On lenses for small cameras: a data-driven counterargument
Andy Westlake of dpreview.com takes apart the current lens offering for lightweight interchangeable lens cameras (LILC) like the micro four thirds and related mirrorless designs, but I was unconvinced. Let's see what the data says.
Andy Westlake is a photographer and camera reviewer at dpreview.com and his opinions carry some weight in the digital photography community. His position, heavily summarized, is that the new type of LILCs such as the micro 4/3 format cameras need to maximize portability to differentiate themselves from standard DSLRs. To do so, they need to be coupled with lightweight, bright prime lenses. Many readers chimed in in the comments and I am not sure I am going to add to that. I just looked at the data. Since it's very hard to match one to one micro 4/3 lenses with equivalent say Canon EF lenses, because of limited available range and several differentiating parameters, I used stats. I started with a google square search for lenses in different mounts (Panasonic micro 4/3, Canon EF and EF-S and Nikon F) and their optical and weight data and I fitted a robust linear model to it (caveat: this is not a random sample and I had to clean the data manually: please let me know if you see something wrong with the data). This is what the model does in plain english: it tries to express weight of the lens as a weighted sum of zoom factor, f-number(s), focal length and mount type. Why weight? Because life is too short, but it would be lovely to try the same for length and diameter, and also add price into the equation. This model turned out to fit the data well enough, with the exception of three huge and bright tele lenses. You can find the details here.
The take home message is that being a zoom adds only about 27 grams per 1X factor increase (measured in degrees, not in mm to avoid any crop factor issues). The standard deviation here is 12 grams, so there is some uncertainty on this number. A typical 3X zoom is expected to add 81 grams to a system compared to a prime of similar aperture and focal length (mid-way in between the extremes of the zoom). If you look at mounts, micro 4/3 is some 260 grams lighter than EF. EF-S is only 90 grams lighter than EF. Another big contribution after the mount is the aperture, a hefty 123 grams per f number. Keeping it constant across focal lengths costs another 118 grams (that is, you can save 118 grams by dropping the aperture at tele by 1, keeping the same at wide). Primes are included in the analysis as 1X zooms with constant max aperture.
So how does this translates into comments to Andy's editorial? I think the data does not support the statement that the portability advantage is limited to the combination of micro 4/3 and prime lenses. There is a 260 grams advantage when keeping other factors equal. Going to a prime might take away another 90 grams, no discussion, but in the big scheme of things it doesn't seem like a lot. Adding a camera into the equation, let's say a Panasonic GF-1 at 285 grams body only vs a Canon 60D at 755 grams it's clear to me that the biggest factor is body weight, followed by mount type followed by f-number and finally zoom range. Is my linear model misleading? Is there a threshold at which even 1 gram makes a huge difference? There is, and it's called pocketability, but we are not there yet. I know because I use one of the smallest LILCs with one of the smallest lenses and there is no way I am going to carry them in my pocket. They might fit some jackets, but it's theoretical. Only if the lens were recessed into the body they would have a chance. Maybe we'll se that one day.
Other assumptions in Andy's analysis are harder to check with hard data. He says serious photographers buying LILCs as second cameras are "half the story", but how about half the market? How many people are going to fork $800 for a GF-1 as a second camera, and buy expensive lenses that they can not share with their main system? And how about people wanting to take pictures of friends and family indoor and with existing light? I am one of those but we are called amateurs, and we have only one system, and many of us use the flash liberally and they are happy with the results. We have one system that has to represent a compromise, which sometimes includes using a zoom and a flash. Sometimes we don't have time to change lens and sometimes we don't own all the lenses to replace a zoom range. Sometimes we shoot stopped down to f-4.5 for a whole sunny day, on the beach, with sand ready to fly onto that delicate CMOS sensor, so a bright prime lens would not make much of a difference, maybe a tad on sharpness if it is high quality. Disclosure: I don't own a zoom and I use the flash sparingly. I hope that Andy is right and camera makers will start catering to people just like me, I am not holding my breath just yet. And I like to throw data behind my opinions.
And to finish on a colorful note, a chart summarizing the data used in this post.
