not all things are equal, but in a perfect world this would be true. FF, like medium format, will always win in terms of resolving power in the end. So to keep claiming that you can take pictures with FF quality on a m4/3 sensor is misleading at best. Of course this “guru” wants to sell you something. My advice: don’t pay a dime for his drivel.

  • Rob S

    I was with you until you said the FF image was “better.’ I think “different” is a….better term.

  • sdq

    are you saying that ff camera has the same result as apsc camera at the same ISO rating? if so, why would anyone buy ff camera?

    you know whats equivalent setting of apsc camera to ff camera? its about 1stop less light.

    in order to bring back the performance of 1 stop you need to decrease ISO, right? but decreasing 1 ISO stop would make image darker. So only thing you can do is to increase the aperture. And to increase the aperture by 1 stop, will almost double the size of the lens.

    Suddenly, your apsc lens now has the same dimension as ff lens. And this is physic and math that you cannot change and also the point of his video

  • Johan Robertsson

    Actually you’re wrong. Yes both of the images are as bright, but the APS-C image is noisier, even at ISO100 the fullframe one will be cleaner. This is because the ISO100 on APS-C is actually closer to ISO320 on fullframe. Then you ask, why would you need to boost the ISO on APS-C to get an equivalent exposure if you’re using the same optics. Because you’re throwing a large part of the incoming light away with APS-C because it doesn’t cover as large part of the image circle as fullframe. What this really means essentially that you get worse signal to noise ratio on the APS-C with the same optics which is why the image is noisier. But in truth this has more to do with the size of the pixels, but the effect is the same.

  • Robert Mark

    He didn’t need anywhere close to 40 minutes to convince me that he doesn’t know what he is talking about.

  • Jenny Wong

    Honestly, who cares? just go out and take great photos with whatever camera/gear you got.

  • SAR_Admin

    Of course not! This is because the FF has a larger sensor with larger pixels. That’s why Image Quality is usually better on FF cameras.

  • SAR_Admin

    NOpe I am not wrong. Both images are as bright like you said. But obviously the pixels of the FF sensor are usually much larger. Therefore Signal to noise ratio is much better on FF. But that has nothing to do with focal length or aperture equivalence between formats!

  • SAR_Admin

    S/N ratio is better on larger pixel sensors (on sensor with same tech).

  • OtterMatt

    I stopped reading when you made it apparent that you’re stuffed up your own ass.
    It’s a Japanese word. That’s how it is SUPPOSED to be pronounced.

  • spiralphoto

    No, actually, no it’s not. It’s prounced “bo-keh”, not “boo-kay”.

  • OtterMatt

    Man, people are awfully quick to disregard basic physics in favor of… um… I guess the assumptions of their own massive space-brains. Myself, not beholden to such a gargantuan logical powerhouse in my head will just trust that physics and math are still right after the last few thousand years.

  • Pitergabriel

    I agree!!

  • Jack McKechnie

    Nikon took a nose dive this morning in their financial report..I don’t know how Canon is doing but the morrorless companies are getting stronger and stronger….I have a camera with an optical viewfinder and a mirrorless and honestly I love the images,the freedom in lenses and most importantly the size…I find myself using my mirrorless more and more…and I can’t see reversing that in the near future!

  • http://www.bobcooleyphoto.com/ bob cooley

    Actually, both seem to be acceptable:
    http://en.wikipedia.org/wiki/Bokeh

  • Marcin

    Actually, both guys (Northrup and Ale) say the same thing, i.e. they agree on the effects of different sensor sizes and the corresponding lenses. They just use different nomenclature (accepting or rejecting the term “equivalent apertures”). A storm in a teacup.

  • Marcin

    Wow, what a brilliant piece of argumentative writing.

  • Frank Martinez

    Nailed it.

  • Robert Mark

    Not trying to be argumentative. Facts is facts — the other poster is correct: F1.8 on a Nikon 50mm = F1.8 on any other lens. Yes, depth of field is different. Some people prefer greater depth of field, which would seem to negate 39 minutes of his video.

