zodiacfml: another high quality article from you guys.
a bit off topic but from my perspective noise is becoming less of an issue these days especially with the CMOS sensors.
from these noise control developments, I discovered that i place more value now on highlight behavior than noise. viewing an image on an excellent but small screen such as smartphone or tablet, noise is not much of problem. in short, I appreciate the highlight handling of CCD cameras regardless of screen size or sensor format.
i hope you could also look into this as I believe there is a difference between CCD and CMOS regarding this aspect. i'm sure Leica fans would love to read such an article.
I don't know of any technical reason why CCD behave differently, in terms of hightlights.
When either type of sensor clips to white, it clips to white and neither technology, as they're currently used, has any means of rolling-off that transition (it's a linear response all the way up to clipping). Any recoverable highlight detail beyond this point comes down to slight differences in color filters.
Dynamic range is the gap between the hard cut-off of highlights and the darkest usable tone before the signal-to-noise ratio drops to an unacceptable level. Consequently, caring about dynamic range is to care about noise.
nutraman: in the second image, how could both shots be 1/100s and F8? One shot was at ISO100 and the other at ISO3200... so the incorrect label should have a shutter speed 5 stops different.
The whole point is to give both photos the same amount of light but then apply different levels of amplification to what's captured by the sensor.
The ISO3200 shot has applied 32x amplification, whereas the second is at base ISO (minimal amplification) and was then selectively brightened (amplified) in software.
armandino: I still do not get this:The Seattle sunset shot above is wonderful, but the aggressive NR is quite a turn off. Why not bracketing instead? Why not choosing for such articles truly meaningful photos for real practical situations? Cameras and applications like Lightroom make true HDR shots painless and vastly superior for this sort of compositions.
We hope to replace the second image with a more compelling example. As always, what you do with the potential your camera offers is where the interesting aspects of photography begin.
GRUBERND: can dpreview please moderate (= delete) all comments containing the words "equiv*", "match*", "fullframe" et al on all news items about any sensor smaller than 8x10" fullframe? thank you.
or the team could push the thought of "it's a tool, choose wisely and use accordingly" a bit more. ;)
back on topic: wow, that 7-14 looks like an awesome lens. =)
I agree, it's not an absolute *measure* of performance, but it does accurately represent how much light is available to the sensor, which is a useful thing to know.
It's a meaningful property of the lens.
coyot3: I tested on my nikon d3100 and the results are at iso 100 1" f4 underexposing 4 ev and later recovering in Darktable(linux) give less noise than iso 3200 1/30 f4 no exposure changes.
I dont know if i understand right the article xD im going to re read ._.
¿Some thoughts on my results?
The difference in shutter speed (and hence [shot noise](http://bit.ly/shotnoise)) will explain the results you're seeing.
Coheritier: Ok. I understand what you are saying. However, I have a technical question. Since photons travel in random patterns and according to your explanation; the more volume of photons that hit the sensor the better.
What happens when one adds time into the equation? Do I benefit more from having a longer exposure (slower shutter speed) than I do with a larger aperture (for any given correct exposure, whether it is TTR or not)?
My thinking is this; There is more change of having more photons hitting your sensor with a longer exposure than there is with a larger aperture. Is this correct?
Both result in more photons hitting your sensor - this is the reason you get the same brightness increase by opening up your aperture by 1 stop as you do for doubling the shutter time.
Discovery Of Light: How would an ISO-invariant vs variant react with some body cap on testing. It would be interesting to see the pushed results.
Dark frame subtraction is offered by most cameras for long exposures and is also used (I believe) in astrophotography.
To the best of my knowledge, this is mainly used to remove thermal noise a fixed pattern noise.
You're right that you'd avoid photon shot noise, but the problem is that you end up with an image made from an artificially low signal which will mean even the most minor source of noise would have a devastating effect on the signal-to-noise ratio, so you'd get *a* result, but not necessarily one that tells you anything useful.
Mike FL: Total light is fundamental to image quality, since it defines [the biggest source of noise in your images](http://bit.ly/shotnoise).
That isn't to say it (or equivalent apertures) should necessarily be used for calculating exposure, but to ignore total light capture is to ignore a useful means of understanding the behaviour light - which seems relevant to photography, to me.
Equivalent aperture is the f-number of the equivalent focal length. If you are happy to say that a 45mm F1.8 Micro Four Thirds lens is a 90mm equivalent, then it's similarly reasonable to then say that the effective aperture of an equivalent 90mm lens would be F3.6.
It's not true to say the 45mm lens *is* an F3.6 but it is reasonable to say its is *equivalent to* a 90mm F3.6 (on full frame).
Equally, you could say it's *equivalent to* a 58mm F2.3 on APS-C (there's no reason full frame has to be the comparison point, it's just the best understood one because it's become the standard by which fields-of-view are compared).
However, since each format has its own size/price/convenience/capability/quality balance, the easiest thing to do is to recognise/appreciate the balance of your system and not take all comparison to other systems as an affront.
@sans culotte - What aspect do you feel is a misconception? Which detail do you feel is incorrect (bearing in mind we've always said it's useful only as a means of comparing *potential* when deciding between formats)?
helltormentor: @ Richard Butler
Can you please explain the reason why you think fixed pattern noise and banding are the same. As far as I know, they are two different kind of noise. I would have attached some photos showing them if I could.
