Resolution isn't coverage of an area, it's linear

Started Apr 9, 2014 | Discussions
RichRMA Veteran Member • Posts: 4,073
Resolution isn't coverage of an area, it's linear
5

A lot of people seem to equate double the resolution with doubling the pixel count.  It doesn't work that way.  Want to double the resolution (resolve detail twice as small) with a camera?  You have to double the pixel count in both x and y dimensions.  So, if the camera has a 3000x2000 pixel array (6 megapixels), you need 6000x4000 or 24 megapixels to double resolution.  To double that, you need 12000x8000 (96 megapixels).  So, it's four times the pixels that allows doubling resolution.  That is why (when you really look at an image) resolution gains going from say 24mp to 36mp are really not that impressive.

AndreaV Regular Member • Posts: 343
Re: Resolution isn't coverage of an area, it's linear

RichRMA wrote:

A lot of people seem to equate double the resolution with doubling the pixel count. It doesn't work that way. Want to double the resolution (resolve detail twice as small) with a camera? You have to double the pixel count in both x and y dimensions. So, if the camera has a 3000x2000 pixel array (6 megapixels), you need 6000x4000 or 24 megapixels to double resolution. To double that, you need 12000x8000 (96 megapixels). So, it's four times the pixels that allows doubling resolution. That is why (when you really look at an image) resolution gains going from say 24mp to 36mp are really not that impressive.

Do you mean that going from a 24Mp camera to a 36Mp camera you increase your resolution by an almost negligible 12.26%?? A few people could probably kill you for such a statement in this forum!!

Anyway, that's totally correct!

 AndreaV's gear list:AndreaV's gear list
Canon EOS-1D Mark III Canon EOS-1Ds Mark III Canon EOS 350D Sony Alpha NEX-5 Fujifilm X-Pro1 +14 more
darklamp Senior Member • Posts: 3,567
Even quadrupling won't work normally
1

On paper quadrupling the number of pixels sounds like it might work, but it won't get you twice the resolution.

The reason is that lenses do not conveniently increase their resolving power just because you increase the sensor pixel count.

Even with the very best lenses you may find that you can only get the required resolution in a limited range of apertures ( typically an SLR lens will peak at f8 ) and perhaps only in the center of the frame.

PerL Forum Pro • Posts: 13,958
On the other hand
2

RichRMA wrote:

A lot of people seem to equate double the resolution with doubling the pixel count. It doesn't work that way. Want to double the resolution (resolve detail twice as small) with a camera? You have to double the pixel count in both x and y dimensions. So, if the camera has a 3000x2000 pixel array (6 megapixels), you need 6000x4000 or 24 megapixels to double resolution. To double that, you need 12000x8000 (96 megapixels). So, it's four times the pixels that allows doubling resolution. That is why (when you really look at an image) resolution gains going from say 24mp to 36mp are really not that impressive.

If you translate the figures to print sizes, hanging on your wall, you can see that there is a difference in impact.

AndreaV Regular Member • Posts: 343
Re: Even quadrupling won't work normally

darklamp wrote:

On paper quadrupling the number of pixels sounds like it might work, but it won't get you twice the resolution.

The reason is that lenses do not conveniently increase their resolving power just because you increase the sensor pixel count.

Even with the very best lenses you may find that you can only get the required resolution in a limited range of apertures ( typically an SLR lens will peak at f8 ) and perhaps only in the center of the frame.

Absolutely right, but at this point we were considering only the sensor itself, in the hypothesis of an ideal lens if front of it.

If the lens is limiting the resolution, then the sensor pixel number doesn't count anymore...

 AndreaV's gear list:AndreaV's gear list
Canon EOS-1D Mark III Canon EOS-1Ds Mark III Canon EOS 350D Sony Alpha NEX-5 Fujifilm X-Pro1 +14 more
MoreorLess Veteran Member • Posts: 4,570
Re: Even quadrupling won't work normally

AndreaV wrote:

darklamp wrote:

On paper quadrupling the number of pixels sounds like it might work, but it won't get you twice the resolution.

The reason is that lenses do not conveniently increase their resolving power just because you increase the sensor pixel count.

Even with the very best lenses you may find that you can only get the required resolution in a limited range of apertures ( typically an SLR lens will peak at f8 ) and perhaps only in the center of the frame.

Absolutely right, but at this point we were considering only the sensor itself, in the hypothesis of an ideal lens if front of it.

