The upper definiton limit for the four thirds

[frode_inge_helland]

<p>Does somebody know what is the theoretical upper definition limit for the four thirds sensor? I guess the optics are very close to their limit already.</p>

[joe_c5]

<p>Theoretically four thirds cameras could be diffraction limited.</p>

<p>Because the sensor is half the dimensions of a 35mm camera, diffraction will be equal to that of a 35mm camera at apertures two f-stops smaller. For example, f/22 on a 35mm camera will look like f/11 on four thirds.</p>

[frode_inge_helland]

<p>I seems that I have to confess two personal shortcomings. The first is that I do not understand the ansver. Diffraction is an optical phrase and is closely connected to the wavelengths of visible ligth (e.g. microscopes). This can be expressed in resolution of the lens eg. linepairs per mm or inch.<br>

The second shortcoming is that I may have expressed myself clumsy. What I ment was: How many megapixels is it theoretically possible to put into a four thirds sensor. I have heard the present sensors are far from that limit. Then, how many megapixels can the best present lenses make use of. I personally do not think it is possible to push the optics very much further. The optical technich has been around for hundreds of years and developed slowly. The use of computers made it possible to calculate complex constructions as extreme wide angles and zooms.<br>

The sensor technology is a newly born baby one can expect to develope considerably but less in sharpness and contrast. The pushing of megapixels in a sensor is nevertheless of limited use if the optics cannot follow.<br>

But where is the limit?</p>

[frode_inge_helland]

<p>The end oft he second last sentence ha jumped from a sentence far above by accident.</p>

[joe_c5]

<p>Ok, to elaborate, for any particular lens aperture there is a smallest spot size (called an Airy disc) that can ever be focused by the lens due to diffraction. The smaller the aperture, the larger the spot.</p>

<p>At f/1.4, the Airy disc is roughly 1.9 microns in diameter. It might take as many as 400 megapixels before increasing the number of pixels stopped increasing the effective resolution. For a more realistic example, at f/5.6 the Airy disc is roughly 7.5 microns in diameter. Now 25 megapixels is roughly the point where adding more pixels no longer helps.</p>

<p>At the moment the smallest pixels seem to be 1.43 microns wide. A four thirds sensor with that pixel pitch would be nearly 110 megapixels.</p>

<p>The downside to smaller pixels is lower dynamic range due to higher noise. Even with a perfect sensor that added no noise of its own, there would still be photon shot noise.</p>

<p>With actual lenses instead of theoretically perfect lenses, four thirds may start to be diffraction limited at apertures smaller than f/5.6 or f/8. The exact value will vary with the lens. The onset is gradual, too, rather than a specific aperture where diffraction effects suddenly become visible.</p>

[steven_f1]

<blockquote>

<p>How many megapixels is it theoretically possible to put into a four thirds sensor.</p>

</blockquote>

<p>that depends on how samll you can make the pixals. They could probably right now srink the pixals down to 1/4 the current size and therefore achieve a 38MP sensor. But that said no one would want it. With current technology such small pixals produce images with a lot of noise. Maybe in the future with different technology we will exceed 50MP but that will probalby be of little use to many of us. We already have enough pixals to make 11X14 prints or16X20 prints. Most people never make prints that large.</p>

<p>For myself I don't have a need for more than 12 MP right now. However if someone introduced a camera with a dynamic range of 15 to 20 stops I might be one of the first to buy it.</p>

 

<blockquote>

<p>I personally do not think it is possible to push the optics very much further.</p>

</blockquote>

<p>I can agree with you on the lenses. Today lens quality is more about how much you are willing to spend than anything else. Today size, the/weight, max aperture, ability to use filters, zoom/prime, and fucus range (macro/normal) are often more important than the ability of the lens to resolve detail. Yes we can make a lens with much better optics than what is on the market now but if no one can afford it, no one would make it. </p>

[frode_inge_helland]

<p>Thank you very much both of you; Joe and Steven, for clearing up this mist. I am now more than happy than ever for choosing lenses before sensor size. :o)))</p>

[berg_na]

<p>Smaller pixels do not cause higher noise. The lower dynamic range is because the pixel full well capacity is reduced, resulting in lower dynamic range and signal-to-noise ratio. Photon shot noise is unavoidable, but hardly an issue, however, the dark current shot noise may be an issue if not properly corrected. Note that sensors with sub 2-um pixels are typically used in cell phones, not digital still cameras where image quality is more important than size.</p>

<blockquote>

<p>The downside to smaller pixels is lower dynamic range due to higher noise. Even with a perfect sensor that added no noise of its own, there would still be photon shot noise.</p>

</blockquote>

[joe_c5]

<blockquote>

<p>Smaller pixels do not cause higher noise. The lower dynamic range is because the pixel full well capacity is reduced, resulting in lower dynamic range and signal-to-noise ratio.</p>

</blockquote>

<p>This is correct, but since the final image will be scaled so that a well 18% full is somewhere near 18% gray, the noise in the final image will increase. Granted I misspoke, and the higher noise is an effect of the lower dynamic range rather than a cause.</p>

<blockquote>

<p>Photon shot noise is unavoidable, but hardly an issue,</p>

</blockquote>

<p>I disagree with this. Photon shot noise often dominates at higher ISO sensitivity above the darker shadows, which tend to show mostly thermal or even patterned noise. An 18% gray at ISO 1600 might have a read noise of 4 and an electron shot noise of 24, out of 585, for a signal-to-noise ratio of only around 24.</p>

[GerrySiegel]

<p>Lens technology has not reached the limits of what is capable so I question that initial assumption right off, even for the affordable glass, so where do we find that at the boundary line.<br /> Why, I innocently ponder, must sensors, electronic receptors on a chip and processors, instructions on how to store light, now have a <em>theoretical limit.</em><br /> And if,let's say they do, and we can spyglass it, what is the effect of said limit in a practical sense, the sense that matters most...<br /> Meaning this only. If the 4/3 sensor is already capturing HD movies, well, wowzer, that impresses this 4/3 and micro four thirds user who has put some dough into the systems. See the new GH 2 previewed already. No breakthrough imager one will argue. So then, against what industry standard do we seek to know the limit or weigh the offering.<br /> A gedanken ,not a pixelpeeping thinkpiece , no, not in a negative sense at all a fair thinkpiece. May even influence investment in brand X or Y now and then, all else being equal ..which it never is.<br /> Thinking too.Moore's Law was said to be reaching some asymptotic upper limit. Then someone is going ahead and thinking we can use carbon as a substrate, or some other element and push it higher and faster...a weak analogy, but shows the capabilities of pushing a boundary within parameters or constraints. My (devalued lately) two cents...</p>

[profhlynnjones]

<p>Hey folks, just how big does the print have to be? My 8 and 10 MP cameras make prints up to 20x30, every bit as good or better than my 35 film cameras (possible exception, Fujichrome Velvia). </p>

<p>If I want something bigger, buy or rent a 20+MP camera but for real additional quality, a large format camera, 20 to 80 square inches is hard to beat. While I can't prove it to be factual, in a Tony Sweet tutorial, he stated that his LF images well exceeded his 22MP images.</p>

<p>Lynn</p>

<p>Lynn </p>