snubbr.com

So, does size or spacing matter more...
This is confusing out here in novice-land, but fun..

If I compare sensor size/sensor site totals of the Canon G6 @12mp on a 1/1.7" sensor to a Pentax K100 @ 6mp on a 1.8" sensor, it comes out that with a "ratio" of 16/1..

G9= 1/1.7" sensor equals 44 sq. mm, or, .000037 individual sensor sizeK100 = 1.8" sensor equals 370 sq. mm, or, .000062 individual sensor size.

So does that mean the Pentax has individual sensors that are 16x larger than the G9 sesnors, or, that they are 16x further apart? Since photosites generate their own interference (noise?), then it's better to have them smaller and furtherr apart... or does it matter... or is this a variation on angels dancing on the head of a pin arguments?Brian..

Comments (16)

BLawson wrote:.

This is confusing out here in novice-land, but fun..

If I compare sensor size/sensor site totals of the Canon G6 @12mpon a 1/1.7" sensor to a Pentax K100 @ 6mp on a 1.8" sensor, it comesout that with a "ratio" of 16/1..

G9= 1/1.7" sensor equals 44 sq. mm, or, .000037 individual sensor sizeK100 = 1.8" sensor equals 370 sq. mm, or, .000062 individual sensorsize.

So does that mean the Pentax has individual sensors that are 16xlarger than the G9 sesnors, or, that they are 16x further apart?Since photosites generate their own interference (noise?), then it'sbetter to have them smaller and furtherr apart... or does itmatter... or is this a variation on angels dancing on the head of apin arguments?Brian.

It means 16x larger. It matters very much. It is why DSLR's have radically lower noise than P&S. It is also why full-frame DSLR's have less noise than APS-C, which have less that 4/3...

Comment #1

I don't think I asked my question correctly. My concen, as always, it output..

I can't find the SAME real-world image images (not the controlled lab testing environment) in the dpreview "samples" to compare 6 and 12mp cameras, so does anyone imperically know whether...for an uncropped full frame A3 size print @250 dpi, will fewer pixels (6mp) on a larger sensor (extrapolated), or more pixels (10-12mp) on a smaller sensor with less exterapolation, produce the best resolution print? Is there a formula/algorthym/guide that shows this?Brian..

Comment #2

BLawson wrote:.

I don't think I asked my question correctly. My concen, as always,it output..

I can't find the SAME real-world image images (not the controlledlab testing environment) in the dpreview "samples" to compare 6 and12mp cameras, so does anyone imperically know whether...for anuncropped full frame A3 size print @250 dpi, will fewer pixels (6mp)on a larger sensor (extrapolated), or more pixels (10-12mp) on asmaller sensor with less exterapolation, produce the best resolutionprint? Is there a formula/algorthym/guide that shows this?Brian.

Resolution or output quality? The 12mp camera will resolve more than an otherwise identical 6mp, as long as it's lens is good enough. The thing is that resolution isn't generally a problem; it's noise, noise reduction artifacts, and distortion. Real world, a 6mp DSLR can make an excellent 16x24 print even in moderately poor light. So can a high-end p&s, but only in good light. The megapixels are irrelevant..

Whatever advantage a 12mp p&s may have over a 6mp one is overwhelmingly swamped by the advantages of a DSLR's larger sensor. The most basic 6mp DSLR's out there produce better images than the best p&s, especially under less than ideal light..

You use a p&s for size, cost, convenience, movie modes, etc. You use a DSLR for image quality. I use both for those reasons...

Comment #3

BLawson wrote:.

I can't find the SAME real-world image images (not the controlledlab testing environment) in the dpreview "samples" to compare 6 and12mp cameras, ....

I'm not sure what it is you are asking for here. The whole point of the studio shots is that they are controlled and therefore directly comparable..

You can find samples on other sites, for example the "Far-Field Test" at Imaging resource, example here:http://www.imaging-resource.com/PRODS/E40D/E40DFARI0100.HTMBut as the author says:.

Even though the vagaries of nature mean that notwo shots will ever be directly comparable.

And.

