Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
10-13-2014, 07:54 AM   #16
Pentaxian
jsherman999's Avatar

Join Date: Dec 2007
Photos: Gallery
Posts: 8,228
QuoteOriginally posted by bxf Quote
All this makes sense to me (meaning I find it logical).

So if this is the correct explanation, are we saying that statements pertaining to FF superiority due to "more total light", including that made by DPR, are plainly wrong?
No.

A lot of mixed info in this thread already.

For one thing, the exact term itself "Total Light" isn't something that really came out of photographic history, like the term "Exposure" did - so it's not something you missed earlier It's just a term that wraps a very important concept that prior to this only carried some vague explanations involving the observations that "larger formats are better". That went back all the way to view cameras. Very basically it describes why you can get more IQ out of larger formats.

As we moved into digital some people attributed this IQ bump in noise/DR directly to "Larger pixels", or more pixels, or better pixels - but really it has much more to do with the larger physical aperture used on these larger formats at the same FOV and exposure.

For a larger sensor area, we require a larger physical aperture (lens pupil diameter) for the same exposure and FOV. This larger physical aperture is dumping more total photons into the image - so even though we have the same FOV and when displayed at the same display size we have an image that was created with more light. This is going to result in an image with less noise and probably more DR, with a very big caveat in that the efficiency of both sensors in the comparison also affect the noise - in other words, a much older, larger sensor isn't going to do much better than a newer, better smaller sensor even if it's getting more light (or it could even do worse if it's old enough.)

I don't know if it can be explained any more simply than that. If it can, someone more articulate than I needs to do it

There are a few links you can look at and some threads on dpreview that feature appearances by folks like Bob Newman and Eric Fossum where total light, photon shot noise and read noise are discussed in detail in relation to sensor technologies. Those are enjoyable to read - free education. More than once I've had a few of my preconceptions challenged there.

Here's a thread that has a nice IMO simple object field representation by The_Suede that illustrates the concept pretty well: Total LIght Gathered? ("Great Bustard" has some very nice contribution to that thread as well)

.


Last edited by jsherman999; 10-13-2014 at 08:08 AM.
10-13-2014, 10:30 AM   #17
Pentaxian
KevinR's Avatar

Join Date: Jan 2010
Location: Cape Town, South Africa
Photos: Gallery | Albums
Posts: 646
All light source variables being equal (lens, aperture, SS), the total light per unit sensor area will be identical between FF and APS-C.

If sensor technology is the same (pixel/photo site size and composition, and electronics performance etc), then the individual pixel signal to noise ISO performance will be identical.

However, typically practically FF either has larger photo sites (same Mpix case) in which case photo site signal to noise ratio is improved, or FF has more photo sites (same Mpix density case) in which case photo site signal to noise ratio is the same, but resulting image can use smoothing to improve the resulting overall image quality.

My 2c worth to try to be simple but very clear.
10-14-2014, 12:47 PM   #18
Site Supporter




Join Date: Jul 2011
Photos: Gallery | Albums
Posts: 677
It is so hard to equate some things... I don't claim to have the answers... but I have come to my own conclusions that make me happy enough to move on and not worry about it.

Personally, I like to imagine we still "paint with light". In digital, we paint with light, even on our sensors.

If you ever saw the first camera obscura, and think about what was going on in there, it might click.

If the hole in the wall of the box the guy was sitting inside and doing the painting was projecting a pinhole on the wall the size of a postage stamp, that guy would have been sitting in a pretty dark box.

If the guy sitting in that box had a hole in the wall big enough to paint a 20x30" canvass, he had a lot more scattered light in the box with him, even if he still decided to use a canvas the size of postage stamp,then the postage stamp would only have a very small field of view compared to the entire scene reflected on the wall made by the larger hole, and he would still have all that extra stray light in the box with him with the larger hole.

The additional light is wasted and of no benefit to the smaller canvas because it is spread out and if anything, probably causes more work for internal coatings and anti-reflective surfaces to have to deal with absorbing and getting rid of, which is why some lenses (even FF) have a rectangular baffle at the rear, to help increase contrast.

The reason the newer generation APSC lenses are better on APSC now is probably not only the better coatings and computer designed stacked elements of rare minerals, but also the smaller image circle optimized for APSC for reasons mentioned above.

