FOV AND LENS RESOLVING POWER WITH NORMAL TO WIDE ANGLE LENSES

Since we have DLSR cameras with three or more sensor sizes there is no longer a 'standard' for a normal FOV focal length lens. The 20D 1.6 crop camera needs a 31mm lens to equal the FOV of a 50mm on full frame camera. The 1D Mark II 1.3 crop camera needs a 38mm lens to equal the FOV of a 50mm on full size sensor camera.

CROP FACTOR SENSORS AND LENS PERFORMANCE
The crop factor, FOV and perspective rules follow each camera through it's entire range of photographic capabilities from wide angle to telephoto. If you want the equivalent of 24mm on a 1.6 crop factor camera you will need a 16mm lens.

Wide angle lenses usually aren't as good optically as normal and telephoto lenses. Optical resolution from the center to edge of the image degrades a little quicker with wide angle lenses than normal and telephoto lenses. Consumer grade wide angle zooms suffer from center to edge resolution loss even more than wide angle primes. This is because of optical design issues with 24X36mm SLR format wide angle lenses.

To some extent these optical design issues can be overcome. However, Canon just hasn't found it necessary from a marketing standpoint to make lenses wider than 35mm that have great resolution center to edge. This situation puts the 1.6 and 1.3 crop size cameras at some disadvantage when it comes to using wide angle lenses to achieve a 50mm equivalent normal FOV or super wide angles to achieve a wide angle FOV. However, this problem is mitigated to some degree by the fact that the cropped sensors use the most central part of the lens's coverage area where it is optically at its best.

IMAGE COMPONENT SIZE ON SENSOR
Another issue that factors into overall image quality is the size of the projected image on the sensor. All lenses have a limited resolving power. The best lenses can resolve about 100 line pairs per mm (lpp/m). Smaller 1.6 and 1.3 crop sensors that require wide angle lenses for a normal FOV and require a super wide angle lens for a true wide angle view are at a disadvantage. The disadvantage is that the wide and super wide angles are projecting smaller image components on the smaller sensor to get the same field of view as the full frame sensor cameras with corresponding FOV lenses. The higher pixel density in the 20D doesn't help when the FOV is the same. Remember, if the output pixel resolution is the same and the FOV is the same, image content will be drawn with the same number of pixels.

With the smaller sensor equipped cameras we actually need wide angle lenses with superior optical performance to resolve the fine details in these smaller projected image components. However, in practice we are getting just the opposite. The wide angle lenses (28-35MM) necessary for the normal FOV on the small sensor are not going to perform optically as well as the 50mm lens on the full frame sensor. And the super wide angle lenses (16-20mm) necessary for a real wide angle view on the small sensor equipped camera are not going to perform optically as well as the 'normal' wide angle (24-35mm) on the full frame camera.

FOV AND LENS RESOLVING POWER WITH NORMAL TO WIDE ANGLE LENSES
Our variables are sensor size, pixel density and lens resolving power. How will these factors affect comparative image quality between these three cameras when using normal to wide angle lenses? Will using the lens's central optical sweet spot on the 1.6 and 1.3 crop sensors overcome the smaller projected image component and lower over all lens resolving power of wide angle lenses? Will the higher pixel density of the 20D be a factor in this photographic scenario? It is time for some more comparison images.

This next series was shot with my Canon 24-70 f2.8 L lens at f8 @ 1/320sec and ISO 100. This is about as good a zoom lens as you can find in Canon's wide to normal zoom lens lineup. I started with 24mm on the 20D and used focal lengths for the 1D Mark II and 1Ds Mark II that would equal that same FOV. The point of focus was the content on the horizon line, close to actual infinity.

edge

With the same FOV and same pixel resolution in the 20D and 1D Mark II we should expect to find that any particular image component will have the same pixel dimensions so pixel density is not going to help the 20D in this case. I believe I got the 20D and 1D Mark II FOV to within about 2% of being the same. If you look closely, you'll see that the 1D Mark II just barely edges out the 20D in image sharpness and contrast. If you are having problems seeing the difference, take a look at the lack of texture in the water in the 20D 'center' example. Since we've got the same number of pixels describing the image components the only thing left to account for this slight image quality improvement in the 1D Mark II is differences in the lens's resolving power and contrast between 24mm and 32mm and the larger size of the projected image components on the 1D Mark II sensor.

It looks to me like my 24-70 just isn't quite as good at 24mm as it is at 32mm for the 1D Mark II and at 42mm for the 1Ds Mark II. In fact, I already knew this to be the case from pervious testing of my 24-70. My 24-70 is at its best optically between 35-55 mm. I have also personally tested my 24-70 against Canon's 24mm f2.8 prime and at f8 couldn't see a bit of difference in center to edge sharpness between those two lenses. If you question how well this lens performs compared to Canon's prime, see this lens review comparing the 24-70 to the 24 f2.8 prime

In review, for any given FOV the bigger sensors will have larger image components projected on them and will require less lens resolving power to get equal detail in the final image. In the examples above of the two 8mp cameras the 20D's 1.6 crop smaller sensor looses out here because of the fact that the lens at 24mm would need to resolve more detail to stay equal with the larger 32mm projected image on the larger sensor in the 1D Mark II. This image series seems to corroborate our earlier remarks that this is usually not possible and certainly not possible with Canon's current wide angle lens lineup. Generally speaking the wider the angle of view of a lens, the harder it is to get good lens resolution and particularly to get equal resolution center to edge.

Optical performance at wide angle and image component size on sensor are two small strikes against the 1.6 crop sensor in this scenario. My experience with Canon's 16-35 f2.8 L lens leads me to believe that had we shot a 3 camera series at 16mm for the 20D, 21mm for the 1D Mark II and 25mm for the 1Ds Mark II we would have seen similar results, if not slightly worse comparative performance from the 20D.

It really isn't fair to compare the 1Ds Mark II in this series. However, it is worth noting that with 16.7mp to describe image detail and tonality and using the lens at 42mm (right in the middle of this lens's sweet spot) the 1Ds Mark II is way ahead in image quality in this series even considering its slight depth of field disadvantage in the image 'center' example of this series.

In the next and last page of this article we will have a look at dynamic range and digital noise.

NEXT

ARTICLE LINKS:
PAGE 1 - INTRODUCTION, TERMINOLOGY EXPLANATIONS AND PIXEL DENSITY WITH TELEPHOTO LENSES
PAGE 2 - DEPTH OF FIELD, PERSPECTIVE AND SAME FOV WITH TELEPHOTO LENSES
PAGE 3 - FOV AND LENS RESOLVING POWER WITH NORMAL TO WIDE ANGLE LENSES
PAGE 4 - DIGITAL NOISE AND DYNAMIC RANGE