  • sdq

    larger pixel cannot accept more light if the lens doesnt pass enough light.

    i guess your explanation can be rewrite to make it correct too.

    at the same shutter speed, ISO and aperture. both camera will have the same output. but why ff lens is bigger? because it passes more light to fill in those larger pixel to make a good quality image.

    so in order for crop camera to have the same quality, it needs the same amount of light passing through. this is results in using bigger aperture and lower ISO setting.

  • MIchael Casimir

    Actually Tony was right.. Smaller sensors capture less light since their surface area is smaller. Therefore to compensate the manufacturer sets ISO 100 to be more sensitive than ISO 100 in full frame. Therefore the test will come up showing the same results but the image will be noisier as smaller sensors are more sensitive to light at ISO100 than full frame which in turn produces a lower signal to noise ratio.

  • http://www.imajez.com imajez

    Easy to tell it’s not a Japanese word as the only consonant that can be used at the end is ‘n’. :)
    Adding a ‘k’ to the end means English speakers are more likely to pronounce it like the original Japanese word.

  • http://www.imajez.com imajez

    I know English spelling bears little resemblance to pronunciation, but boo-kay seems like a way to make the word seem even more pretentious.

  • Elisabeth Bjorsta

    Being half japanese Im offended that you guys even use the word, why not just stick to english.

  • Shawn Aikey

    I’ve done a similar test… only I kept the APS-C camera body constant and changed between a full frame lens and a lens designed for APS-C. I had the camera in auto-ISO and used a fixed shutter speed, then adjusted aperture until ISO was the same using both lenses.
    Guess what, despite everyone saying aperture is aperture, I had to use F2.8 with the full frame lens and f4 with the crop lens to keep the ISO the same. And the histograms looked very comparable as well.

    Granted I’ve only done this once, but I do plan to investigate this further, because every article I’ve read says this shouldn’t be the case.

  • Adrian S

    I think he is assuming sensors have the same resolution.

  • ChrisPerc

    I found Tony’s discussion is totally correct and really clarifies the issues with regard to aperture/crop cameras for me. His main issue is very easy to verify as he did in his video; you don’t get the same result with a 12-40 f/2.8 on MFT as you do with a 24-70 f2/8 on FF. I know, I went from a 5D to Oly MFT (with no regrets since I understand these issues).

    Using sensor size is not a confusion. He’s considering overall image S/N, IMO, the proper consideration. When viewing an image, you don’t see noise in the individual pixels, but in the overall image.

  • ChrisPerc

    Wrong reason to stop viewing. He explains ISO, and ‘Americans’ all have to say “bouquet” ;-). He understands the technololgy very well.

  • Lukas Prochazka

    I qiuet didnt get it with ISO. In my mind it’s like when some pixel receive a same amout of brigtness the pixel can produce same amout of sygnal no matter if besides are 2 pixels or 200…I think the issue is with the size of each pixel and because they differ at full-fram and 3/4 ..pixels are smalles so they can produce less signal so they more noisier. I think if we had 4 pixels put on full frame senzer at ISO 100 it would produce same amout of noise as 1 pixel pu on 3/4 senzor at 100 ISO…pixel stay same size and also brigtness….I am right?

  • https://plus.google.com/115197896619361127605/posts Aaron W

    I checked out when he got through the entire 13-minute section on ISO without mentioning how pixel count/photosite size affects sensitivity and noise. He didn’t disclose the pixel density of the sensors he used in the demo, which makes the whole thing a pointless discussion, and starts to smell like a smarmy infomercial or used car sales pitch.

    Or to try out a thought exercise — a theoretical full-frame 400 megapixel sensor will be MUCH noisier than a theoretical 2 megapixel MFT sensor if you point them at the same grey card, because you’ll have to crank the ISO on FF sensor to capture the same number of photons per pixel.

    His totally made-up equation involving the square of the crop factor makes no sense when you have different pixel counts involved.

    Alternately: Look at the low-light performance of the Sony A7s (12 MP) vs. the A7r (36 MP). They both have full-frame sensors, but the A7s has much lower noise at high ISO because the photosites are roughly 70% bigger.

  • sshoihet

    The density of light that falls on the sensor at the same f-ratio is not affected with different sensor sizes but the total amount of light collected is affected. A full frame sensor collects about 4x more light at f/2 than a m4/3 sensor does.

