Luminous Landscape Forum
Equipment & Techniques => Mirrorless Cameras => Topic started by: Mosccol on May 31, 2017, 01:05:18 pm
-
Hello
I've been shooting with APS-C for the past 10 years or so and can get all the image quality that I need by using suitable prime lenses.
However, like many users in this forum, I increasingly 'forget' to pack my SLR on my trips because of weight/bulk etc. I find my telephone is clearly insufficient for my needs so I have bought a second-hand Canon mirrorless to get a feeling about these 'new' cameras' and, assuming I can find one with the right ergonomics, it's a no brainer to move to mirrorless these days.
However I find the choices even more daunting than in the the old days. I guess the first decision I need to make is sensor v lens bulk. It may only be an impression, but when using my daughter's 4:3 Lumix the depth of field is not so great (in fairness she doesn't have the best lens set). The beauty of course is the size and weight of the lenses.
Assuming that, over time, every major manufacturer will have a decent lens line-up, what are would you choose and why?
- Micro 4:3 - tiny lenses
- APS-C - Medium lenses
- Full frame - Any point?
Thanks!
-
You're using the mid-sized choice (APS-C) today and finding it to be too large. Why would you consider full frame unless you went in the Sony mirrorless direction. You'd probably have a smaller body, but lenses would still be too large. Therefore, your question comes down to should I downsize to Micro 4/3? I suggest you rent a high quality Micro 4/3 body and lens or two and test them out yourself. What difference does it make what others think?
-
By "full frame" I think you mean 35mm size sensors.
Since 2000 I've shot DSLRs with APS-C, APS-H and "full frame".
My current stable is a Fuji X-T2 and X-T1 with a number of Fujinon lenses. I'm very pleased with what I can do with this gear. I feel no constraints versus "full frame" as I don't print living room size monster images. I don't show my work in galleries or museums where very large prints are in vogue.
My friend has an Oly OM-D EM1 and hopefully the EM1-Mark II soon. It's a fine camera but there are limitations. I would be reluctant to consider shooting high school Friday night football with one.
If I get to the point where I might need humongous prints I would likely move up to medium format. Perhaps something like the new Fuji GFX.
-
I sold all my DSLR gear years ago. I'd never been really happy with the bulk, weight and in your face attention grabbing qualities of DSLR+lens combinations, so much bigger and heavier than the 35mm film kit I had.
I now have a "FF" Sony A7 and three AF lenses, the 28-70mm kit zoom, the 35mm f2.8 and the 55mm f1.8. With either prime the A7 + lens is a compact package capable of excellent quality. I also use old film era lenses.
I also have a couple of Panasonic MFT RF style cameras with 17, 25 and 45mm f1.8's, 14-45mm kit zoom, 9-18mm, 45-200mm and a 12-35mm f2.8. With any of the primes or the prime sized kit zoom these cameras are small and capable and I'm sure I get better quality than I got from 35mm film and although DoF is a criticism some level at the system I'm quite happy that I can get shallow DoF when I want it, ultra shallow DoF isn't everything, here's a nice little piece on the subject :D
http://theonlinephotographer.typepad.com/the_online_photographer/2012/06/in-defense-of-depth.html
I tend to use the Sony A7 for fun with old lenses and for when I want the best quality I can get. I tend to use MFT for when I want an even more compact system and when I want a faster responding camera than my A7.
If you can manage with the more compact lenses and the relatively leisurely speed of operation I think that the A7 is excellent. I also like the MFT system and I wouldn't spend the money I'd have to spend to build up an equivalent FF A7 lens set to my MFT set.
-
- Micro 4:3 - tiny lenses
- APS-C - Medium lenses
- Full frame - Any point?
I was using m43 (from GH2 to GH3 to E-M1 - only normal bodies with proper grip) and then switched to FF Sony dSLM ... I use only few primes (28/35-50-85-low 100s) and always take only one lens (mounted on a body) with me and relevant m43 bodies are simply not smaller enough to make a difference vs A7R2 ... in either case I take a shoulder bag for camera+lens ... or I am not taking camera at all... and more often than not I am making photos where I will use a strobe or two, so camera/lens is simply a minor addition to the rest of things
PS: (28/35-50-85-low 100s) in my case currently are FE35/2.8, FE55/1.8, FE85/1.8, Sigma MC11 + Canon EF 100/2.0
-
if you're satisfied with APSC then you should be (in my experience) equally happy with the latest 16 and 20mpix micro 43 - so long as you use primes and the high quality zooms. the space and weight saving with a kit of lenses is very substantial (amazingly so with lenses like the Pany 12-35 f2.8 and 35-100 f2.8).