As to why this is an advantage to FF, that is probably very real, even if we don't fully understand why other than less magnification to the final output... but I think the full use of the sensor area potential inside the box, with less scattered light going to waste inside the walls of the box to contend with is a good thing since we are still using what started life as a FF K-mount. The larger sensor, even with lower density and same total pixel count seems to preform better because of less crosstalk between the photosites and better heat dissapation, etc. as others have mentioned already. All the advantages seem to go to FF if the final ultimate IQ under a wider variety of lighting is all that matters.



https://en.wikipedia.org/wiki/Camera_obscura#mediaviewer/File:Camera_obscura2.jpg

Last edited by Erictator; 10-14-2014 at 12:58 PM.
10-14-2014, 01:00 PM   #19
Pentaxian
Sagitta's Avatar

Join Date: Feb 2011
Location: Maine
Photos: Gallery
Posts: 2,912
If you want to think about this abstractly (and somewhat incorrectly), imagine you have a window in your house in a room that's painted over with black paint. Its the ONLY light source for that room - there are no other windows or anything.

Open the window halfway, and you'll wind up with a nice, bright half-window sized patch of light on the floor, and the room will brighten somewhat as a result.

Smash the window out entirely, and you'll have a full window-sized patch of light hitting your floor, and the rest of the room will illuminate even more. Obviously you have the same light source, with the same quality of light coming in, so you would think that while the patch on the floor would indeed get larger, the rest of the room would only illuminate the same amount... (ie, stuff in the corners won't brighten up any more than they have) but it doesn't work that way. The larger sheer amount of light coming in makes the rest of the room brighten even more, because the total light coming in is greater.

10-14-2014, 01:28 PM   #20
Loyal Site Supporter
clackers's Avatar

Join Date: Jul 2013
Location: Melbourne
Photos: Albums
Posts: 7,721
QuoteOriginally posted by Sagitta Quote
so you would think that while the patch on the floor would indeed get larger, the rest of the room would only illuminate the same amount... (ie, stuff in the corners won't brighten up any more than they have) but it doesn't work that way.
Don't know where to start with this, Sagitta.

By your puzzling analogy (window=lens, room=sensor), the nicely lit, sharp image circle on the ground is the cropped factor one, and the problematic outer shading is full-frame.

And that's what happens.

APS-C gets the best part of the lens, FF gets worse vignetting and more distortion.

The DA15 is well loved on this forum, but any 15mm on FF is likely to end up a fisheye.
10-14-2014, 01:37 PM   #21
Site Supporter
dakight's Avatar

Join Date: Sep 2014
Location: Phoenix, AZ
Photos: Albums
Posts: 1,138
Thinking about it in terms of film photography I see it this way. Take for example Tri-X film which has known, reproducible characteristics of grain, saturation and contrast. Grain is analogous to sensor noise if you think about it. If a 35mm negative is enlarged to say 8x10 it will have a certain graininess but if a medium format negative of the same tri-x film is enlarged to the same 8x10 final image it will appear much smoother because the inherent graininess of the negative is not as enlarged and the resulting appearance of grain will be correspondingly finer. The exact same result will be evident with equivalent digital sensors. Grain is to film what noise is to a digital sensor.
10-14-2014, 03:19 PM   #22
Pentaxian
ChristianRock's Avatar

Join Date: Jul 2013
Location: Marietta, GA
Photos: Gallery | Albums
Posts: 4,048
Ok, if it's all about sensor size within the same generation, how come the Pentax MX-1 is only 3.5 stops behind a full frame camera in terms of ISO noise, and 2.5 stops behind an APS-C camera? The difference is even smaller if you look at dynamic range.

If you look at the sensor area, FF is 864 mm≤ and the MX-1 is only 41,5 mm≤. That's well over 20 times the sensor size. If it's all about total light, the difference should be much bigger.
10-14-2014, 03:21 PM   #23
Pentaxian
Sagitta's Avatar

Join Date: Feb 2011
Location: Maine
Photos: Gallery
Posts: 2,912
QuoteOriginally posted by clackers Quote
Don't know where to start with this, Sagitta.

By your puzzling analogy (window=lens, room=sensor), the nicely lit, sharp image circle on the ground is the cropped factor one, and the problematic outer shading is full-frame.

