Which Digital Camera Sensor Size Should I Choose for Sharper Pictures?

[zoobyshoe]

I'm confused about this. I'm looking at two point and shoot cameras. One is 10MP and has a sensor of 1/1.7. The other is 12MP and has a sensor of 1/2.3.

According to this wiki article:

http://en.wikipedia.org/wiki/Image_sensor_format,

the 1/1.7 sensor is physically larger than the 1/2.3 sensor. Yet, the latter has more megapixels.

The 1/1.7 camera has a faster lens; 1.8. It's a 3x zoom. The lens on the 1/2.3 camera is 2.8. It's a 5x zoom. Wide angle in both cases is 28mm.

I'm trying to figure out which should give me sharper pictures.

I'm thinking that the faster lens and larger sensor should, in principle, be the sharpest.

 

[Andy Resnick]

The simple answer is that the pixel size on the 1/2.3 sensor is smaller than the pixel size on the 1/1.7 sensor.

Two differences are: smaller pixels = less signal = more sensor gain required = more noise, and smaller sensor size = smaller field of view.

I estimate the 1/1.7 sensor to have 4 um^2 pixels, while the 1/2.3 sensor has 2 um^2 pixels, so the 1/2.3 sensor has to have at least 2x the gain as the other sensor. Taking the lenses into account, the f/1.8 lens is 1.333 stops faster than the f/2.8, meaning 2.5x as much light is incident on the sensor- now the 1/2.3 sensor-lens combination has to provide 5 times the gain to give you the same signal as the 1/1.7 sensor-lens combination.

Since both lenses are specified as 28mm (most likely that means 'the focal length of this lens provides a field of view equivalent to a 28mm lens with a 35mm format image size'), the actual focal lengths of the lenses are different.

Both cameras will likely give you acceptable images. Both cameras have pixel sizes slightly larger than diffraction-limited airy disks from the lens (again, Bayer filters greatly complicate this issue); if the lens is well-made, image quality will be limited by the amount of light and your technique- in principle, either sensor can provide high-quality 8x10 prints.

Just curious- what are the cameras you are considering?

 

[zoobyshoe]

Thanks very much, Andy!

The two cameras are the Samsung TL500 and the Samsung TL320.

My reason for looking at these two is that they are among the few point-and-shoots that have manual control of shutter speed and aperture capability, and I want a physically small camera that I can always carry in my backpack without much bulk or weight for unanticipated photo ops. I'm looking also at the Nikon Coolpix 310. These come up used on ebay, amazon, craigslist.

I'll be shooting portraits, using a tabletop tripod and the timer for a steady camera.

 

[zoobyshoe]

I should ask you about the Nikon. It's 16MP, but has the smaller sensor, 1/2.33. The lens is 1.8. Do you think the pixel um^2 is too small? I have no idea what "diffraction limited airy disks" are".

 

[Andy Resnick]

"Diffraction limited" is typically a shorthand way of saying "the optical performance of this instrument is limited by fundamental physics, and not by design/manufacturing/assembly error". For monochrome sensors and film, the longer statement can be put into a precise analytic expression, by comparing how a subresolution object is imaged by the system under test to an idealized (aberration-free, perfectly aligned, dispersion-free, etc.) system.

Idealized systems usually image point objects as 'Airy disks' (or 'Airy footballs', if you are concerned with 3-D imaging), and system performance is specified in terms of the hape and spatial/angular extent of the Airy disk- for example, the Rayleigh criterion.

For example, the Airy disk radius of a diffraction-limited system (visible light) is approximately r = 0.6* f/#, where 'r' is in microns and f/# is the f-number of the lens. If your pixel size is much larger than 'r', the sensor limits your system performance, while pixel sizes much smaller than 'r' mean the system is limited by the lens.

It may be tempting to always use the largest possible aperture (smallest f/#), but there are several compelling reasons why this is not always the best strategy. First, lens performance is best if it is slightly stopped down- spherical aberration is maximal at maximum aperture, often resulting in 'halos' around objects. Second, the depth of focus is reduced as the aperture increases, meaning less of your image will be in focus (which can be good or bad, depending on the application). There are online depth of focus calculators that your should spend time getting familiar with.

I am less familiar with what is considered state-of-the-art for sensor gain: companies do a better job every year of making quieter amplifiers and senors, resulting in smaller and faster devices. The whole concept of ISO doesn't translate well to digital, frankly.

As always, the best strategy is to try out the different cameras and see what works best for you.

 

[zoobyshoe]

Thanks very much, Andy!

I find it a bit frustrating that you can't know without physically testing the cameras, but at least I have a better sense of why that's the case.