Welcome! Log In Create A New Profile
Show your ad here! Mail admin@walletfriendlywatchforum.com for rates

Resurrecting this thread regarding sensor size even though it's been dormant for a few months >>>
avatar May 16, 2018 07:10PM
The physical size of the sensor is only one piece of the equation. The other two are the resolving power of the lens being used, as defined by its modulation transfer function, or MTF, and the number of pixels the sensor contains. I will presume the camera can focus properly and the lens mount is on axis with the sensor, etc.

Larger physical sensor makes life easier regarding system resolution, but only to a point. It also demands larger glass for the same aperture (f-stop) compared to a smaller sensor, which is more difficult to make with high precision compared to smaller diameter lens elements to achieve the same largest aperture (lowest f-stop number), and that can affect the lens MTF (optical resolving power). More pixels in the sensor only works to a point as eventually the pixels are significantly smaller than the resolving power of the lens and nothing further is gained. In the ideal system, the sensor physical size, its total number of pixels (which gives pixel physical size) and the resolving power of the lens are all well-matched. Doesn't mean you'll get high resolution imagery, just that you won't be wasting money on excess sensor resolution, or conversely on excess lens resolving power for the number of pixels and their physical size.

In the era of the "pixel wars" the big thing being touted was the number of pixels with manufacturers touting how many pixels their sensors had. Some of the small inexpensive pocket camera tiny sensors and mediocre lenses have an absurdly high number pixels, well beyond the capability of the optical train of glass in front of it. For all those pixels, many of their owners bemoaned they were still getting photos with crappy resolution. There is also a phenomenon called diffraction limiting with physically small lens apertures. Light diffraction starts to limit resolving power. I won't consider that much here, but it's yet another problem with the bitty little lenses on the tiny pocket cameras. The aperture behind the lens is physically quite small.

Wikipedia entry on MTF (aka Optical Transfer Function)

I have three digital cameras, one of which isn't used much now. Of the other two, one of them is a Sony NEX-7 with a 24.3 Megapixel (6000 x 4000 pixels) APS-C (23.4 × 15.6 mm) CMOS sensor. One might wonder if that's pixel overkill for a sensor that size. Only if you've got crappy glass in front of it. Put glass on that has exceptionally high MTF (high resolving power) across most of its aperture range, and it can fill all 24.3 Megapixels with extremely high res images. Did that with a Carl Zeiss Vario-Tessar T* 16-70mm F/4 (equivalent of 35-105 in 35mm format). As compared to the Sony branded lens it came with, which was good, the Zeiss glass was a quantum leap.

This was taken hand held using the NEX-7 with the Zeiss glass. The main photo has been cropped some from the original 6000x4000 and then resized to 1024 x 682 for upload and storage on a photo hosting site. Made four 150x100 pixel crops from the original showing the detail level the NEX-7 with the Zeiss lens captured. The combination of the sensor and lens are well matched.

The second one is a Sony DSC-RX100 pocket camera which has a non-removable Carl Zeiss Vario-Sonnar T* on it. The sensor is a 20.2 Megapixel (5742 x 3648 pixels) 1-inch type (13.2 × 8.8 mm) CMOS sensor. This photo was taken with one hand while riding a bicycle at about 17mph with a group of other riders. As with the other one I resized it from the original 5472x3648 to 1024 x 682 for upload and storage on a photo hosting site. Made a similar 150 x 100 crop from the original photo to show the detail of the USA flag sock on on the male rider directly in front of me. You can see the tufts of sock yarn depicting the stars on the flag. As with the NEX-7, the lens and sensor are very well matched.

The bottom line is sensor physical size isn't the only important parameter. How many pixels it has and the dimension of one pixel on the sensor coupled with lens optical resolving power as quantified by its MTF are the three things that will determine the level of detail possible by the system. Spent a good number of years with film in 35mm small format and 645 medium format (6cm x 4.5cm frame size using 120 and 220 roll film on spools). The advantage of medium format was clearly similar to having a CMOS sensor that's physically larger. To leverage on that one must excellent glass in front of it with excellent MTF, and high resolution (small grain size) film. The sensors in these two cameras, particularly the RX100, are not physically that large. For the lenses in front of them, they don't need to be physically large, they simply need enough pixels to capture the resolving power of that glass. It's the entire optical train from lens through sensor that produced the results above. Notwithstanding doing both hand held and the latter one-handed while moving 17mph on a bicycle.


Edited 5 time(s). Last edit at 05/20/2018 08:42AM by John Lind.
Subject Author Views Posted
Do you feel satisfied with your camera sensor size? BF Hammer 811 February 11, 2018 01:08PM
Resurrecting this thread regarding sensor size even though it's been dormant for a few months >>> Jpeg Attachments John Lind 484 May 16, 2018 07:10PM
I felt the urge a couple of years ago... Micha 560 February 15, 2018 03:05AM
Sensor low light performance has improved significantly plus >>> John Lind 468 May 23, 2018 04:17AM

Sorry, only registered users may post in this forum.

Click here to login

  Our thanks to our WF-WF Sponsors!
Show your ad here! Mail admin@walletfriendlywatchforum.com for rates Show your ad here! Mail admin@walletfriendlywatchforum.com for rates Show your ad here! Mail admin@walletfriendlywatchforum.com for rates