And that's what happens.

APS-C gets the best part of the lens, FF gets worse vignetting and more distortion.

The DA15 is well loved on this forum, but any 15mm on FF is likely to end up a fisheye.
On a lark I stuck my Sigma 10-20 on my old SF10 just to see where the vignetting would start - at least as per the viewfinder, I could probably get down to around 14 to 13mm before the vignetting started in. This actually impressed me somewhat, as if there is ever a compatible FF I would have a pretty crazy wide angled lens for the format.

10-15-2014, 02:08 AM   #24
bxf
Pentaxian
bxf's Avatar

Join Date: Jul 2010
Location: Lisbon area
Posts: 1,041
Original Poster
QuoteOriginally posted by jsherman999 Quote
No.
...
Here's a thread that has a nice IMO simple object field representation by The_Suede that illustrates the concept pretty well: Total LIght Gathered? ("Great Bustard" has some very nice contribution to that thread as well)

.
I've been somewhat under the weather the last few days, so I haven't been responding to any comments here, for which I am grateful nevertheless. But I have been reading...

jsherman999's link above is recommended reading on the subject, although it gets somewhat heavy in spots. I have yet to re-read and digest everything discussed there, but so far, if my currently foggy brain is not mistaken, it does make one realize that the common sense, logical view, which "dictates" that more light on a larger area should not result in any benefit on any individual spot, is not correct. Quite interesting.
10-15-2014, 03:16 AM   #25
Pentaxian




Join Date: Apr 2007
Location: Romania
Posts: 8,426
QuoteOriginally posted by ChristianRock Quote
Ok, if it's all about sensor size within the same generation, how come the Pentax MX-1 is only 3.5 stops behind a full frame camera in terms of ISO noise, and 2.5 stops behind an APS-C camera? The difference is even smaller if you look at dynamic range.

If you look at the sensor area, FF is 864 mm≤ and the MX-1 is only 41,5 mm≤. That's well over 20 times the sensor size. If it's all about total light, the difference should be much bigger.
That's easy enough so I could try to give an answer I won't try to make it look too scientific, I promise (working with the proper level of abstraction is very useful); it's not my area of expertise anyway.

First, just to state the obvious: exposure is the amount of light per unit of area. "Total light", i.e. the total amount of light for the entire sensor area is exposure multiplied by sensor area.

As I'm sure you already know, noise has several sources. Significant in normal conditions is the photon/shot noise, which depends solely on this "total light"; that's why a larger sensor will always start with an advantage of about the square root from the "area multiplier". Technology can do nothing to improve this.
There are however other sources of noise which do depend on technology; e.g. read noise which do not give the larger sensor an advantage. Read noise - since I've mentioned it - tends to become significant in very low light.

So my answer would be:
1. overestimating the advantage re. photon noise
2. other sources of noise will also influence the result.
10-15-2014, 05:48 AM   #26
Site Supporter
dakight's Avatar

Join Date: Sep 2014
Location: Phoenix, AZ
Photos: Albums
Posts: 1,138
QuoteOriginally posted by ChristianRock Quote
Ok, if it's all about sensor size within the same generation, how come the Pentax MX-1 is only 3.5 stops behind a full frame camera in terms of ISO noise, and 2.5 stops behind an APS-C camera? The difference is even smaller if you look at dynamic range.

If you look at the sensor area, FF is 864 mm≤ and the MX-1 is only 41,5 mm≤. That's well over 20 times the sensor size. If it's all about total light, the difference should be much bigger.

That's not correct. The are different implementations of the APS-C sensor but none are as small as 41.5 mm≤. A Canon sensor is 329 (22.2x14.8) and a Nikon sensor is 370 mm≤ (23.6x15.7). I don't know the exact size of a Pentax sensor but it would certainly be in that ballpark. The actual area difference is going to be the square of the crop factor (1.6≤ - 2.56) or about 2 1/2 times the area, not 20+!


*edit* I missed the "MX-1" when I read the original message so I have to say that I was mistaken in that regard. My apology. My remarks were regarding the APS-C sensor which apparently is not used in the MX-1, a camera with which I am unfamiliar.

