Equivalence of fnumbers which relate to areas and equalisation of input NOT output
Your game is really childish. Of course this is intentional to keep this ridiculous history, maybe you think it is funny and you want to break a record but this is not nice.
Dpreview is going to save these useless discussions for many many years, this is unreadable and counter productive.
I agree. This thread and some others get completely trashed and so off topic as to render them useless.
The perpetrators of this behaviour should start a separate thread somewhere and continue to show off and argue there.
This current behaviour is REALLY REALLY boring and REALLY REALLY destructive.
I posted some data about a camera I have and I have taken measures
At some point someone made a cross system comparison and that kicked off a total beating of photonstophotos which is not justified and unsubstantiated
the method has got some specific metrics which is something you can or not agree however the full data sets are there if you want to read them
lens fstop dont match transmission an iso doesnt match the standard values and so what?
nobody bothers an f/1.4 lens is f/1.8 but talk about iso and people get a fit
weird
Personally i find photonstophotos really useful as a repository not just for sensor but for lenses too and I thought this was shared view here but looks like it isnt
i am guilty myself of taking this into a rabbit hole responding to off topic with more off topic however the data is valuable to users out there and hopefully people keep taking those measures
You misunderstand the sentiment here.
Photons to photos is extremely valuable. For all kinds of research. There is sensor information you find nowhere else. Same actually for their not very well known lens tool.
It is simply not designed to compare the high ISO performance of cameras, which is sadly what people are trying to do most with it.
Sorry but photonstophotos allows a user to make the right decisions on low light high iso better than it does for bright scenes in fact.
Not that I want to argue with you (again), as it makes no sense, but some people might be reading this and believe you.
photonstophotos' DR measurements are completely useless for comparing noise of various cameras at high iso because of two main reasons. First, as panther fan says, due to the unknown amount of gain the results for read noise can't be correlated to exposure levels. Simply said, the same ISO on the x axis can mean very different exposure levels in practice and thus you can't say how much the read noise will impact the image, as you have no idea about the SNR ratio of the signal to read noise. For that, normalizing nominal ISO (gain) to exposure levels is needed, which is obviously not possible for these simple measurements. And that is connected to the second issue.
No because the average user relies on what the camera says nobody adjusts anything. So who cares. What matters is the settings you can control as a photographer not what it should be especially as raw converters adjust for the tuning.
An average user (typically relying on auto ISO) would see two possible effects (or their combination):
- an explicitly different ISO in the same condition with different cameras, even with the same brightness of the OOC jpegs (i.e. a different ISO implementation). Comparing at the same ISO thus makes no sense, one image would be darker for both jpg and raw.
- the same brightness of the OOC jpegs, but darker raws from one camera with more highlight headroom (i.e. a partly digital ISO implementation). Comparing converted raws without matching their brightness thus makes no sense.
That's why the dpreview comparison scene not just tests cameras at the same ISO, but quite crucially adjusts the brightness of the converted raws. This normalizes the output for different analog gain, which photonstophotos does not do (and can't do, as he does not have images taken at controlled exposure levels from the various cameras - they are provided by users around the world).
, but it seems this way