-
When I transitioned from FF DSLR to MILC, I tried m43, Fuji APSC, and finally settled with Sony A7. For what I shoot (landscapes and travel), Sony gives me the quality I was used to (EOS 6D plus Zeiss 21, for example), in a much smaller package (A7 plus Loxia 21, for example).
In terms of lenses, I think that Fuji X and Sony A7 are the systems to go to. Not to say that m43 does not have a full complement of lenses (it does), but I think it is an overpriced system; once you go to the better cameras and lenses, I can't rule out that I would be paying a lot for a system that has a sensor the 1/4 size of FF, and for expensive f1.2 lenses if I want shallow DOF.
-
I think your needs are the determining factor. In my experience, FF Sony gives you a RAW image that has a lot of room to correct errors, be it shadow detail or whatever. That said, I have Oly m 4/3 and they are fun and always the one I pick up for walkabout.
-
You're using the mid-sized choice (APS-C) today and finding it to be too large. Why would you consider full frame unless you went in the Sony mirrorless direction. You'd probably have a smaller body, but lenses would still be too large. Therefore, your question comes down to should I downsize to Micro 4/3? I suggest you rent a high quality Micro 4/3 body and lens or two and test them out yourself. What difference does it make what others think?
Sorry if my question wasn't clear. I'm happy with the small Canon M toy I have purchased, but it has pretty lousy glass. Obviously I can use my L glass but that defeats the purpose. I can either sell it or use it special circumstances.
If I buy a full frame mirrorless camera, are the lenses that much smaller/lighter? (at equal quality) It seems that lens size decreases with sensor size, hence the tenor of my question.
-
although DoF is a criticism some level at the system I'm quite happy that I can get shallow DoF when I want it, ultra shallow DoF isn't everything, here's a nice little piece on the subject :D
http://theonlinephotographer.typepad.com/the_online_photographer/2012/06/in-defense-of-depth.html
Great article. I can't speak about very large sensors (anyway it seems that everything in a studio is shot at f8...) I find that APS-C is a decent compromise. You can regain deep DoF by playing with the ISO when needed of course. I guess I need to broaden my experience by borrowing/renting proper glass with 4:3 to get a more informed opinion....
-
If I buy a full frame mirrorless camera, are the lenses that much smaller/lighter? (at equal quality) It seems that lens size decreases with sensor size, hence the tenor of my question.
I see no reason why comparable lenses would be smaller for a larger format, except maybe if you are comparing lenses designed for mirrorless system in the larger format to SLR lenses in the smaller format (or worse still using those SLR lenses plus an adaptor on a mirrorless body). Otherwise, it will if anything be slightly the other way: the larger format needing lenses of proportionately longer focal length, balanced against being able to use lenses of somewhat higher minimum f-top to get comparable DOF wide open and so on, if such lenses exist for the larger format.
Since you seem to care about getting sufficiently little DOF, note that the minimum f-stop needed to get equal DOF wide-open scales in proportion to format size and the focal length needed for equal DOF, such as f/4 on your Canon EF-S gear comparable to about f/6 in 35mm and f/3 in Four Thirds.
-
If I buy a full frame mirrorless camera, are the lenses that much smaller/lighter? (at equal quality) It seems that lens size decreases with sensor size, hence the tenor of my question.
Lenses for full frame mirrorless, ie, Sony A7 camera, are the same size as lenses for full frame Nikon or Canon or any other brand. Check out the weight, length of various camera/lens options at this site:
http://camerasize.com
Brad
-
Re: lens size comparison, here are three cameras with "comparable" lenses. Left to right, full frame Canon 5D Mk IV w/70-200 f2.8, wgt = 2290 g; full frame mirrorless Sony A7RII w/70-200 f2.8, wgt= 2105g; Micro 4/3 Oly OM-D1 w/Pana 35-100 f2.8, wgt= 857g.
Note that the Sony mirrorless full frame body is much smaller than the Canon, but overall size and weight between these 2 are more or less the same. But look at both the size and weight reduction by going to Micro 4/3. The much smaller Micro 4/3 lenses are what really make the difference. Smaller lens design to cover smaller sensor. I left a Canon APS-C system for the Oly and am very happy. 80% of happiness is size/weight. 20% is improved image quality.
