Part 1, Just the Facts
Author Andrew S, June 2013
 |
| Gasherbrums from Concordia, Pakistan. Original photo on medium format film, Mamiya 7. Scanned with Epson V700 |
Introduction Way back in the days of film, there was a limited selection of sizes from which to choose. Most people used the ubiquitous 35mm film, giving an image size of 24x36mm. Some cameras such as the original Olympus Pen, made half frame images on standard 35mm film. Just at the end of the film era the APS photo system was introduced, but failed to thrive under the onslaught of the digital revolution. This system used a film size smaller than 35mm and produced images a little smaller than half frame. Professional studio workers and a few enthusiast amateurs used medium format rollfilm, giving an image size of 6x4.5, 6x6, 6x7 or 6x9 cm. An even smaller number of hardy souls used large format film usually in the 4x5 inch size. One of the advatages of film in rolls was that dedicated panorama cameras could utilise a super wide image format.
The digital era By my count there are 9 sensor sizes in common use for cameras at the time of writing. Is there any sense in which one of them can be regarded as "ideal" ? Can one predict which, if any, will prevail in future ? These are not entirely frivolous questions particularly for the photographer seeking to make an investment in a system with one or more camera bodies, several lenses and other accessories. Nobody wants to be left holding equipment from a photo system which has failed in the marketplace.
Sensor size comparison Below is a diagram showing comparative sizes and dimensions of sensor sizes. Various medium format sizes, all larger than full frame, are not shown.
Sensor size vs image quality We know that larger sensors deliver better image quality. That is the principal selling point for the larger sizes, which of course are more expensive and presumably deliver more profit margin for the manufacturer and vendor. I have put some numbers on this for consideration. As a proxy for image quality I have used the DXO Mark score [dxomark.com] for the best sensor at each size, published at the time of writing. I know there is some discussion about the validity of the DXO Mark score, but in my experience using many cameras there is broad general agreement between DXO Mark score and real world image quality. In the table below I have compared the DXO Mark score to the area of each sensor. This makes no pretence at being "scientific", in the sense that the ratio doesn't have any specific meaning, but I did the exercise to show that on a score per image area basis, smaller sensors dramatically outperform larger ones. You get a lot more performance per square millimeter from small sensors than large ones.
The table below summarises these relationships.
Sensor Name | Diagonal mm | Area square mm | Camera | Sensor maker | Score | Ratio score/area |
1/2.3" | 7.75 | 29 | Lumix FZ200 | ? | 37 | 1.3 |
1/1.7" | 9.35 | 42 | Nikon P330 | Sony ? | 54 | 1.3 |
2/3" | 11 | 58 | Fuji X10 | Fujitsu ? | 50 | 0.86 |
1"/Nikon CX | 15.9 | 116 | Sony RX100 Nikon 1 V2 | Sony Aptina | 66 50 | 0.57 0.43 |
Micro 4/3 | 21.6 | 225 | Lumix GH3/ Olympus Pens | Sony | 71/72 | 0.32 |
APSC Canon | 27 | 332 | EOS 60D | Canon | 66 | 0.12 |
APSC Nikon | 28 | 367 | Nikon D5200 | Sony | 84 | 0.23 |
Full Frame | 43 | 864 | Nikon D800 | Sony | 96 | 0.11 |
* Sony appears to have the upper hand as sensor maker for consumer cameras at the present time. To illustrate this I included two cameras at the 15.9 mm sensor size. The Sony RX100, presumably using a Sony sensor, is 16 points ahead of the Nikon V2 with Aptina sensor. The Nikon V2 can do a great many things of which the RX100 is incapable but the sensor performance difference is clear, at least at base ISO levels.
* You get a lot more bang for your buck with the smaller sensors.
* The march of progress is bringing the performance of smaller sensors up to the same level as that of larger sensors just a year or few previously. Of course the larger sensors are also improving. This raises the question of "good enough" image quality which I will explore in later articles.
* I hasten to point out that image quality is only one criterion one might use when considering which camera to buy. For instance the Nikon P330 has a very good DXO Mark score [54] equal to some of last year's Micro Four Thirds cameras which use a sensor two size steps larger. However CNet Australia reports that the RAW shot to shot time is 6 seconds. Good Grief. I could operate much faster than that in 1967 with an all manual Pentax Spotmatic DSLR.