    ISO says nothing about noise, it only indicates sensitivity to light. At a given ISO for similar technology, noise will relate to the difference in area; if you double the sensor area you will halve the noise.

  • Ridgecity

    some of the “debates” are really weird here…

  • ChrisPerc

    Huh?

  • sshoihet

    You forgot to account for noise in your little experiment. The brightness of the image is the same but the APS-C camera has twice the noise.

    F-ratio is the density of light that falls on the sensor; at the same f-ratio, the FF sensor collected twice as much light as the APS-C sensor.

    Assuming everything else, sensor technology, QE, etc is similar of course…

  • http://www.bobcooleyphoto.com/ bob cooley

    Because it’s a lot easier than saying “the aesthetic quality of the blur produced in the out-of-focus parts of an image produced by a lens” multiple times in a paragraph… :)

    Besides, Bokeh is not a Japanese word, it comes from the Japanese root ‘boke’ – if you are going to be offended by every English word that has a foreign root, you might as well stop speaking the language, as English is rooted in Latin, German, Japanese, French, Spanish, etc.

    This has to be silliest thing I’ve ever seen anyone be offended by, and we see some pretty silly things written on PP…

  • sshoihet

    Then maybe you’d care to explain the decrease in noise between the D800 and D7000 which have the same size pixels or even the decrease in performance for the same sensor in DX vx FX mode? It has everything to do with sensor area and equivalent aperture and nothing to do with pixel size.

  • http://www.bobcooleyphoto.com/ bob cooley

    Not my preference, either – I say Bo-kah; but tomato, tomato (that really doesn’t work when you spell it, does it) ;)

  • https://plus.google.com/115197896619361127605/posts Aaron W

    Not true at all. The APS-C images *might* be noisier because he’s using higher pixel counts on the APS-C cameras. He doesn’t disclose it, so there’s no way to know. ISO 400 is ISO 400 — it measures light sensitivity per square inch. Sensors with larger photosites (and thus lower pixel counts) will have better performance at high ISO because each pixel captures more photons. By ignoring pixel density, his entire ISO argument is pointless.

    (Exhibit A: Sony A7s vs. Sony A7r — identical size sensors with vastly different low-light performance because the A7s has photosites that are ~70% bigger.)

  • Lukas Prochazka

    but what is catching light is pixels not the sensor…if one pixel would make entire sensor it would capture x amout of light but when you would make 2 pixels from that one on same area the amout of light caputured by one pixel would be x/2 …that means if you cover 1 pixel the other is still geting same amout of light x/2…..well I dont know this seems just logic to me..

  • GeniusUnleashed

    That’s cool. But unless you’re other half is Mexican, you’re no longer allowed to say guacamole.

  • Lukas Prochazka

    yes but his point I think was yes the appature stays but appature is just that matematical expression…and not that appature is a way how to calculate dof, because that changes with sensor size ..

  • GeniusUnleashed

    And the fact that THAT offends you, means you have one too many sticks up your arse. I’m half Irish, so I’ll say arse. :{

  • Lukas Prochazka

    yes, but the point was how to deal with eqivalent….1.8 will be 1.8 but the eqvivalt to get same result must change…

  • GeniusUnleashed

    The whole point of the video is that they SHOULD be pissed off because they’re being charged for it. Did you even watch it all the way through?

  • http://www.bobcooleyphoto.com/ bob cooley

    One problem here – all ISOs are not the same. Unlike film, digital camera manufacturers can use a number of algorithms or standards to determine ISO; one commonly used is the “Recommended Exposure Index” which allows for an arbitrary determination of the ISO based on a number of standards, only some of which are determined by physics. Sometimes they equalize a senor’s ISO rating by turning up the gain. Merely cutting down a FF sensor into the size of a m43 sensor, or determining the optical physics between the sensor, lens, etc. doesn’t determine the ISO value – its not done by a strict scientific measurement the way film was.

    Shorter version: manufacturers can label ISO arbitrarily, based on ‘perceived’ ISO standards in comparison to other sensors and brands. There is as much marketing as science in today’s ISO ratings.