Brad
-
For me... the sheer size of M43 with lenses factored in with the quality the em1mkII delivers... it just no longer find a reason for anything bigger. I do use medium format...and lately even that has been a lot less because of the new 20mp sensors. Few clients NEED the size of MF. It only gets used when I need true camera movements, leaf shutters, or insane detail at a high MP for specific subject matter.
The size and quality won me over. The trade offs are few from what i've seen in the way that I work, and the images that I prefer to make. And the close focus distance of almost the entire olympus lens lineup beat out most others.
-
Re: lens size comparison, here are three cameras with "comparable" lenses. Left to right, full frame Canon 5D Mk IV w/70-200 f2.8, wgt = 2290 g; full frame mirrorless Sony A7RII w/70-200 f2.8, wgt= 2105g; Micro 4/3 Oly OM-D1 w/Pana 35-100 f2.8, wgt= 857g.
Note that the Sony mirrorless full frame body is much smaller than the Canon, but overall size and weight between these 2 are more or less the same. But look at both the size and weight reduction by going to Micro 4/3. The much smaller Micro 4/3 lenses are what really make the difference. Smaller lens design to cover smaller sensor. I left a Canon APS-C system for the Oly and am very happy. 80% of happiness is size/weight. 20% is improved image quality.
Brad
For the same DOF, light on the sensor, etc, that MFT lens would have to be an f/1.4.
Jim
-
Jim,
true IF that very shallow DOF is a goal, rather than just an unfortunate side effect of the need for adequate shutter speed with the far more limited ISO speeds of the film era — which I suspect was the most common reason for using such bulky lenses back then!
For those of us for whom the DOF options and the lens and sensor speed combinations of formats like Four Thirds are quite adequate, a more relevant comparison might be 35-100/2.8 vs 70-200/5.6: the latter lens or the closest available approximation will be smaller than those 70-200/2.8 options, but the kit is still likely to be somewhat bulkier and distinctly more expensive than the MFT kit.
-
For the same DOF, light on the sensor, etc, that MFT lens would have to be an f/1.4.
For the same amount of total collected light and DOF, a M4/3 sensor needs two extra aperture stops over a FF sensor, that is right. But in that case, noise statistics dictate that the M4/3 sensor will double the SNR over the FF sensor, so the comparison at equal collected photons and DOF (2 extra aperture stops in the M4/3) is not fair either.
Same DOF and collected light:
(http://www.guillermoluijk.com/article/ffmyth/mazinger2de3.png)
(http://www.guillermoluijk.com/article/ffmyth/mazinger3de3.png)
My PEN beats my A7 II.
Regards.
-
For the same DOF, light on the sensor, etc, that MFT lens would have to be an f/1.4.
Jim
Jim,
You're technically correct. But I put "....comparable lenses" in quotes in recognition of your point. I think that most (yes, I know.....not all) photographers looking at quality equipment and results who are comparing these systems for lenses would look first at comparable field of view. Then they'd consider the fastest lens in the high end series of lenses for that body with that field of view. That is what I did.
Brad
-
I think one thing worth mentioning when talking about MFT is that some of the lenses are perfectly good from wide open, I don't think twice about using any of my MFT lenses from wide open. Thinking back to my SLR/DSLR days that wasn't always the case,
-
FF lenses are larger and heavier no matter what body they are in, so even FF-mirrorless would not be a good option if you are trying to save on weight.
Take a moment to consider the 1" sensor cameras. The Sony RX10iii has excellent Zeiss glass, is a 24-600mm equivalent at f/2.4-4. The DR is as good as any Olympus or Panasonic M4/3 and you never need to change lenses. I carry it around all day on my shoulder and in hand and never think twice. DoF at f/5.6 is the same as f/15 on FF, so even hand holding is much easier.
Almost everything I've shot in the last year has been with it, so have a look at my Flickr page to see what's possible. And they print to 13x19, no problem.
More on the Sony experience (https://luminous-landscape.com/sony-rx10iii-ultimate-travel-camera/) here, too.
-
I recently upgraded my Sony RX10 to the RX10III, and my experience perfectly matches Terry's. I don't miss my Canon 5DII gear at all, and I don't miss having to clean the sensor after almost every photo shoot.
-
Lenses for full frame mirrorless, ie, Sony A7 camera, are the same size as lenses for full frame Nikon or Canon or any other brand. Check out the weight, length of various camera/lens options at this site:
http://camerasize.com
Brad
Depends on the lens. With a Sony system you can get a very nice compact kit if you want. For example the Sony 28 and 35 lenses are very compact as well as the 55. The Zeiss Batis line of lenses are relatively small and deliver exceptional quality. Sure the GM lenses are large, but if you want compact they are available.