Last edited by dakight; 10-15-2014 at 05:51 AM. Reason: misread
10-15-2014, 12:42 PM   #27
Site Supporter




Join Date: Feb 2011
Posts: 4,892
QuoteOriginally posted by ChristianRock Quote
Ok, if it's all about sensor size within the same generation, how come the Pentax MX-1 is only 3.5 stops behind a full frame camera in terms of ISO noise, and 2.5 stops behind an APS-C camera? The difference is even smaller if you look at dynamic range.

If you look at the sensor area, FF is 864 mm≤ and the MX-1 is only 41,5 mm≤. That's well over 20 times the sensor size. If it's all about total light, the difference should be much bigger.

2^3.5 = 11.3. So the projection is 'off' by 'only' one stop if your numbers are correct. I'm not sure where you got the 3.5 stops number though.


Edit: just compared the D810 vs the MX-1 at DxO. The difference at 30.7 dB of SNR is 13.3. At 24.6 dB it's 13.62.

So projected difference with area is 4.38 stops ('over 20 times').
Measured difference from DxO at 30.7 dB is 3.73 stops.
Measured difference from DxO at 24.6 dB is 3.77 stops. Not much difference.
"ISO Score" from DxO mark (compilation of SNR, Dynamic Range, etc) is 3.78 stops.

So the projection is 'off' by a bit more than half a stop.

Newer, better technologies is usually seen in smaller chips first... for instance, the MX-1 is reportedly BSI. The fundamental assumption for the theory to be applied without thought is that the chips have to be of the same technology. Clearly that assumption is invalid and thought is required. Even with that assumption invalid, though, it's not like the results are out of left field.

It would be interesting to see if they have the same results for the same chip in an ILC camera, as well.

Last edited by ElJamoquio; 10-15-2014 at 12:55 PM.
10-15-2014, 01:49 PM   #28
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 23,772
This is only an issue if noise at the ISO you need to shoot is an issue for you.

What the numbers mean is, if you're not happy with the noise in your images, you may be better off with a larger format. Rent something try it out.
It's quite possible that you just learn to live with what you have. If you're not happy with the noise of a 300mm on a K-3, it doesn't mean you're going to be happy with the weight of 600mm lens on your 645z. The problem with these kinds of questions , is, you can't get everything you want. You have to decide what's most important to you. For absolute IQ you want the 600mm on the 645z,,,, but honestly, most of use are not interested in that. To much weight, too much cost, all the sudden a 400mm lens on a FF looks better. Less weight less cost, or maybe you decide given the alternatives a 300mm lens on a K-3 isn't so bad.

You just have to find what you can live with. To me, part of that is what your customers want. I'll buy what they'll pay for.

With the 645z, D3s and A7s, you certainly have more choices now than ever before.
Until there's a K-3s APS-c is not giving you a workable APS-c option if you're having to bump up your ISO to get the images you want, and that's causing too much noise. But keep it practical. Base it on your images, not on a bunch of numbers.

Last edited by normhead; 10-15-2014 at 02:00 PM.
10-15-2014, 04:36 PM   #29
bxf
Pentaxian
bxf's Avatar

Join Date: Jul 2010
Location: Lisbon area
Posts: 1,041
Original Poster
QuoteOriginally posted by normhead Quote
This is only an issue if noise at the ISO you need to shoot is an issue for you.

What the numbers mean is, if you're not happy with the noise in your images, you may be better off with a larger format. Rent something try it out.
It's quite possible that you just learn to live with what you have. If you're not happy with the noise of a 300mm on a K-3, it doesn't mean you're going to be happy with the weight of 600mm lens on your 645z. The problem with these kinds of questions , is, you can't get everything you want. You have to decide what's most important to you. For absolute IQ you want the 600mm on the 645z,,,, but honestly, most of use are not interested in that. To much weight, too much cost, all the sudden a 400mm lens on a FF looks better. Less weight less cost, or maybe you decide given the alternatives a 300mm lens on a K-3 isn't so bad.

You just have to find what you can live with. To me, part of that is what your customers want. I'll buy what they'll pay for.

With the 645z, D3s and A7s, you certainly have more choices now than ever before.
Until there's a K-3s APS-c is not giving you a workable APS-c option if you're having to bump up your ISO to get the images you want, and that's causing too much noise. But keep it practical. Base it on your images, not on a bunch of numbers.
Too practical, Norm

Still, you must admit that there is at least the occasional situation where one wishes one could shoot at an ISO step lower, no?