The character of the light is unavoidably going to changequite a bit, depending on the atmospheric humidity andthe time of year. - You thus shouldn't rely on it forabsolute comparisons between cameras, since it'sunlikely that conditions will be identical from one testto the next..

Regards,Peter..

Comment #4

Sherwoodpete wrote:.

I'm not sure what it is you are asking for here. The whole point ofthe studio shots is that they are controlled and therefore directlycomparable..

Again, I didnt' phrase the question well..

Since I won't be shooting in a controlled studio environment, I wanted to compare images of the same subject taken under real world conditions. I was thinking it would be good if dp "samples" for all cameras contained a baseline outdoor mid-day photograph at standard settings, e.g., a building at 2 pm at ISO 200 5.6/250 or something like that.Brian..

Comment #5

The day to day changes in weather and bandwidth needed to serve full size images just doesn't seem to make it sound real useful way to compare..

It'd be easier to borrow the cameras you were interested in comparing or bring a memory card to a camera store and try their demo units and look at the images at home...

Comment #6

BLawson wrote:.

Again, I didnt' phrase the question well.Since I won't be shooting in a controlled studio environment, Iwanted to compare images of the same subject taken under realworld conditions. I was thinking it would be good if dp "samples"for all cameras contained a baseline outdoor mid-day photograph atstandard settings, e.g., a building at 2 pm at ISO 200 5.6/250 orsomething like that.Brian.

Yep, you still didn't. .

In simple terms, a camera with fewer MP per square inch has bigger pixels...NOT dead space between them. That would be a foolish design that would lose detail..

So the bigger pizel camera will have better dynamic range and work better in low light with less noise..

Not much else to say..

Greg..

Comment #7

BLawson wrote:.

This is confusing out here in novice-land, but fun..

If I compare sensor size/sensor site totals of the Canon G6 @12mpon a 1/1.7" sensor to a Pentax K100 @ 6mp on a 1.8" sensor, it comesout that with a "ratio" of 16/1..

G9= 1/1.7" sensor equals 44 sq. mm, or, .000037 individual sensor sizeK100 = 1.8" sensor equals 370 sq. mm, or, .000062 individual sensorsize.

So does that mean the Pentax has individual sensors that are 16xlarger than the G9 sesnors, or, that they are 16x further apart?.

The pixel areas differ by a factor of 16 (so the big pixel gathers 16 times more light for the same exposure time & f-stop.).

The center-center distance of the pixels (pixel pitch) differs by the square root of 16 or a factor of 4. This also implies the Signal to Noise ratio for the big pixel is 4 times better (for the same exposure time & f-stop.).

Ultimate resolution depends in part on pixel pitch. The best sensor resolution possible is twice the pixel pitch but often the lens can't match that depending on quality & f-stop..

The Airy formula relating lens resolution and f-stop is roughly:p~f/3 in microns for greenish light..

For the two cases you mention,p=1.9 microns for the G9 so it is "Diffraction Limited" above about f:5.7p=7.8 microns for the K100D so it is "Diffraction Limited" above about f:23.

"Diffraction Limited" means the size of a spot made by a perfect lens is too big for the pixels to discriminate (ie. a soft image.) Again the ratio is a factor of 4..

These are not exact numbers but give an idea of what's going on, other things being equal..

I think what it implies is that a K100D's photo at f:23 will be about the same as the G9's photo at f:5.7 (both diffraction limited, same Signal/Noise ratio for the same exposure time.).

Dave..

Comment #8

OOPS - In my answer make that K100, not K100D..

Dave Martin wrote:.

BLawson wrote:.

This is confusing out here in novice-land, but fun..

If I compare sensor size/sensor site totals of the Canon G6 @12mpon a 1/1.7" sensor to a Pentax K100 @ 6mp on a 1.8" sensor, it comesout that with a "ratio" of 16/1..

G9= 1/1.7" sensor equals 44 sq. mm, or, .000037 individual sensor sizeK100 = 1.8" sensor equals 370 sq. mm, or, .000062 individual sensorsize.

So does that mean the Pentax has individual sensors that are 16xlarger than the G9 sesnors, or, that they are 16x further apart?.

The pixel areas differ by a factor of 16 (so the big pixel gathers 16times more light for the same exposure time & f-stop.).