I find it funny that every time people bring up the big and heavy GM lenses when making a point that Sony lenses are no smaller than DSLR lenses and totally ignore the ones that are smaller.
-
The combo I'm finding most rewarding in terms of quality v. size/weight is the Sony a7r, Kolarivision Ultra Thin cover glass replacement, and Leica M-Mount optics. The widest, most symmetrical M-mount optics are still unduly affected by the cover glass, as they are on the M9. However, some, like the Voigtlander 35mm Ultron are amazingly good -- better than the OEM Sony with the thicker cover glass. The Voigtlander VM-E adapter can be used to precisely set the infinity stop, which is important to me for fast, hand held, dual-focus landscape shots. Note that the infinity stop is somewhat unique to each lens (or ray angle) where there is a cover glass, but the VM-E is easy to put a calibration system/scale on so that I just adjust it to the indicated mark when a lens is changed.
Paul
www.PaulRoark.com
-
Lenses for full frame mirrorless, ie, Sony A7 camera, are the same size as lenses for full frame Nikon or Canon or any other brand. Check out the weight, length of various camera/lens options at this site:
http://camerasize.com
Brad
What a great site! :)
-
Take a moment to consider the 1" sensor cameras. The Sony RX10iii has excellent Zeiss glass, is a 24-600mm equivalent at f/2.4-4. The DR is as good as any Olympus or Panasonic M4/3 and you never need to change lenses. I carry it around all day on my shoulder and in hand and never think twice. DoF at f/5.6 is the same as f/15 on FF, so even hand holding is much easier.
Almost everything I've shot in the last year has been with it, so have a look at my Flickr page to see what's possible. And they print to 13x19, no problem.
More on the Sony experience (https://luminous-landscape.com/sony-rx10iii-ultimate-travel-camera/) here, too.
Thanks Terry. Great LuLa article too
-
Jim,
true IF that very shallow DOF is a goal, rather than just an unfortunate side effect of the need for adequate shutter speed with the far more limited ISO speeds of the film era — which I suspect was the most common reason for using such bulky lenses back then!
Another way to say what Jim said is that equivalent images could be produced with the Sony lens having a minimum aperture of f/5.6. How small would that lens be? It turns out that it could be about the same size as the 4:3 lens.
As a matter of fact if one compares a larger format to a smaller one the main difference is that the larger format is not necessarily better but it gives the photographer more options, and many of us like to have options (e.g. to have an equivalent f/5.6 lens, or if we wish to have an equivalent f/1.4 lens). If you don't need options then stick with the minimum that will do what you need.
Jack
-
Another way to say what Jim said is that equivalent images could be produced with the Sony lens having a minimum aperture of f/5.6.
Indeed: perhaps you missed my second paragraph, where indeed I said "a more relevant comparison might be 35-100/2.8 vs 70-200/5.6". But:
How small would that lens be? It turns out that it could be about the same size as the 4:3 lens.
These "equivalent" lenses (with the same effective aperture diameter, a.k.a. entrance pupil size) are similar in front element size, but typically somewhat longer due to the greater focal length, and as I said there is also likely to be a more expensive body, so it hardly seems the best way to achieve a given DOF goal. The differences become more extreme when one can get the job done in the smaller format with a lens in the sweet spot for good, flexible lens design, like f/2.8-4 or constant f/4 zoom lenses, because then there are often no "equivalent" lenses for the larger format, so one instead is forced to using larger, heavier lenses that are "overkill" for the task.
Add to this the observation of Guillermo Luijk in the post after mine that this "equivalence at equal effective aperture diameter" breaks down in low light situations, due to the generally greater total internally generated noise in a larger sensor. Very rough modeling of electronic noise sources suggests that in terms of low-light (high ISO speed) SNR, the equivalence is about a one stop gain in usable ISO speed per doubling of linear sensor size, so f/2.8 in Four Thirds roughly matches f/4 at twice the focal length in 35mm format, and so with 1.4x larger effective aperture diameter ND twice the effective aperture area. Note that this means that equalizing SNR goes with somewhat shallower DOF in larger format.