If the lens is limiting the resolution, then the sensor pixel number doesn't count anymore...

Realistically its normally a combination of them both, lenses don't generally hit a wall where extra pixels don't aid resolution, its just a question of how much extra resolution they offer.

This is likely the reason why though larger formats will always be the best way to get better resolution. 24mp on ASPC is hitting the realms of diminishing returns for most lenses outside of the centre of the frame, on FF your probably not reaching that level until you get to 40+ MP.

AndreaV Regular Member • Posts: 343
Re: On the other hand
1

PerL wrote:

RichRMA wrote:

A lot of people seem to equate double the resolution with doubling the pixel count. It doesn't work that way. Want to double the resolution (resolve detail twice as small) with a camera? You have to double the pixel count in both x and y dimensions. So, if the camera has a 3000x2000 pixel array (6 megapixels), you need 6000x4000 or 24 megapixels to double resolution. To double that, you need 12000x8000 (96 megapixels). So, it's four times the pixels that allows doubling resolution. That is why (when you really look at an image) resolution gains going from say 24mp to 36mp are really not that impressive.

If you translate the figures to print sizes, hanging on your wall, you can see that there is a difference in impact.

Keeping a fixed ppi number in the printing your printed area increase as much as the number of pixels, and the width of the print as the square root of the number of pixels.

Anyway I print in A3+ (30x45cm) images shot with both my Canon 350D (8Mp) and Canon 1DmkIII (10Mp) and still you can't see a pixel even with a magnifier lens.

 AndreaV's gear list:AndreaV's gear list
Canon EOS-1D Mark III Canon EOS-1Ds Mark III Canon EOS 350D Sony Alpha NEX-5 Fujifilm X-Pro1 +14 more
MoreorLess Veteran Member • Posts: 4,570
To me this is a bit of a non issue...

RichRMA wrote:

A lot of people seem to equate double the resolution with doubling the pixel count. It doesn't work that way. Want to double the resolution (resolve detail twice as small) with a camera? You have to double the pixel count in both x and y dimensions. So, if the camera has a 3000x2000 pixel array (6 megapixels), you need 6000x4000 or 24 megapixels to double resolution. To double that, you need 12000x8000 (96 megapixels). So, it's four times the pixels that allows doubling resolution. That is why (when you really look at an image) resolution gains going from say 24mp to 36mp are really not that impressive.

For me this semantic debate is a bit of a none issue to me because most people who hold one view or another on what doubling resolution means will do so for input AND output.

That's to say someone who thinking moving from an 18 MP camera to a 36 mp camera(ignoring of course the issues of format size, lens performance, noise performenace etc where are very real) is doubling there resolution also probably thinks that moving from say printing A3 to printing A2 will double the resolution of their printed output so ultimately the two issues balance themselves out.

To me this also highlights that the real need for a doubling in resolution is likely beyond the needs of all but a select few people. Not only are A1 prints rarely needed but when they are produced the kind of resolution that's needed for them often diminishes as viewing distance increases.

Glen Barrington
Glen Barrington Forum Pro • Posts: 19,658
I don't care, as long as . . .

you don't try to tell me what REAL photographers should believe or do.  This is only tangential to good photography in my opinion.

The major manufacturers seem to do a good enough job at keeping sensor technology up to date, though I do think they aren't pushing optical quality hard enough.

-- hide signature --
 Glen Barrington's gear list:Glen Barrington's gear list
Canon PowerShot G3 Olympus OM-D E-M10 Olympus Zuiko Digital ED 9-18mm 1:4.0-5.6 Olympus Zuiko Digital ED 40-150mm 1:4.0-5.6 Olympus Zuiko Digital 14-54mm 1:2.8-3.5 +12 more
Andreas Stuebs Contributing Member • Posts: 924
It's a case of the law of diminishing returns
1

My first digital camera had a 2MP sensor. My next camera had 5. That was a noticeable difference in one step. From Pentax K5 to K3 is a difference of 6MP but relatively it is much less of a deal.

I don't subscribe to the 6MP is enough notion,  but there comes the times when the increments become less and less relevant and improvements should be looked for elsewhere.

 Andreas Stuebs's gear list:Andreas Stuebs's gear list
Samsung Galaxy Note 10.1 Panasonic Lumix DMC-LX3 Panasonic Lumix DMC-G2 Pentax K-3 Pentax smc FA 50mm F1.4 +14 more
Keyboard shortcuts:
FForum MMy threads