The center-center distance of the pixels (pixel pitch) differs by thesquare root of 16 or a factor of 4. This also implies the Signal toNoise ratio for the big pixel is 4 times better (for the sameexposure time & f-stop.).

Ultimate resolution depends in part on pixel pitch. The best sensorresolution possible is twice the pixel pitch but often the lens can'tmatch that depending on quality & f-stop..

The Airy formula relating lens resolution and f-stop is roughly:p~f/3 in microns for greenish light..

For the two cases you mention,p=1.9 microns for the G9 so it is "Diffraction Limited" above aboutf:5.7p=7.8 microns for the K100D so it is "Diffraction Limited" aboveabout f:23.

"Diffraction Limited" means the size of a spot made by a perfect lensis too big for the pixels to discriminate (ie. a soft image.) Againthe ratio is a factor of 4..

These are not exact numbers but give an idea of what's going on,other things being equal..

I think what it implies is that a K100D's photo at f:23 will be aboutthe same as the G9's photo at f:5.7 (both diffraction limited, sameSignal/Noise ratio for the same exposure time.).

Dave..

Comment #9

This is way beyond me, but if I read it right, the ratio stays constant along the f:stop scale... so that the larger sensor at f:11 will gather as much light as the smaller sensor at around f:2.8 ? .

Does that then correlate with ISO... the larger sensor can be set higher/faster by a factor of 4 to produce the same image as the smaller sensor, i.e., ISO 200 on the G9 will be app. the same as ISO 800 on the Pentax?.

If that's correct, I think I just had what my students called an "ah ha" moment when I finally get it... Brian.

Dave Martin wrote:.

The pixel areas differ by a factor of 16 (so the big pixel gathers 16times more light for the same exposure time & f-stop.).

The center-center distance of the pixels (pixel pitch) differs by thesquare root of 16 or a factor of 4. This also implies the Signal toNoise ratio for the big pixel is 4 times better (for the sameexposure time & f-stop.).

These are not exact numbers but give an idea of what's going on,other things being equal..

I think what it implies is that a K100D's photo at f:23 will be aboutthe same as the G9's photo at f:5.7 (both diffraction limited, sameSignal/Noise ratio for the same exposure time.).

Dave.

Brian..

Comment #10

BLawson wrote:.

This is way beyond me, but if I read it right, the ratio staysconstant along the f:stop scale... so that the larger sensor at f:11will gather as much light as the smaller sensor at around f:2.8 ? .

Does that then correlate with ISO... the larger sensor can be sethigher/faster by a factor of 4 to produce the same image as thesmaller sensor, i.e., ISO 200 on the G9 will be app. the same asISO 800 on the Pentax?.

Aha! Got it!.

The k100 pixel has 16 times the area as the G9; since light intensity is photon flux per unit area, in a given exposure time the K100 collects 16 times as many photons for the same light intensity. Signal to noise ratio is proportional to the square root of number of photons so the k100 is 4 times better than the G9..

Or, if the light illuminating the k100's sensor is 1/16th that illuminating the G9's they will both collect the same number of photons per unit time (and the k100's potential well will be only 1/16th a full as the G9's.) Since the number of photons collected is the same for the two cases the signal to noise ratio is the same..

The K100 at ISO 1600 looks about the same as the G9 at ISO 100 from a signal to noise ratio standpoint. That's because ISO increases in software are a simple multiplication of the sensor's data..

Now, please don't take this as the literal truth; there are other factors involved like how much pixel area is obscured by circuitry lines, how good the microlenses are, how much noise is inherent in the electronics, etc..

See Clarkvision.com for good detail...

Comment #11

What I think you are asking is whether or not a good p&s can take the same shot as a 6mp dslr, and when printed at a 8x10ich how do they compare..

The answer is yes no and maybe. I can illustrate this by the following real story which I found to interesting..

Last summer in 2006 a pro working for a website and photo magazine went on a trip to hawaii. he took his canon20d, a dslr, and a canon s3IS, a good p&s super zoom. his intent was to shoot as many identical shots as possible with the 2 cameras and compare the results when he got home. before he went he said that the dslr should win in a walk. if for no other reason that the dslr has more mps..