Let me float an idea for discussion: for the sake of low DOF and high shutter speed/low light needs, there is value to using a larger format when the smaller format is falling short even with moderately fast lenses like f/2.8 or f/2.8-4 zooms and f/1.4 or faster primes. (Resolution and dynamic range are other factors of course; the ones most relevant to choosing larger formats in the film era.)
-
I guess the first decision I need to make is sensor v lens bulk.
Or, you may want to decide first about the range of focal lengths being of interest.
The balance of sensor size vs. bulk shifts considerable when long focal lengths are required.
Peter
--
-
Indeed: perhaps you missed my second paragraph, where indeed I said "a more relevant comparison might be 35-100/2.8 vs 70-200/5.6". But:These "equivalent" lenses (with the same effective aperture diameter, a.k.a. entrance pupil size) are similar in front element size, but typically somewhat longer due to the greater focal length, and as I said there is also likely to be a more expensive body, so it hardly seems the best way to achieve a given DOF goal. The differences become more extreme when one can get the job done in the smaller format with a lens in the sweet spot for good, flexible lens design, like f/2.8-4 or constant f/4 zoom lenses, because then there are often no "equivalent" lenses for the larger format, so one instead is forced to using larger, heavier lenses that are "overkill" for the task.
Sure Bill, lens availability at either end of the usability spectrum depends on manufacturer's marketing focus. However most folks who do this type of thought experiment do not realize how much marketing has to do with it. If you bring equivalence into the picture at the design stage you get more options to make larger format lenses smaller and lighter than one would guess. One can easily trade off f-number for focal length as demonstrated by the addition of, for instance, a 1.4 teleconverer which transforms a 300mm f/4 lens into 420mm f/5.6 in just 20mm. That's with an outside adaptor, imagine what could be done by a good optics engineer.
Add to this the observation of Guillermo Luijk in the post after mine that this "equivalence at equal effective aperture diameter" breaks down in low light situations, due to the generally greater total internally generated noise in a larger sensor. Very rough modeling of electronic noise sources suggests that in terms of low-light (high ISO speed) SNR, the equivalence is about a one stop gain in usable ISO speed per doubling of linear sensor size, so f/2.8 in Four Thirds roughly matches f/4 at twice the focal length in 35mm format, and so with 1.4x larger effective aperture diameter ND twice the effective aperture area. Note that this means that equalizing SNR goes with somewhat shallower DOF in larger format.
Guillermo's demo, though interesting for what it is, has too many hidden assumptions to be a good reference for this kind of discussion. I do not understand what he means by in that case, noise statistics dictate that the M4/3 sensor will double the SNR over the FF sensor: that's not what I know, all other things equal. What I know is that, in general, if two cameras of different formats are set up equivalently in theory the signal in photoelectrons out of both is pretty well the same, as shown below (http://www.strollswithmydog.com/equivalence-and-equivalent-image-quality-signal/). Forget the Aspect Ratio (AR) and lens transmittance (q*) terms for now, if the signal out of both sensors is the same, they both have the same number of pixels and read noise is the same, why would the SNR be different?
(http://i.imgur.com/Lor8nSB.png)
Jack
-
Guillermo's demo, though interesting for what it is, has too many hidden assumptions to be a good reference for this kind of discussion. I do not understand what he means by in that case, noise statistics dictate that the M4/3 sensor will double the SNR over the FF sensor:
My demo just demonstrates how noise statistics make invalid the FF vs M4/3 equivalence just giving the M4/3 a lens with two extra stops of max aperture. It's a counter example.
For the case you cited: same amount of collected light and same DOF, the smaller sensor provides better SNR than the FF sensor.
Just think of two sensors with the same photosite size (or any sensor vs a crop of irself):
- Two extra stops of light multiply per-pixel SNR in the shadows of the M4/3 sensor by 4.
- The four times extra surface of the FF sensor, once rescaled to the same final pixelcount of the M4/3 just doubles the SNR.
The result is that the M4/3 sensor still doubles the SNR of the FF sensor for the same collected light and DOF.
In a real case like my demo smaller sensors usually have smaller photosites so the advantage is less than one whole stop, but still remains some advantage.
In low light conditions with a minimum DOF requirement, smaller sensors (or crops from a bigger sensor) perform better in SNR at the cost of pixelcount.
I tried to explain it here without much success:
http://forum.luminous-landscape.com/index.php?PHPSESSID=badlu6vmb2c0gv55njahfbanj5&topic=102356.0
Regards
-
My demo just demonstrates how noise statistics make invalid the FF vs M4/3 equivalence just giving the M4/3 a lens with two extra stops of max aperture. It's a counter example.
