jrista said:
There is noise, and there is editing latitude. In pretty much every thread on here that ends up with the DR debate, what do people talk about? The amount of noise they can see, or the amount they can lift the shadows? In all the threads I've been party to, it all ultimately comes down to how much you can lift shadows.
In none of the DR debates I've ever been party to has anyone ever said "You don't see as much noise with an Exmor sensor." No, the thing everyone always says, and the thing everyone always tries to demonstrate, is "Look how much I lifted the shadows! Look, I have a fully detailed sun, and detailed shadows, in this landscape photo. Oh, and look over here, the Canon sensor has tons of nasty red banding noise when I lift the shadows."
As far as I can tell, as far as most consumers are concerned, DR all boils down to EDITING LATITUDE. It means more shadow lifting.
Yeah and the greater the engineering DR Print DR measurement for one camera compared to another the greater the degree you have in editing latitude when the goal it to compare a final result at the same scale.
So, why do I treat them differently? First, you are not wrong about comparing cameras...you need to normalize. And normalization affects read noise as much as it affects photon shot noise. But that is comparing final IQ. That's fine and dandy...but I believe people are misusing final IQ (of UNEDITED images) to refer to editing latitude. It's THAT, the use of normalized images to refer to editing latitude, that I believe is wrong. I believe DXO's publishing of Print DR exclusively on their ratings, completely ignoring Screen DR entirely, has lead to the misinterpretation of REAL WORLD editing latitude in a RAW.
Once again as we say and they say if you want to view at 100% and see what you can do use their Screen DR.
Sure you are correct about that.
But then you turn around and call anything BS when it is not. If you want to compare one camera to another and find out the relative difference in what you could get out of it you use the Print plots instead.
And you realize that you can use the same arguments you try to use in a weird way against Print DR against PrintSNR. Why not just say they edit the RAW at 100% view so the amount of noise they have to deal with and NR applied and sharpening applied is determined by that so forget about PrintSNR and just use ScreenSNR? Why not penalize high MP cams for that too? Why penalize high MP cams for DR but not for SNR? (or better why penalize them for either!)
Since everyone always ultimately arrives at "pulling shadows" in DR debates at one point or another, I'm always harping on that point. More DR means less noise, and normalizing means even less noise for larger sensors, but we cannot edit normalized images. We only edit full size images.
You don't have to, your first step could be to normalize the image and then edit, if you wanted to.
So, from an editing latitude standpoint...I believe Print DR is invalid.
With that twisted point of view why from an editing noise appearance is PrintSNR not also invalid then?
And in that case though you magically change your mind and agree that using ScreenSNR would unfairly penalize high MP cams. Is that because you'd then end up having to claim that a very low MP ASP-C camera might be 'better' for noise than a very high MP FF camera (both using the same tech)? But with the way read noise tends to work you manage to less easily hit such paradoxes???
For one thing, the Print DR numbers at DXO are only valid if you downsample an image to exactly that size.
Of course.
Nobody says to care about the ABSOLUTE normalized numbers each in isolation. All we care about with normalized values is carrying out relative comparisons. To be fair they have to pick some size to normalize to unless they were to make up tables for every single sensor difference (and to what point?), what size and what absolute numbers come out who cares? All you care about is the difference.
I don't think many people actually downsample their images to exactly 8x12 @ 300PPI all the time...hell, I think it's actually probably quite rare. So always referring to 14.4 stop of DR when discussing shadow lifting ability is plain and simply wrong in all the infinite other possibilities for image size.
Where I am talking about 14.4 stops of shadow lifting latitude? First of all engineering DR doesn't account for banding so if any of that was around the usable amount would be less. Second (really first), you probably wouldn't be satisfied with the engineering level base term and probably wouldn't want to push that low. And finally the absolute number you can push to also depends upon the output scale and detail you are going after.
So sure, in probably very few cases would you feel you had 14.4 stops to work with.
But that is NOT what we are talking about, we are talking about relatively how one camera fairs vs another.
And that's not even mentioning that when it comes to the types of photography that tons of DR are most useful for, say landscapes, your probably UPsampling, rather than downsampling, which makes it even more invalid.
not if you want to know how one camera would relatively do comapred to another
sure if you are talking waht absolute value of stops you get in editing latitude the absolute PrintDR isn't so useful, but then again, you probably won't agree that you have the ScreenDR number either even if working at 100% view on a 30" 1080p monitor since you'd likely think the low parts that the engineering value uses are total junk.