What happened was the following: in 80% of the shots it was difficulkt to impossible to tell which camera shot which pic. the 80% however all had something in common. they were pics of everyday scenes or people or landscapes. there was not any closeups, heavy telephoto, or any other specialzed images. for everyday shooting the super zoom did extremely well. BUT, when the dslr was able to switch lenses for a special use or add other gear to get a more optimized photo gear setup; there was simply no comparison.

In the remaining 20% the p&s could not compete on a one for one quality image basis. the p&s beyond the builtin features could not adapt well enough in very difficulkt conditions, but the dslr could..

So therefore as long as the user stays inside of the limits that a good p&s super zoom has the pics will generally be very good. once you try to take images that are pushing the design limits then the super zoom simply has no way to change to get any more performance..

Please note that the above is talking about a max print size of 8x10inches. this was the comparison used. if you increase the print size to 11x14 or 16x20 or 20x30 THEN the p&s would not be able to compete in even more situations and would be doing this faster as the print size went up...

Comment #12

Dave Martin wrote:.

Now, please don't take this as the literal truth; there are otherfactors involved like how much pixel area is obscured by circuitrylines, how good the microlenses are, how much noise is inherent inthe electronics, etc..

I was hoping that someone would inject this point in this thread. The operative expression is "fill factor". It is simply the % of the area occupied by the photosensitive part. The support circuitry (a bunch of transistors) are masked so that light doesn't get on them. Manufacturers don't tell us much about it..

As the size of the photosite is reduced, the fill factor drops because the size of the transistors are fixed. Thus very small sensors have a really bad fill factor, because most of the APPARENT sensor area is NOT photosensitive..

The second part of this is that CMOS sensors generally have a better fill factor than CCD sensors..

The third part is that everybody uses microlenses to improve the fill factor. These microlenses do more good for small CCD sensors than for big CMOS sensors..

Your question is complex...there is no single correct answer..

The best rule I can state is: "Buy the biggest piece of Silicon you can afford and one with as few photosites as you can put up with.".

Charlie DavisNikon 5700, Sony R1, Nikon D300HomePage: http://www.1derful.infoBridge Blog: http://www.here-ugo.com/BridgeBlog/..

Comment #13

Chuxter wrote:.

Your question is complex...there is no single correct answer..

The best rule I can state is: "Buy the biggest piece of Silicon youcan afford and one with as few photosites as you can put up with.".

Charlie,.

Nice...that sounds like the single correct answer. .

Greg..

Comment #14

Gregory King wrote:.

Chuxter wrote:.

Your question is complex...there is no single correct answer..

The best rule I can state is: "Buy the biggest piece of Silicon youcan afford and one with as few photosites as you can put up with.".

Charlie,.

Nice...that sounds like the single correct answer. .

Greg.

Not really...it's more complex than that simplification. Increasingly, in-camera NR blurs the picture (pun intended). If in-camera NR really can dull the "grain" in big areas like the sky, yet preserve detail in busy areas, then we may see small sensors become acceptable for serious photography. Oh, sure...the old guard will never accept this, but they will die off..

But even with excellent in-camera NR, there will still be a small advantage to a camera that follows my simple Rule..

Charlie DavisNikon 5700, Sony R1, Nikon D300HomePage: http://www.1derful.infoBridge Blog: http://www.here-ugo.com/BridgeBlog/..

Comment #15

Brian,BOTH matter and they impact each other..

The real measure of SENSOR quality is S/N ratio. (Little used but the same as in audio.) Isolating the OF S/N from bit depth, processing software and other possible influencers is a bit tricky. So, bear with me....

In any given sensor, placed in a given camera, every thing about the design matters. Particularly the shielding of surrounding components that generate systemic noise. (Later, thermal noise and ambient light noise come into play, too.).

NOise: The systemic noise is roughly constant for a given system, all other things being equal. It comes from stray electrons floating around contaminating the sensor signal. Then, if you use the camera in hot temperatures or for along time, as the camera heats up, thermal noise is also induced. Also, if you increase ISO, (which is a gain pot really) the nosie is increased from it's floor level too. All of these can combine: a hot day, a high ISO, poor shielding to make a REALLY noisy image. Conversely, on a really cold day, only shielding is likely to be a gremlin..