I tried to compress a lot of information in this graphics, so bear with me. On the axes are the min and max field of view for each lens. That puts tele lenses in the lower left corner and wide angles in the upper right. Extreme zooms (wide to tele) are at the bottom right and the diagonal line that you can guess upper left of center has all the primes. The symbol is a square for micro four third lenses, and a circle otherwise. The color is weight, with blue being the lightest and orange the heaviest (I used topo map colors). The symbol size is inversely proportional to the f-number: larger means brighter. I printed a few lens names just as examples, I hope they are readable if you go to the full size version. Despite all my efforts, the results of the linear model are not immediately apparent from the graphic. Your best bet is to focus on the the micro 4/3 lenses and look for comparisons in the surroundings, but there's seldom a closely equivalent lens to compare.
Andy Westlake is a photographer and camera reviewer at dpreview.com and his opinions carry some weight in the digital photography community. His position, heavily summarized, is that the new type of LILCs such as the micro 4/3 format cameras need to maximize portability to differentiate themselves from standard DSLRs. To do so, they need to be coupled with lightweight, bright prime lenses. Many readers chimed in in the comments and I am not sure I am going to add to that. I just looked at the data. Since it's very hard to match one to one micro 4/3 lenses with equivalent say Canon EF lenses, because of limited available range and several differentiating parameters, I used stats. I started with a google square search for lenses in different mounts (Panasonic micro 4/3, Canon EF and EF-S and Nikon F) and their optical and weight data and I fitted a robust linear model to it (caveat: this is not a random sample and I had to clean the data manually: please let me know if you see something wrong with the data). This is what the model does in plain english: it tries to express weight of the lens as a weighted sum of zoom factor, f-number(s), focal length and mount type. Why weight? Because life is too short, but it would be lovely to try the same for length and diameter, and also add price into the equation. This model turned out to fit the data well enough, with the exception of three huge and bright tele lenses. You can find the details here.
The take home message is that being a zoom adds only about 27 grams per 1X factor increase (measured in degrees, not in mm to avoid any crop factor issues). The standard deviation here is 12 grams, so there is some uncertainty on this number. A typical 3X zoom is expected to add 81 grams to a system compared to a prime of similar aperture and focal length (mid-way in between the extremes of the zoom). If you look at mounts, micro 4/3 is some 260 grams lighter than EF. EF-S is only 90 grams lighter than EF. Another big contribution after the mount is the aperture, a hefty 123 grams per f number. Keeping it constant across focal lengths costs another 118 grams (that is, you can save 118 grams by dropping the aperture at tele by 1, keeping the same at wide). Primes are included in the analysis as 1X zooms with constant max aperture.
So how does this translates into comments to Andy's editorial? I think the data does not support the statement that the portability advantage is limited to the combination of micro 4/3 and prime lenses. There is a 260 grams advantage when keeping other factors equal. Going to a prime might take away another 90 grams, no discussion, but in the big scheme of things it doesn't seem like a lot. Adding a camera into the equation, let's say a Panasonic GF-1 at 285 grams body only vs a Canon 60D at 755 grams it's clear to me that the biggest factor is body weight, followed by mount type followed by f-number and finally zoom range. Is my linear model misleading? Is there a threshold at which even 1 gram makes a huge difference? There is, and it's called pocketability, but we are not there yet. I know because I use one of the smallest LILCs with one of the smallest lenses and there is no way I am going to carry them in my pocket. They might fit some jackets, but it's theoretical. Only if the lens were recessed into the body they would have a chance. Maybe we'll se that one day.
Other assumptions in Andy's analysis are harder to check with hard data. He says serious photographers buying LILCs as second cameras are "half the story", but how about half the market? How many people are going to fork $800 for a GF-1 as a second camera, and buy expensive lenses that they can not share with their main system? And how about people wanting to take pictures of friends and family indoor and with existing light? I am one of those but we are called amateurs, and we have only one system, and many of us use the flash liberally and they are happy with the results. We have one system that has to represent a compromise, which sometimes includes using a zoom and a flash. Sometimes we don't have time to change lens and sometimes we don't own all the lenses to replace a zoom range. Sometimes we shoot stopped down to f-4.5 for a whole sunny day, on the beach, with sand ready to fly onto that delicate CMOS sensor, so a bright prime lens would not make much of a difference, maybe a tad on sharpness if it is high quality. Disclosure: I don't own a zoom and I use the flash sparingly. I hope that Andy is right and camera makers will start catering to people just like me, I am not holding my breath just yet. And I like to throw data behind my opinions.
And to finish on a colorful note, a chart summarizing the data used in this post.
 |
| Lens specs chart |