If someone uses ScreenDR and says oh look this digital sensor here is giving me 13 stops so much more than I felt I could ever get working with film, that is misleading. But were did I mention saying that?
If you are pointing to ScreenDR or PrintDR and say I can get that many stops latitude it's not likely you'd feel that working with the image.
But if after normalizing that high MP camera comes out a few stops better under PrintDR but about the same under ScreenDR then viewing both at 100% ScreenDR would tell you the relative story, but if you want to know if that high MP would give you a worse result printed to the same size and viewed from the same distance then it would be misleading and imply you wouldn't do better when the PrintDR would tell you you'd do better.
So yes, normalization will reduce all noise, including read noise, we can't downsample our RAWs...we must edit them at native size. If we take DXO's Print DR numbers as a reference for editing latitude, they would have you believe that you have more than an additional stop of editing latitude with a D800, and nearly two additional stops of editing latitude with a D810. That is plain and simply false. Hence the reason I treat read noise uniquely in the case of editing latitude.
You must edit them in terms of the mid-tone noise measurement they take at full RAW too (unless foveon in which case that is not true, you could downsample and still keep RAW, even for Bayer you could if you downsampled to very particular sizes but whatever). And here you agree you can't fairly compare them at 100% view. So why can you not fairly compare them at 100% view for noise at say mid-tone (just using mid-tone gray since that seems to be what DxO uses for their SNR plots) but you fairly compare viewing pulled shadows at 100%?? What is the difference??
So, I don't believe that read noise frequencies are consequential to normalization...you can't see them anyway.
First why are you calling them read noise frequencies? You could sample over large blobs or individual photosites in the darks or mids or brights.
And second if you say that read noise only happens in such dark tones that you can't see any difference regardless of what the read noise is then how come even you admit that you can pull shadows better with an Exmor sensor??
That leaves photon shot noise as the primary noise culprit we are dealing with when normalizing images for comparison.
Yeah much of the images tones are affected more by shot noise, but not at all. So you look at two different things and they give you two different plots. What does this have to do with anything under discussion?
Sure, read noise frequencies get normalized as well, but only a computer algorithm can tell the difference,
First, I think you are getting confused by mixing up spatial frequencies with something or other? Not sure why you are referring to it in that way.
If the signal you are looking at is so low that the normalization has no effect that you are looking deeper than you are using anyway since you won't spot any non-normalized differences anyway I'd think.
You might say well maybe the eye finds anything in the realm where read noise starts affecting the signal too deep and dark and a mess to care about after raising shadows, which could be, only it doesn't seem to be since the differences in the shadows can very clearly be seen so the read noise is stomping over stuff that is visually useful and it seems to match up to about what is suggested. So it doesn't seem to be the case that the measurement is looking in too deep at this point.
The only time read noise becomes a meaningful factor is when your pulling shadows. THEN, and only then, does the advantage of having LESS read noise really become a meaningful issue.
Well, yeah of course.
In that case, the D800, and any other Sony Exmor based camera, wins, hands down, no contest. However, and here is where DXO comes in again...A D800, D810, A7r, A7s, etc. don't have an 8x light gathering advantage over a Canon camera (as DXO's PrintDR numbers would have you believe). In fact, its about HALF that, one stop less, or a 4x advantage. Personally, I think being off by 100% is a meaningful thing. If DXO was saying the D800 had a 4.1x advantage over Canon cameras, I'd have never said a peep. But saying the D800 has nearly an 8x advantage over Canon cameras...yeah, I have a problem with that.
Because you decide to not normalize for shadows but only for upper tones for some reason.
The difference seems pretty major to me at ISO100. I mean really quite, quite, jump out at you so and 1 stop certainly doesn't produce that and even 2 stops not really, so the getting close to 3 stops seems reasonable to me although I haven't tried to do a careful study with custom uniform RAW converter that treats them all them same and then visually comparing by stop. But just what you see using say ACR it seems reasonable although it's tricky to get clear measurements out of that because so many variables.
So...I keep read noise levels in the context of discussions on DR (which pretty much ALWAYS end up referring to shadow lifting ability at some point), distinct from the whole concept of normalization. Because were talking about editing latitude, something that cannot be compared in a normalized context (at least, as far as I see it.)
Well, I don't think I can explain my stance any better than that. I'm guessing you still disagree, but that's ok. Nothing either of us can do about that at this point.
:'(
Anyway I'm sick of this, so I will quit, you can have any last word if you wish, have at it. I've become bored and didn't mean to waste so much time on this response.