For the case you cited: same amount of collected light and same DOF, the smaller sensor provides better SNR than the FF sensor.
Hi Guillermo,
I did not cite a specific case, I made a general statement: With an Equivalent setup - leaving aside differences in lens transmission, effective QE and sensor aspect ratio for simplicity - two sensors of different formats will produce exactly the same signal in units of photoelectrons. Therefore, if the Signal is the same and assuming similar read Noise performance, normalized SNR out of the two will be about the same.
Just think of two sensors with the same photosite size (or any sensor vs a crop of irself):
- Two extra stops of light multiply per-pixel SNR in the shadows of the M4/3 sensor by 4.
- The four times extra surface of the FF sensor, once rescaled to the same final pixelcount of the M4/3 just doubles the SNR.
In my opinion it is easier to start by assuming that both sensors are made up of the same number of pixels. All else being equivalent/equal, the FF and mFT pixels will see/collect more or less the same number of photons/e-, as shown in the link in the earlier post. Does that make sense? If so the rest, including situations where the number of pixels is different, follows from there.
Jack
-
In my opinion it is easier to start by assuming that both sensors are made up of the same number of pixels. All else being equivalent/equal, the FF and mFT pixels will see/collect more or less the same number of photons/e-, as shown in the link in the earlier post. Does that make sense? If so the rest, including situations where the number of pixels is different, follows from there.
Hi Jack, I think the opposite because the effect of different photosite sizes is much more complex to model than just changing the pixelcount, which is pure resize noise statistics.
For instance the Canon 7D II is a perfect APS crop of the 5DS. If one looks at DxOMark DR and SNR figures for both cameras will find that noise advantage of the FF sensor is just 2/3EV, not 1+1/3EV which is the difference needed in aperture to match DOF.
4dB in SNR (2/3EV eq.):
(http://www.guillermoluijk.com/misc/5ds_vs_7d2_db.png)
2/3EV in DR:
(http://www.guillermoluijk.com/misc/5ds_vs_7d2_ev.png)
My personal cameras (16Mpx E-P5 and 24Mpx A7 II) fall in between the equal photosites and equal pixelcount cases, being closer to your preferred situation, equal pixelcount. Still the M4/3 beats the A7 II in SNR for equal amount of collected photons. A photosite collecting four photons is better than four photosites collecting one photon each.
Regards
-
Here are two cameras with roughly 20MP, roughly Equivalently setup, roughly same number of photons on the image, roughly the same FOV, roughly exposed for same DOF (cheating here, but it works for SNR*): from the second cc24 gray square from left, Em1-II SNR 10.8, D5 SNR = 10.2, compatible with many small differences and non idealities.
(http://i.imgur.com/4FF7DCQ.png)
Similar story and same Exposure with two older cameras with different pixel counts (16 vs 24MP), normalized to the same size by the 'comp' button:
(http://i.imgur.com/78dhmYS.png)
Where is this SNR doubling?
Jack
* These are captures from DPR's dim sudio scene, you can see them there. The Em1s were shot at 1/10s f/5.6, which for SNR purposes is equivalent to 1/40s f/2.8.
-
The SNR doubling is always there when photosites are the same technology, it is dictated by simple statistics. This counterexample alone makes false the argument that "a FF sensor is equivalent to a M4/3 sensor with a lens two stops faster".
Regarding your examples, mixing technologies you'll find all possible comparisons (D5 is the King of high ISO at the cost of base ISO DR). Even M4/3 Sony sensors with better DR than Canon FF sensors. But again this makes false that "a FF sensor is equivalent to a M4/3 sensor with a lens two stops faster".
(http://www.guillermoluijk.com/misc/comp80d_6d.jpg)
Regards
-
I will probably attract a lot of criticism and other comments: But is there not more to this question then DOF, equivalent this and that? I always admired the 120 roll film photos being taken with cheap cameras in the 60s. They had this special appearance I never got with 35 mm. And I did admire Polaroids because they all had this big-format feeling. There is more to large vs small format then just the numbers that have been cited here. It is the "Anmut" of these Images. Same as comparing these tiny Iphone photos to FF and eventually M43. This might all be imagination and not backed up by hard data, or is it not?
-
The SNR doubling is always there when photosites are the same technology, it is dictated by simple statistics. This counterexample alone makes false the argument that "a FF sensor is equivalent to a M4/3 sensor with a lens two stops faster".