Aside from that, it could be said that this is simply academic - understanding for its own sake.

---------- Post added 16-10-14 at 00:42 ----------

Here is my simplified take on this entire matter. If what Iím saying is nothing that hasnít been said before, then all I can say is that at least Iíve clarified things for myself Also, these are my thoughts, and are being presented as such. I donít claim to have verified the validity of everything that follows.

Imagine a box, one meter on each side, with a window of 25cm square in the centre of one of the sides. Outside the box, directly opposite this window is a uniform light source that is larger than one meter. i.e. larger than the side of the box. The window will permit a certain amount of light to enter the box, and the centre of the side opposite the window will be hit by light of a certain intensity.

If we now increase the size of the window to 50 cm square, there will obviously be more light inside the box. However, the intensity of the light at the centre of the opposite side will be the same as it was with the smaller window (this is an assumption on my part).

I believe the above illustrates the difference between ďlight intensityĒ and ďtotal amount of lightĒ. What this means is that one can have more light, but equal exposure.

Now, a 150mm/f4 lens has an aperture of 37.5mm. This size aperture will permit a certain amount of light to enter the camera. On a FF body, this 150mm lens will have an angle of view (AOV) of 15 degrees. To get the same AOV using a cropped sensor body, you need to use a 100mm lens. In order for this 100mm lens (or any other lens, for that matter) to permit the same quantity of light to enter the camera as would the 150mm f4 lens, the aperture has to be the same size i.e. 37.5mm. This means that the lens would be 100mm f2.67 lens (100/37.5). Hence the argument for applying the crop factor to the f stop as well as the focal length. We can say that the 100mm f4 lens is equivalent to 150mm f4 on APS-C for the purposes of exposure (light intensity), but not in terms of total amount of light. As stated above, a 100mm lens would have to be f2.67 in order to permit the same amount of light to enter the camera as a 150mm f4 lens.

Now I ask myself the following: what if donít care about perspective or DOF in some situations, so I simply move back from the subject until I get the same AOV with my APS-C 100mm f4 as the FF would in the original position. Would this not effectively retain my 100m f4 as a 100mm f4? My guess (and thatís all it is) is that in this example, increasing the subject-to-camera distance reduces the light that enters the camera, and you are no better off.


Sometimes I find that I can be satisfied with a logical explanation, even if i's wrongI don't know if this is the case here.
10-15-2014, 05:09 PM   #30
Loyal Site Supporter
clackers's Avatar

Join Date: Jul 2013
Location: Melbourne
Photos: Albums
Posts: 7,721
QuoteOriginally posted by bxf Quote
In order for this 100mm lens (or any other lens, for that matter) to permit the same quantity of light to enter the camera as would the 150mm f4 lens, the aperture has to be the same size i.e. 37.5mm. This means that the lens would be 100mm f2.67 lens (100/37.5).
Whoa, BXF ... why do you need the same quantity of light when the frame is half the size?

That's why for exposure you don't have to think about sensor format or length of lens - shutter speed, ISO and lens aperture are the only three camera factors.

When a D800 shoots in cropped DX mode it won't be metering a stop lower or switching on an electronic ND filter!

The 'total light' thing is a furphy.

Last edited by clackers; 10-15-2014 at 05:24 PM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
24x36mm, answer, aps-c, camera, d800, density, equivalence, exposure, factors, ff, full-frame, heat, image, iso, k5, lens, light, noise, pentax, pixel, pixels, reason, sense, sensor, sensors, size, surface
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Total newby here mungobrick Troubleshooting and Beginner Help 9 03-19-2014 06:21 AM
Total Newbie Question jonatizzle Troubleshooting and Beginner Help 2 08-02-2013 10:02 AM
For Sale - Sold: Total Sell-Off kevwil Sold Items 10 05-21-2013 10:34 AM
Landscape Total eclipse 2012 AussieTrev Post Your Photos! 14 11-15-2012 04:45 PM
Total Amateur Jackie51 Troubleshooting and Beginner Help 6 10-19-2012 06:02 PM



All times are GMT -7. The time now is 10:25 PM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top