You can never display as much as the camera can capture. It's one reason why DR is over-rated. Not to say it's useless, it comes in handy if you want to do some creative editing on your files.
March 23, 2024
40
-
Mostly I agree with Doug.
-
you cant gain 2 stops more DR by using an extended gain. its not possable.
-
First, it's not extended gain, it's the same gain as 250 ISO. More gain would give higher ISO, all else being equal. Second, the read noise measures a bit lower (you can see that on Bill's chart too). Third, there's a stop more exposure. Remember, I'm going to ISO 100%, which is what you'd get metering normally, not 'FWC'. So, you get a stop more at the top, half a stop or so at the bottom, for 1.5 stops, which is what it says, not 2 stops.
Whether the RN reduction is due to noise smoothing, I don't know. Bill thinks that it is. -
Just kidding but there's saying here in Texas "if yew had'na tole me that, ah would never have known" ...
Indeed. For example: those who hate blown skies or are into HDR ... "If ah had'na tole you that etc." ... 😉
-
you dont gain 2 stops useing extended iso's 50 vers 100 just tested my a7iv ,in fact as it should you loose dr and information, its suptle but you can see the difference at 3000 % i wouldnt hesitate useing iso 50 on my a74 if needed though as its nearly identical.
-
As I said, it's not 2 stops, it's 1.5. And I explained where it comes from.
Sorry Don, that test tells you exactly nothing about 'DR and information'. It doesn't get to ISO 100% exposure and it doesn't go down to the deepest shadows. As was discussed upthread, you can't display all of the DR that a modern camera can capture - the very best display tops out at about 10 stops.
-
thats where your wrong, i did shoot a 100% DR shot and this is in the middle tones 😁 why ? because thats where DR is most important ,in mid tones, and you can clearly see camera manipulation to the file ☺im tired of seeing in camera raw file manipulated shadow noise and squashed highlights that everyone posts.
-
On the contrary, the middle tones tell you nothing whatsoever about DR. How could they? DR is all about the limits. Maybe you're confusing DR with SNR.
-
have a good think about what you just posted. your not looking at the big picture (pun)
do you actually think DR is black, or white. 😉
a small sensors black is a large sensors grey ,and a small sensors white is a large sensors grey, colours play out the same. -
I did before I posted it.
No, but neither is it what you think it is. DR is simply the ratio between maximum and minimum signal levels. It says nothing about what goes in between. There's room for discussion about what 'maximum' and 'minimum' mean.
Doesn't make any sense whatsoever.
-
+1
Donald, please read and understand:
ChatGPT:
The dynamic range of a digital camera refers to the range of luminance values that it can capture, from the darkest shadows to the brightest highlights, without losing detail. It's typically measured in stops, with each stop representing a doubling or halving of the amount of light. The wider the dynamic range, the greater the ability of the camera to capture details in both very bright and very dark areas of an image simultaneously.
-
I think you're missing the context of Bob's metric. He is only looking at "safe" and "standard/normal" highlight ranges that pretty much are available with all cameras and ISOs, and so the bottom of his DR is measured relative to standards, not the highlight limits of the RAW file. It's just like DPR's special versions of the Studio Comparison Tools that compare under-exposures; DPR never looked at the raw clipping point, and used that for a reference; they use an exposure standard, so they are demonstrating what happens with exposures in an absolute sense, rather than relative to raw highlight clipping levels.
When you allow the wild card of variable headroom into the picture, then you don't know how much of the DR is "footroom", which is a better proxy for exposure-referred noise, which is why I tell people not to use Bill's "PDR" as any kind of proxy for absolute noise. PDR attempts to tell you what you can fit (aesthetically) into the raw levels, independent of absolute, exposure-referred noise.
-
Correct. I developed it for AP magazine. The idea is a simple to use/understand metric of sensor quality, for normal photographic use. Most photographers meter by the camera, so if one bases a metric on the understanding that they will meticulously use all the raw headroom then most photographers will not see all of teh performance indicated. The full 'metric' is shown like this:
The major line shows SNR at 100% ISO, plotted log to relative to 1 (2^0). This gives an indication of how 'noisy' the image will look. The dotted line shows the SNR if the photographer makes full use of the raw headroom. The dashed line shows the 'Usable Shadow Depth' (aka 'DR') plotted relative to the major line. The Y axis legend is a bit non-intuitive for this - the reader needs to look at the difference between the two lines to work out how many stops below 100% is the usable shadow. On the other hand, it's visually apparent just by looking at the graph.
It's normalised, as I said to and A3 print at 360 ppi. The reference point for 'usable shadow' is SNR=2. This sounds very low to many, but as I explained before, you'll never see a lot of detail or tonality in the Zone 1 shadows (Zone 0 is black point), and if your reference is a 360ppi print at normal viewing distance, then SNR=2 looks pretty good (I've tried it), unless you're using an old 5D - then you have to whack all the pattern noise.
Mostly, people won't bother with the technicalities, but the graphic illustration gives a nice illustration of the 'quality envelope' of cameras. For instance, here it's quite easy to see that the extra DR at 125 ISO comes from lower RN (Sony does some crazy things with its gain settings), or 'in the shadows'.
It can even be used to compare, as follows.
Which shows what has been sacrificed for the global shutter. -
from my experience with cameras with multipul size sensors. is if shooting a scene the FF sensor with a wider/larger DR it streches the colour gamet over the whole image recorded where the smaller sensor squashes the colour gamet ( muddy images) including the mid tones, the mid tones are not the same between tiny and large sensors. the larger sensor will always produce a smoother graduation due to wider dynamic range.
-
i just took a very contrasty image with my xz1 and the a7iv, amazing little ccd camera the xz1.
-
It's the higher SNR that's doing that, not the wider dynamic range.
-
What have you done to separate the effects of total light from the effects of sensor size, in your decades-long photographic epidemiology?
What have you done to separate the effects of pixel size from the effects of sensor size?If you can only know about either total light or sensor size, which would you choose if you wanted to predict the smoothness of gradation?
-
"Contrasty" is a function of scene, lens veiling glare, and conversion math, as virtually all sensors are very close to linear in response over most of the raw range of digitized values.
-
neither "larger pixels" which is why i chose the 33 meg a7iv over the 60 meg a7r5 the whole crop factor is just to justify lack of practical sence.
i can shoot extreme macro with my 24 meg apsc a6300 and crop mode a7iv at 14 meg and no one could ever tell the images apart. even viewing at 200% -
define "capture". do you capture 9 stops from camera which is my thinking of DR capture. or do you manipulate the file in Ps and guess what is aceptable noise and call that DR. you have 1 choise, lets see which is right 😁
-
If true, that's a corner case that has no relevance to most detail-prioritizing photography. It may be more like shooting at f/64 at typical distances with the range of pixel densities from the m43, APS_C, and FF formats, all of which should be well into in the "oversampling" range for f/64.
-
I think that john was looking for a systematic approach to the investigation, defining the variables and determining separately the effect of each.