Regarding your examples, mixing technologies you'll find all possible comparisons (D5 is the King of high ISO at the cost of base ISO DR). Even M4/3 Sony sensors with better DR than Canon FF sensors. But again this makes false that "a FF sensor is equivalent to a M4/3 sensor with a lens two stops faster".
I have no idea what you mean by this Guillermo, my post was in reference to your comment "in that case, noise statistics dictate that the M4/3 sensor will double the SNR over the FF sensor", that case referring to equal total collected photons and DOF for two different formats. I cannot see where this SNR doubling comes from all else being equal/equivalent. Can you take a look at the physics in the link in my earlier post? Unless you can point to an error there, it says that they would be about the same and that's my experience as also shown by the images in my previous post.
Jack
-
One can easily trade off f-number for focal length as demonstrated by the addition of, for instance, a 1.4 teleconverer which transforms a 300mm f/4 lens into 420mm f/5.6 in just 20mm. That's with an outside adaptor, imagine what could be done by a good optics engineer.
Yes; in fact I have advocated the "teleconverter equivalence" idea, as far as FOV, DOF and photon shot noise go–if one ignores read noise. The idea is that an "equivalent" 70-200/5.6 for 35mm format could be created from a 35-100/2.8 for Four thirds by adding a 2x TC. Then the size difference need not be nearly as much as comparing to a 70-200/2.8, but the TC would add some bulk. And this is somewhat close to how telephoto lens designs work: by definition, a true "telephoto" lens (any that is physically shorter than its focal length) consists of a shorter, brighter lens design at the front with a diverging group ("integrated TC") at the back. Something like "50–140/4 + 1.4TC = 70–200/5.6".
Hence my "somewhat bulkier" comment.
if the signal out of both sensors is the same, they both have the same number of pixels and read noise is the same, why would the SNR be different?
Your formulas indeed show "teleconverter equivalence" if the only noise is photon shot noise, or if there is an equal amount of noise from other sources. But Guillermo's point and mine is that it is unreasonable to assume equal read noise: instead, the greater size of the larger sensor and its electron wells, transistors, ADC units and so on increases electronic noise, and this theoretical observation is supported by Guillermo's experimental result, as well as other "equivalent exposure in different formats" comparisons that I have seen.
-
In fact in my opinion the importance of photon noise is overestimated. No picture gets ruined because of photon noise, everytime noise is a problem there is a read noise problem.
That is why I consider irrelevant the fact that SNR increases by sqrt(2) per exposure stop in well exposed areas where photon noise is dominant, because noise is not a problem for the photographer there. The game is played in read noise dominant areas where SNR improves by a factor of 2 everytime exposure is increased by one stop. That is why four times more surface (SNR differs by a factor of 2) can't compensate for the loss of two stops in aperture (SNR differs by a factor of 4).
Regards
-
Your formulas indeed show "teleconverter equivalence" if the only noise is photon shot noise, or if there is an equal amount of noise from other sources. But Guillermo's point and mine is that it is unreasonable to assume equal read noise: instead, the greater size of the larger sensor and its electron wells, transistors, ADC units and so on increases electronic noise, and this theoretical observation is supported by Guillermo's experimental result, as well as other "equivalent exposure in different formats" comparisons that I have seen.
I see, so we all agree that in an equivalent setup (with same FOV, number of photons collected, DOF, number of pixels etc.) as far as SNR is concerned in the end it mostly boils down to read noise in e- and other non idealities (like differences in eQE, lens transmission etc.). And we also all agree that "The SNR doubling is always there when photosites are the same technology, it is dictated by simple statistics" or that "same amount of collected light and same DOF, the smaller sensor provides better SNR than the FF sensor" makes no sense when different formats are compared with everything else including read noise equal/equivalent.
Once read noise and the non idealities are accounted for, I can see small deviations from equivalent theory and practice in typical mFT vs FF SNR comparisons (examples shown above) but nowhere near approaching 2x with same grade kit of the same generation.
Jack
-
In fact in my opinion the importance of photon noise is overestimated. No picture gets ruined because of photon noise, everytime noise is a problem there is a read noise problem.
