Pages: 1 2 [3]   Go Down

Author Topic: Pixelsize and depth-of-field  (Read 36639 times)

Gordon Buck

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
    • LightDescription
Pixelsize and depth-of-field
« Reply #40 on: May 26, 2008, 10:45:52 pm »

It appears that doubling the number of pixels on a full frame 35mm sensor only costs about one f-stop with respect to diffraction, see

http://www.cambridgeincolour.com/tutorials...sensor-size.htm
Logged
Gordon
 [url=http://lightdescription.blog

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Pixelsize and depth-of-field
« Reply #41 on: May 27, 2008, 02:16:10 am »

Quote
It appears that doubling the number of pixels on a full frame 35mm sensor only costs about one f-stop with respect to diffraction, see

http://www.cambridgeincolour.com/tutorials...sensor-size.htm
[a href=\"index.php?act=findpost&pid=198223\"][{POST_SNAPBACK}][/a]

That makes sense to me provided one is working within a range of diffraction limited f stops. A doubling of pixel count increases the resolution potential of the sensor by 1.4x. Moving up from a diffraction limited F22 to a diffraction limited F16 increases the resolution of the lens by 1.4x. However, stopping up from F4 to F2.8 might not provide any resolution increase. With some lenses one might lose resolution at F2.8.

I think the concept here is, one presumably uses a camera with a greater pixel count in order to achieve greater resolution. If one fails to achieve greater resolution because one's lenses are not good enough, or because the f stop chosen is too limited by diffraction, or because it's too limited by various types of aberrations at wide apertures, then the increased pixel count will not affect the perception of DoF. You might as well have taken the shot with the lower pixel camera, regarding over all sharpness and DoF.
Logged

Nick Rains

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 705
    • http://www.nickrains.com
Pixelsize and depth-of-field
« Reply #42 on: May 27, 2008, 05:35:50 am »

Quote
I think the concept here is, one presumably uses a camera with a greater pixel count in order to achieve greater resolution. If one fails to achieve greater resolution because one's lenses are not good enough, or because the f stop chosen is too limited by diffraction, or because it's too limited by various types of aberrations at wide apertures, then the increased pixel count will not affect the perception of DoF. You might as well have taken the shot with the lower pixel camera, regarding over all sharpness and DoF.
[a href=\"index.php?act=findpost&pid=198256\"][{POST_SNAPBACK}][/a]
This almost makes sense but if it was true there would never have been any reason to shoot large format, film or digital.

What you are forgetting is that large format lenses have longer focal lengths for a given FOV. This means that at any given aperture (ratio of iris size to focal length) the iris size is bigger and thus proportionately less affected by diffraction error. The same diffraction error would be found at a much smaller f-stop with correspondingly greater DOF. I suspect the two cancel out leaving the result that the larger the format the better image Q at any given output size and at any given pixel density or film type.

That's why LF, sensor or film, gives generally better results.

Edit : I misread your post, too hasty - you don't mention sensor size and thus I agree with you.  
« Last Edit: May 27, 2008, 06:44:47 am by Nick Rains »
Logged
Nick Rains
Australian Photographer Leica

Nick Rains

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 705
    • http://www.nickrains.com
Pixelsize and depth-of-field
« Reply #43 on: May 27, 2008, 06:41:35 am »

Quote
It appears that doubling the number of pixels on a full frame 35mm sensor only costs about one f-stop with respect to diffraction, see

http://www.cambridgeincolour.com/tutorials...sensor-size.htm
[a href=\"index.php?act=findpost&pid=198223\"][{POST_SNAPBACK}][/a]

Yes, I played with that too and found about 1.5 stops advantage to a 39mp 645 sensor over a 35mm size 22mp sensor with respect to diffraction limits.

Using the DOF calculators it seems that this is more than enough to counter the effects of longer focal length on the bigger sensor so at any given print size and viewing distance the bigger sensor will have twice as many pixels representing the image with much the same apparent DOF. This will equate to better quality prints over a certain size, with roughly equal DOF .
Logged
Nick Rains
Australian Photographer Leica

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Pixelsize and depth-of-field
« Reply #44 on: May 27, 2008, 08:50:40 am »

Quote
Edit : I misread your post, too hasty - you don't mention sensor size and thus I agree with you. 
[a href=\"index.php?act=findpost&pid=198269\"][{POST_SNAPBACK}][/a]

Well, thanks Nick   . I think there is some confusion in this thread between the two sets of circumstances, (1) a bigger sensor with more pixels and (2) the same size sensor with more pixels. The OP is asking about the DoF consequences of increasing pixel count on the same size sensor.
Logged

Gordon Buck

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
    • LightDescription
Pixelsize and depth-of-field
« Reply #45 on: May 27, 2008, 11:17:40 am »

Here's an amusing and nearly pertinent quotation from the Leica Manual, of 1938:

"If you really want to look at the picture you will never hold an enlargement of 8 x 10 inches closer than 10 inches from the eye.  Only grain fiends have a habit of smelling their pictures, regardless of size."
Logged
Gordon
 [url=http://lightdescription.blog

thsinar

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2066
    • http://www.sinarcameras.com
Pixelsize and depth-of-field
« Reply #46 on: May 27, 2008, 11:37:29 am »

that's exactly the point: the viewing distance. The same when I used to see photographers using a loupe to check their sharpness on a 4x5" transparency (or on any other format), when there is an optimal viewing distance for a film/print which defines the CoC/DoF. Of course, with digital it is impossible to control or view an image on the sensor, but the optimal viewing distance of the output is still part of the definition of sharpness/DoF.

Thierry

Quote
Here's an amusing and nearly pertinent quotation from the Leica Manual, of 1938:

"If you really want to look at the picture you will never hold an enlargement of 8 x 10 inches closer than 10 inches from the eye.  Only grain fiends have a habit of smelling their pictures, regardless of size."
[a href=\"index.php?act=findpost&pid=198335\"][{POST_SNAPBACK}][/a]
Logged
Thierry Hagenauer
thasia_cn@yahoo.com

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Pixelsize and depth-of-field
« Reply #47 on: May 27, 2008, 01:13:15 pm »

Using more, smaller pixels will not reduce decrease depth of field or decrease resolution when using the same aperture ratio so long as one compares with prints of the same size viewed from the same distance. Diffraction only reduces the extent of the resolution/sharpness/detail improvements that one gets at any given aperture from increased sensor resolution (smaller pixel size.)

Any apparent DOF increase at equal aperture attributed to using fewer, larger pixels is really an effect of viewing a smaller version of the lower pixel count image, like prints at the same PPI. You can increase DOF equally well by making equally small prints from a sensor with more, smaller pixels.


Pardon me if I repeat so often the dogma that the only image quality comparisons that matter are ones based on viewing at equal apparent image size (e.g. same print size and viewing distance, or to put on my propellor beanie, equal angular size for the image of the same subject.)
Ignoring this and instead comparing at different image sizes (like equal PPI, as with 100% on screen) leads to various misleading claims for benefits from lower sensor resolution, the likes of which I never heard in the film era.


P. S. Gordon, thanks for the wonderful Leica manual quote!
(It resembles my personal guideline of not closer than 15" from prints 11"x14" and bigger. Going closer is mainly to compensate for "undersized" prints, as 15" is a rough typical minimum comfortable viewing distance; a favorite for reading, for example. Maybe fitting a comfortable 15" or more viewing distance is part of why print sizes like 13"x19" and 16"x20" are popular for galleries.)
« Last Edit: May 27, 2008, 01:23:54 pm by BJL »
Logged

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
Pixelsize and depth-of-field
« Reply #48 on: May 27, 2008, 03:09:00 pm »

Quote
the likes of which I never heard in the film era
This one takes an eminent place amongst the arguments made on this thread. I wonder, why you have not distributed this post per snail mail or telex - there were no internet forums in those days, were there?
Logged
Gabor

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Pixelsize and depth-of-field
« Reply #49 on: May 27, 2008, 04:08:49 pm »

Quote
This one takes an eminent place amongst the arguments made on this thread. I wonder, why you have not distributed this post per snail mail or telex - there were no internet forums in those days, were there?
[a href=\"index.php?act=findpost&pid=198374\"][{POST_SNAPBACK}][/a]
You have taken a fragment out of a paragraph, thus rendering it meaningless without the full context. Whereas in the complete paragraph it was actually quite a sensible thing to say and was certainly not anti-progress in meaning, as it simply put pixel peeping in context.
« Last Edit: May 27, 2008, 07:09:29 pm by jjj »
Logged
Tradition is the Backbone of the Spinele

Kirk Gittings

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1561
    • http://www.KirkGittings.com
Pixelsize and depth-of-field
« Reply #50 on: May 27, 2008, 05:36:10 pm »

Quote
when there is an optimal viewing distance for a film/print which defines the CoC/DoF.
Thiery

I am one of those guys. It is that depth of detail in Large Format that is part of its attraction. I frankly could care less about any so called "optimal viewing distance". I look at my finished prints with a loupe or strong reading glasses and do the same at shows by Ansel Adams or Weston. There is a real different experience between a fine print and a billboard.
Logged
Thanks,
Kirk Gittings

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Pixelsize and depth-of-field
« Reply #51 on: May 27, 2008, 09:27:23 pm »

Quote
Any apparent DOF increase at equal aperture attributed to using fewer, larger pixels is really an effect of viewing a smaller version of the lower pixel count image, like prints at the same PPI. You can increase DOF equally well by making equally small prints from a sensor with more, smaller pixels.
[a href=\"index.php?act=findpost&pid=198360\"][{POST_SNAPBACK}][/a]

Yes, but it is also true that the apparent increase in DoF in the smaller image and smaller print still holds if the smaller image is interpolated to the same size as the larger image, unless all parts of both images appear equally sharp whatever the size, which might be the case if one had used a very small aperture for both shots and/or the scene photographed did not have great depth.