I'm surprised you're the first person to highlight this. I'll revise the text.
The problem with lens cap tests is that they end up testing unusually low signals and, if you expose for a long time, end up giving too much emphasis to thermal noise, which might not otherwise be a significant issue.
The question is: what situation are you hoping to model with a lens cap test?
stephenmattiola: I'm confused on the extra highlight part where it says it captured the same extra highlight but ultimately had to clip it to fit into the raw file, i can not seem to find where this happens. please help.
It happens every time you increase the ISO setting on your camera.
Set your camera in P, A or S mode and shoot an image at low ISO. Now increase the ISO setting by 1 stop. The exposure will be reduced, yet your Raw file and final image still clip at the same point as your original, low ISO exposure.
Now switch the ISO back down and match the exposure values that the camera chose for the second shot. You should notice an extra top of highlight information (and everything else 1EV underexposed). Why weren't those extra highlights in your second image?
Detail Man: Richard Butler and Rishi Sanyal:
These two articles about sources of image-noise are well presented and explained in your statements of clarification made in comments sections.
I like the way that DPReview writers appear to have the adopted the terminology "ISO-invariant" / "ISO-variant" in lieu of the (IMO, less descriptive) terms "ISO-less" / "ISO-full", and recall suggesting the alternate use of those particular terms in this post replying to "gollywop":
... when he was in the course of editing his to be published DPReview article here:
... and which he continued to use in his subsequent DPReview published article here:
The more information regarding these subjects that is made accessible to readers within articles published on the DPReview site (in addition to appearing in forum posts), the better !
That usage certainly precedes ours, even if I was unaware of it.
The awkwardness of talking about a camera being 'more ISO-less' or 'less ISO-less' than another prompted us to look for an alternative. I'm glad we didn't accidentally create a less good alternative than the one you'd already established.
Lhermine: Very interesting article ! Thanks very much DPR.
As many people here, I was wondering if the number of photons can be so low that shot noise would be significant.
Based on the so-called "sunny 16 rule", I've found that for a proper exposition on a 24 MPix FF sensor, around 50,000 photons should hit each pixel. This lead to a shot noise with an amplitude of 1.4 %.
This kind of noise amplitude should to small to be noticed. However, in darker conditions, you may have to increase the ISO speed let's say for instance 6400. It means that you have fewer photons. The corresponding shot noise amplitude will be around 3.5 %. We reach up to 14 % for ISO 100,000!
And that's the bad news: 1,000,000 ISO speed will never be as good as 100 ISO whatever the quality of the sensor because of the shot noise.
(for those who are interseted in the computation, please mail me, you may point out some mistakes ;-) )
WaltFrench - I believe 50-70% is a better assumption.
SolidMetal: Great article! Though I still cant really understand why high ISO images have less dynamic range. That orange line which is overamplified in the picture wouldnt be out of the raw file even if its straight not curved? I mean its still brighter than the highlights marked in yellow on the low ISO picture.
Note what happens when you roll your mouse onto the main image of the third diagram and then onto the blue text link below it. Follow the path of the brightest tone in your final image, back to its source (Marked 'Highlight' in the scene).
The result is that you still see highlights clip at the same point in your final image (if you change ISO setting and exposure together). But you've been throwing away any extra highlights that your shorter exposure was capturing.
So you *see* the change in DR in increasing image noise, particularly in the shadows, but it's mainly happening *because* you've thrown away a lot of what your sensor captured.
io_bg: What a great article!The edited RAW file does show a lot more than the unedited one however in my opinion both the highlights and shadows are pushed beyond the point of looking good. The whole image has a greenish tint and contrast is quite low.
There are elements of taste at play (and you're entitled to your opinion, of course), but the images in the article are merely examples.
The point is that you can gain access to that broad range of tones. How you use them is entirely up to you.
mapgraphs: Richard, Rishi,
When you make statements like "most cameras have low upstream read noise" it would be helpful if you cited your source for the statement. And the same goes for every other statement of implied fact. Your sources should be a readily available so that anyone can easily access them.
Otherwise this is just an editorializing opinion piece. If it is intended as an opinion piece, that's fine, but that should be set out at the beginning.
Upstream read noise is a variable value, which can be affected by heat and the software used to interpolate sensor pixels:
"Interpolation works by using known data to estimate values at unknown points."
"Image interpolation works in two directions, and tries to achieve a best approximation of a pixel's color and intensity based on the values at surrounding pixels." [ibid.]
mpgxsvcd - I don't think Rishi will take it as any slight if I refer to [the man who developed the CMOS image sensor](http://en.wikipedia.org/wiki/Eric_Fossum) as having more experience.
Mssimo: Question: As far as shot noise... Would a APS-C camera with a speed booster have the same low noise performance as a full frame assuming the same lens is used?
Not quite. For a fixed light source, the same f-number will project the same light intensity across any size of sensor (the same light per unit area).
The only time the small sensor is like the boy with the magnifying glass is when a speedbooster is used to concentrate more light down onto the smaller sensor. As I say, this is effectively an increase in f-number, since the aperture diameter is unchanged but the speedbooster has the effect of reducing the focal length (so represents an increase in f-number).