Ok, it may have been worth mentioning in your demo and previous posts as quoted to Bill that you were referring specifically and only to read noise differences in a disaster scenario when you suggested "For the same amount of total collected light and DOF, a M4/3 sensor needs two extra aperture stops over a FF sensor, that is right. But in that case, noise statistics dictate that the M4/3 sensor will double the SNR over the FF sensor". That was not obvious to me. And of course it's not true when all else is equal/equivalent - in fact in practice it is mostly not true at all :)
In practice the impact of read noise in this type of comparison is limited for the typical photographer. For instance Bill Claff has a neat way of defining the ratio of maximum to minimum useful/acceptable signal in typical photography. He calls it Photographic Dynamic Range and has calculated this value and much more for a large number of cameras at his site http://www.photonstophotos.net/ (http://www.photonstophotos.net/). That clever metric is based on Bill's realization that most photographers will find image data below a certain SNR unacceptable when viewed in typical conditions. If interested in the subject I encourage folks to read up on its details there.
Using the earlier example of two cameras from roughly the same generation and number of pixels to guide us through this, PDR says that image data with an SNR below 5 would be unacceptable for photographers with 16MP cameras like the EM-1. His fine site also indicates that at ISO 3200 the EM-1 has a read noise of 1.4e- and the D610 at ISO12800 of 2.4e-*. The Signal at which the normalized D610 would hit a SNR of 5 is about 29.7e-, this is the minimum acceptable Signal. As a result of its lower read noise, theory says that at that Signal the EM-1 would instead produce a SNR of 5.28, not substantially different from the D610's.
Theory is confirmed by looking at actual images and measuring SNR off them, as I did in the earlier post**. The EM-1 at ISO3200 1/40s f/5.6 below received two stops more Exposure than the D610 at ISO12800 1/10s f/5.6:
(http://i.imgur.com/78dhmYS.png)
Even in a disaster scenario, with the D610 limited to an SNR of 1, the EM-1 would have an SNR of 1.33 - the difference wholly accounted for by the EM-1's lower read noise and still very, very far from 2x. I would also like to point out that all formats are fast converging to the current read noise limit of just over 1e- per pixel in 'dark' ISOs (see FF cameras released over the last year or so), therefore it's unlikely we are going to see such big differences in disaster scenarios in the future.
Jack
*Normalized to 16mp, but normalization is not a determining factor for this calculation.
** From DPR dim studio scene, 'comp' button pressed, many non idealities and provisos apply.
-
Even in a disaster scenario, with the D610 limited to an SNR of 1, the EM-1 would have an SNR of 1.33 - the difference wholly accounted for by the EM-1's lower read noise and still very, very far from 2x. I would also like to point out that all formats are fast converging to the current read noise limit of just over 1e- per pixel in 'dark' ISOs (see FF cameras released over the last year or so), therefore it's unlikely we are going to see such big differences in disaster scenarios in the future.
Jack, are there MFT cameras available with DR-Pix technology or something similar?
Jim
-
Even in a disaster scenario, with the D610 limited to an SNR of 1, the EM-1 would have an SNR of 1.33 - the difference wholly accounted for by the EM-1's lower read noise and still very, very far from 2x.
The 2x factor appears when the compared cameras have the same photosites, as I already said. The whole demo was aimed at making the "FF is equivalent to M4/3 with two extra stops of exposure" false, as I already said too. In general the smaller sensor (or crop from the big one) will beat the larger sensor in SNR, as my cameras test and your example show.
Regards
-
The 2x factor appears when the compared cameras have the same photosites, as I already said.
If you assume the exact same pixel (size and read noise), say that of the 7DII then at the lowest acceptable shadows by PDR's definition SNR is 4.4 and a 5DS would show 4.7, a ratio of 1.07. In the disaster scenario with the 7DII at SNR =1, the 5DS would be at 1.25.
The whole demo was aimed at making the "FF is equivalent to M4/3 with two extra stops of exposure" false, as I already said too. In general the smaller sensor (or crop from the big one) will beat the larger sensor in SNR, as my cameras test and your example show.
Should we try it with more recent cameras, say EM-1II and a9? The a9 is better, the ratio of the two SNRs of 0.96 at PDR and of 0.78 in the disaster scenario.
In my opinion it is misleading to say that a smaller format will automatically have higher (or 2x) SNR than a larger one. That has virtually nothing to do with format size and all to do with the read noise (and other non-idealities) of the sensors and lenses involved. I'll leave it at that.
Jack
-
Jack, are there MFT cameras available with DR-Pix technology or something similar?
Assuming those steps are it, it looks like the latest generation does Jim. From http://www.photonstophotos.net/Charts/RN_e.htm (http://www.photonstophotos.net/Charts/RN_e.htm)
(http://i.imgur.com/Xq33Q9L.png)
Jack