The DoF calculators, such as DoF Master, do not make it clear that the CoC used in their calculations is valid only for a fixed size print (which I believe is 8x10").

When one upgrades one's camera to one with more pixels, keeping the format size the same, the reason is presumably to enable one to make larger prints which are still acceptably sharp.

If one then views the larger print from a proportionally greater distance, the appearance of DoF will remain unchanged. In that sense, pixel count has no bearing on DoF.

One might therefore conclude that the main purpose of using a camera with a high pixel count is so that one can appreciate fine detail from a closer distance than one would normally view the print when appreciating the composition as a whole.
Logged

thsinar

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2066
    • http://www.sinarcameras.com
Pixelsize and depth-of-field
« Reply #52 on: May 27, 2008, 10:09:56 pm »

Yes, like many others, including myself, but not to "check" and adjust DoF.

But in this case any definition of sharpness and depth of field does not make any sense, respectively is changed by the observer/viewer.

Best regards,
Thierry

Quote
Thiery

I am one of those guys. It is that depth of detail in Large Format that is part of its attraction. I frankly could care less about any so called "optimal viewing distance". I look at my finished prints with a loupe or strong reading glasses and do the same at shows by Ansel Adams or Weston. There is a real different experience between a fine print and a billboard.
[a href=\"index.php?act=findpost&pid=198395\"][{POST_SNAPBACK}][/a]
Logged
Thierry Hagenauer
thasia_cn@yahoo.com

henrikfoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 899
Pixelsize and depth-of-field
« Reply #53 on: May 28, 2008, 03:01:31 am »

Thierry!

Can you tell me the size of the pixels of the older Sinarbacks like the 23 HR?
I couldn´t find it on the net.

Henrik
Logged

thsinar

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2066
    • http://www.sinarcameras.com
Pixelsize and depth-of-field
« Reply #54 on: May 28, 2008, 05:16:26 am »

Dear Henrik,

the pixel size of the 6 MPx Philips CCD sensor was 12 micros.

Best regards,
Thierry

Quote
Thierry!

Can you tell me the size of the pixels of the older Sinarbacks like the 23 HR?
I couldn´t find it on the net.

Henrik
[a href=\"index.php?act=findpost&pid=198472\"][{POST_SNAPBACK}][/a]
Logged
Thierry Hagenauer
thasia_cn@yahoo.com

henrikfoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 899
Pixelsize and depth-of-field
« Reply #55 on: May 28, 2008, 07:01:17 pm »

Quote
Dear Henrik,

the pixel size of the 6 MPx Philips CCD sensor was 12 micros.

Best regards,
Thierry
[a href=\"index.php?act=findpost&pid=198487\"][{POST_SNAPBACK}][/a]


Thanks a lot, Thierry!

Henrik
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Pixelsize and depth-of-field
« Reply #56 on: May 29, 2008, 12:09:52 pm »

Quote
When one upgrades one's camera to one with more pixels, keeping the format size the same, the reason is presumably to enable one to make larger prints which are still acceptably sharp.
[a href=\"index.php?act=findpost&pid=198441\"][{POST_SNAPBACK}][/a]
Indeed. Let me put it this way: a sensor with higher pixel count through more, smaller photosites on the same sensor size gives the option of seeing more detail in the image, and that option comes with less DOF. So when one aims to use the greater detail and also wants adequate DOF, this must be accounted for in choice of aperture: smaller apertures are needed. And thus, getting the maximum possible detail from the higher resolution sensor imposes a lower maximum DOF.

But one is never forced to accept lower DOF than is possible with a lower resolution sensor, since there is also the option of not using the extra detail, for example by printing and viewing in the same way as one would with the lower resolution option, perhaps downsampling first.

And therefore, choosing a sensor of lower resolution does not add anything to one's DOF or related IQ options: it simply discards some options for more detail combined with less DOF.


More generally, I still have not seen any good evidence or arguments that higher pixel count sensors cannot match or outdo the results of lower pixel count sensors of the same size with suitable processing or presentation. The processing needed might be as simple as converting to JPEG (or TIFF) at about the same pixel count as the lower resolution sensor.

Or maybe conversion to JPEG at a somewhat lower pixel count than the lower resolution sensor, since in JPEG's of equal pixel count from Bayer CFA sensors, those from a higher pixel count sensor can have a detail advantage, due to less loss of resolution in the demosaicing interpolation.
Logged
Pages: 1 2 [3]   Go Up