It's not been demonstrated that there is a loss of DR.
Well, I cannot prove it, but I think I stand in the middle ground in
this issue. I find the posts of you-know-who aka. "anti-superstar"
ridiculous concerning the total lack of DR below APS-c, but I cannot
help notice the difference immediately when I see s strip from a
small sensor DV-cam. No, I cannot tell the size of the sensor of the
DOF, but the DR is very obvious.
But sensor SIZE does affect image DR, I've never said anything else.
(But less than, yes, I-know-who, says.)
But given a sensor size, then the pixel size may not affect image DR,
particularly in as how much details you can make out from the noise N
stops below clipping in the final image, viewed at a fixed size.
Cameras ought to have hardware support for quick downsizing.
Then you could shoot in a low resolution setting at a higher speed.
Wait, don't the recent Lumixes do this? Maybe the G10 doesn't.
Are you saying that previous cams only used a fraction of the readout
for the low MP settings???
No, not at all, the keyword is "quick". It seems many cameras aren't
faster with their reduced-size files. This could be because the time saved
by having less data to write to the card is lost in the time it takes to
downsize the data.
Of course, reading the data off the sensor also takes time, so that puts
a limit on how fast bursts can be had, and is an argument against high
Mp counts.
But it is better people use this valid argument against high Mp counts, than
unsubstantiated ideas about DR and noise.
The recent Lumixes have extra-high burst rates at
lower pixel count. I hope they use all pixels to start with but I've not
seen anyone test this.
OK, this is where my layman's logic seize to cope. I understand
'image-noise' vs. 'pixel-noise' but what's "image-DR" about? IOW, are
there other 'kinds' of DR?
Image-DR is not a well-defined entity, but it's a useful IQ characteristic,
more relevant than the better-defined pixel density.
Everything else being equal, if we have more, smaller, pixels, each pixel
will have less DR. (DR defined as clipping level divided by the noise floor.)
But if we downsize the pixels to a lower-Mp image, each new pixel will
have less noise, because the noise averages out. So the DR increases.
This is the point people are missing.
But we still not doing the higher-Mp image justice, since the downsampling
is throwing away resolution. And what we care for is how the details
compete with the noise in the deep shadows. More pixels will give more
subject detail and a finer noise pattern.
E.g. at ISO 1600, the 400D has better noise performance, also per area,
than the FZ50, due to Canon's good high-ISO.
http://forums.dpreview.com/forums/read.asp?forum=1018&message=28760503
But still the noisier, higher res image of the FZ50 looks better (step away a
few m from the screen). Sharpening the 400D image (USM 500/1/0) gives
a similar apperance of sharpness, but now the 400D image looks noisier.
This image doesn't have lots of deep shadow detail, but the principle
holds for deep shadow detail/noise and thus DR, too.
other manufacturers don't have a same generation 'high
IQ - low mp sensor' and 'standard IQ - high mp sensor' in the P&S
category. The Nikon 12mp FF vs. 12mp APS-c would also be interesting,
but then again the electronics come into play more than presumably
between the LX3 and the FX150.
Yes and the FF isn't made by Sony.
So essentially you are saying that the industry (still) hasn't
reached the 'optimum' resolution in the MP development, picture DR
and noise is are not harmed, all that's needed is better electronics
and lenses. Fair enough, but this is seen from an engineering POV.
I'm saying that it has not been demonstrated that higher pixel
density (upto today's P&S pixel pitch) is harmful for (raw) IQ. I haven't seen
a single case. Whereas I've seen several that suggest the opposite.
Sure, more tests are needed to draw more general conclusions, but I think
it's high time for DPR and everyone else that thinks that high pixel density is
"terrible", to produce something to support that.
Seen form a economic/ consumer (assuming value for money preferences
towards better IQ) perspective, pushing the current MP further would
require better lenses and electronics to benefit from the additional
sensor resolution obtained by smaller pixels, but the additional cost
of these does presumable not yield anything in customer value/ price.
IOW, the consumer is better off with a lower price then another notch
up in resolution. The lower price could of course be translated to
other attributes like durability, waterproof, LCD, EVF, build
quality, GPS, gizmos etc.
I agree. These are good arguments against higher Mp counts. But when I
see bogus or dubious arguments, like loss of DR, I can't help but protest.
There's not much
extra effective resolution from a 15mp vs. 10mp if these sensors are
behind a lens which is soft already at 10mp.
There's a big distance from the pixel density where we begin to see softness
to the point where no further visual resolution improvement is possible.
But I agree that we seem to get more and more limited by the lenses
in the compacts.
Just my two oere
Erik from Sweden