Now, here is where receptor (pixel) size and array area come in. Larger receptors create stronger signal than smaller receptors, since they generally collect more electrons, because of their larger area, if all other things are constant. Since the noise bed is roughly constant, and influenced by the variables above. Let's talk at optimum low ISO performance and noise..

Larger receptors create stronger signal than smaller receptors. The area of a small receptor (say, 4x4 microns) is roughly one fifth the area of a large receptor (9x9 microns) by a factor of about 5X the area, so there are 5X as may electrons to count. Noise as a percentage of area is 1/5th as much. So, big receptors are contaminated by noise much less than small receptors..

Now, if you have 6MP or 12 MP (of similar manufacture and grade) ARRAY of receptors and both are the same size receptor, using the same bit depth, the image quality would be the same. The physical size of the array would be twice the area for the 12MP array. And it would produce twice the area of larger enlargements at identical quality..

However, if you have the same array size, with one having 6MP and the other 12MP, then the signal to noise ratio will be worse on the 12MP, since the recptors are smaller..

The 6MP will NOT enlarge as far, but will have more dynamic range and perceptibly more perceived sharpness (because of the S/N extra dynamic range and tonal spread) below the threshold limit of changeover to the 12MPs enlargability advantage..

Remember, perceived sharpness, is governed by both resoluteness and contrast. As one goes down and the other goes up, things look either better of worse, depending on which is the challenged component of the equation. The extra tones of the 6MP will carry the day, until they are stretched so far as to no longer calculate a "sharper" perceived image. Then resolution (pixel count) has more value..

Obviously, this does not take into account all of the other influcencers (output devices, processing software, etc)..

Bit depth divides the dynamic range into segments (10-12-14, 16, etc) for counting electrons and calculating tone. Less s/n ratio means the divisions are smaller. Usually bit#1 is assigned to the cover the system noise bed (at optimum, low ISO). Then as ISO increases and/or heat drives up noise, it clips bits two, three and so on, cutting down dynamic range and tone computation. With a noisy system, operating at high ISO on a hot day, even if the camera claims 12 bits, you may actually have only the use of 6-8 bits of data above the noise level..

So, it's amazing this stuff works at all. It's magic I tell you:-).

I hope this helps explain the issues even if it does not answers you question..

John.

BLawson wrote:.

, then it's better to have them smaller and furtherr apart... or does itmatter... or is this a variation on angels dancing on the head of apin arguments?Brian.

Van..

Comment #16

Click Here to View All...

Sponsored Amazon Deals:

1. Get big savings on Amazon warehouse deals.
2. Save up to 70% on Amazon Products.


This question was taken from a support group/message board and re-posted here so others can learn from it.

 

Categories: Home | Diet & Weight Management | Vitamins & Supplements | Herbs & Cleansing |

Sexual Health | Medifast Support | Nutrisystem Support | Medifast Questions |

Web Hosting | Web Hosts | Website Hosting | Hosting |

Web Hosting | GoDaddy | Digital Cameras | Best WebHosts |

Web Hosting FAQ | Web Hosts FAQ | Hosting FAQ | Hosting Group |

Hosting Questions | Camera Tips | Best Cameras To Buy | Best Cameras This Year |

Camera Q-A | Digital Cameras Q-A | Camera Forum | Nov 2010 - Cameras |

Oct 2010 - Cameras | Oct 2010 - DSLRs | Oct 2010 - Camera Tips | Sep 2010 - Cameras |

Sep 2010 - DSLRS | Sep 2010 - Camera Tips | Aug 2010 - Cameras | Aug 2010 - DSLR Tips |

Aug 2010 - Camera Tips | July 2010 - Cameras | July 2010 - Nikon Cameras | July 2010 - Canon Cameras |

July 2010 - Pentax Cameras | Medifast Recipes | Medifast Recipes Tips | Medifast Recipes Strategies |

Medifast Recipes Experiences | Medifast Recipes Group | Medifast Recipes Forum | Medifast Support Strategies |

Medifast Support Experiences |

 

(C) Copyright 2010 All rights reserved.