Usee: Sorry, but within the comparison tool it wipes away the pattern in the green area around the Nike logo on the shirt of the left player - even at the "Large" setting...
OK, it reduces the file size noticably, but I would have liked a comparison with JPEG 2000 and also different jpeg compressions...
...I see only examples of "noise reduction" with a heavy loss on detail.
They should at least improve the comparison tool, or maybe also, the format......b.t.w. which is the reason why JPEG 2000 isn't used that much?
The biggest technical problem with jpeg2000 was that it took a lot more CPU cycles to generate than to view.
Ken Rocket: how many shades of gray does it have? I want more than 256.
14 bit means16,384 levels of grey. Should be enough for casual use.
sportyaccordy: For whatever its worth, in the studio comparison anyway, the D810 gives nothing up to the 645Z on a 100% view basis as far as noise goes. Def want to see its DR capabilities as well as ability to pull info from dark/underexposed areas. DxOMark's data can't come soon enough. Not in the market, just want to see what it can do.
@falconeyse: No problem. It looked like a copy-paste error, I just wanted to figure out which of the two of us you were addressing.
@HowaboutRAW: Not sure what you mean by "weakness" of audio amplification with transistors - it's about as common as sliced bread.
I take it, by "truly digital imaging sensor" you mean putting a single photon counter into every pixel. Yes, that *is* a long ways off.Today's single photon counters are the size of a compact camera, and cost many times more. It will be a while before they figure out how to cram one into each pixel.http://www.excelitas.com/Pages/Product/Single-Photon-Counting-Modules-SPCM.aspx
"Um, I thought transistors step discretely, unlike vacuum tubes"No, that is not the case. A transistor can be fully "open" or fully "closed", but gradually increasing the driving voltage causes its resistance to decrease in a smooth curve, not as an abrupt tradition.
If not for this smooth response curve, it would have been impossible to use transistors in amplifiers.
You should be able to confirm this with a quick Google search, or a wiki article.
Transistors are the basic building blocks of electronics. One can use transistors to build both analog circuits (like amplifiers) and digital circuits (like logic gates). But a transistor, by itself, is analog, and its response is a smooth curve.
HowaboutRAW:> "The photo receptors on sensors are transistors"No, photodetectors are not transistors. The two are fundamentally different. Photodetectors generate electricity when hit by photons. Transistors change resistance when hit by electrons.Photodetectors have linear responses, transistors don't (look at their response curves in Google images).
However, both photodetectors and transistors are analog devices. Their response curves are smooth, not stepped.
falconeyes:You put my name on top of HowaboutRAW's text. Why?
HowaboutRAW:Sorry, I mistyped, I meant ADC, not DAC.
Yes, photons hitting a sensor are discrete events, but sensors *function* as analog devices. There is no counter that would get incremented every time a photon hits it. There's no way to walk through each atom within a sensor element and total up their electrons' energy levels.
Just as they did when were first invented, sensor elements accumulate electric charge, which is then transferred to the edge of the chip, amplified, and then measured by an ADC converter. That charge is proportional to the number of photons that hit the sensor, but there's no way ADC could measure it with enough precision to count individual electrons.
Sure, a sensor is not a solar sell - but an individual sensing element is very similar to one.
"All raws are cooked a bit by the processor" - I must have missed the day this became common knowledge. Could you post a link?
HowaboutRAW: The only way a processor could affect raw DR is by being "quiet" electrically and cold (meaning literally, temperature-wise) while the image data is being captured and DAC-converted.Well, unless we consider "cooked RAW" - with some NR applied before JPEG conversion - but I thought FF cameras didn't do that, as a rule?
falconeyes:Aside from CMOS vs CCD differences, there's the matter of electromagnetic noise. Nikon is very good at keeping the camera "quiet" electrically. There were DSLRs (sorry, don't remember which off the top of my head) where Nikon managed to get a stop more DR than Sony - using the same sensor.To sportyaccordy - I think, the formula is just falconeyes way of saying "DR goes up with sensor area and goes down with ISO". It's a very rough estimate that only works (kind of) because the sensors' base ISO doesn't change very much (unless you want to mention that Kodak DSC 14n that had ISO of 6 :) ).
"What is the denominator in this formula" - I assumed, base ISO.
Jogger: Things will get interesting once the next generation of high resolution Sony FF sensors come online. Scaling a 24mp APSc up would give you about 54mp.
"There are dynamic range advantages to bigger pixels"
For the reference:
D810 pixel width = 0.04891mm645Z pixel width = 0.05329mmD750 pixel width = 0.05984mm
falconeyes:According to your formula, 645D's DR should've been the same (33x44 / 100 = 14.5). And yet, DxO measured it at 12.55:http://www.dxomark.com/Cameras/Pentax/645D---MeasurementsWhat leads you to expect two stops DR improvement in 645Z?
The Davinator: I agree. Fast lenses are more needed for low light and shallow DOF scenarios.badi: Agreed as well. If you have to get those 50Mpix in a single shot, your options are limited.
Chances are, Pentax DR will be behind the D810 by at least a stop. Resolution is the *only* advantage you get.What you lose though, is the fast lenses. I don't think there are any modern Pentax lenses faster than f:2.8 (equivalent to f:2.24 on FF).
fmian: Hopefully they will fix the shutter shock issues.Otherwise what good are 40mp photos if they are blurry?
Electronic shutter fixed that problem a while ago.
So, is this just for backup/storage, or for photo hosting too?
If I have pictures stored on Amazon Cloud drive, can it create a gallery visible to somebody who is not logged in?
To all you complaining about noise in ISO 200 shot (the Seattle view):
On a bright sunny day, the "sunny-16" rule says that for F/8, ISO200 the exposure needs to be 1/400 or thereabout.
This is sunset, so it's a couple of stops darker. Proper exposure should be at least 1/100, probably longer.
The shot used 1/3200. That's FOUR TO FIVE STOPS under (at least).
Then it's pushed up 4 stops in postprocessing.
The comment says it's done "to protect highlights".
Menneisyys: Regarding the new uploads,
3040827 – In the new images too, shadow noise is rather high even in base ISO - see the white/gray rocks' surface in the lower right corner. This camera in no way can match the noiseless-ness of full frame or Fuji crop cameras at base ISO.
3040823 – in sweep panos,
1, exposure changes could be a little bit less obtrusive (see the subtle exposure change steps in the sky) and 2, more vertical pixels used (I assume this camera also shoots at either 1920 or 1080 pixels high panoramas, just like Sony's cameras).
These may mean you still won't want to stop shooting individual images and stitch them on the desktop.
3040827 is coping with extreme contrast. Background is in bright sunlight, the subject is in deep shadow. The face is *massively* underexposed and then pushed up several stops.
It's Panasonic's "Intelligent exposure" DR expansion at work.
stevens37y: "solid block of aluminum"They should use titanium. Looks and feels much better.
But with enriched uranium, you'd get a much brighter flash!