57even: I think I am right in thinking that larger photosites generally have more read noise and less shot noise. They will look better in mid-tones, but if you push the shadows in post, noise will show up.
This would normally mean that the 5DR would be better than the 6D, if it were not for the noisy ADC that Canon's use. However, high ISO (ISO variant) shots indicate that upstream read noise is actually really low. What's messing up the low ISO pushed shots is the downstream read noise.
In other words, the DR problem is only an issue at low ISO, and the reason is the same as for all Canons, namely they don't use dedicated column ADCs like Sony do on the smaller pixel chips (D750 and D810).
Thanks for that. Yes, you are right. But the difference seems quite large. Could it be due to other factors as well. For instance:
The D810 is ISO invariant above ISO 1600, the A7S is not.
The A7S may have switchable gain on the photosite, reducing upstream noise at some point (higher ISO). This seems to be true of the Nikon DF.
The trouble with sensors is that there are so many design considerations that it's hard to make any assumptions.
I assumed the 5DS was capped because the ADC is trying to cope with such high data rates, but there could be any number of reasons I guess.
But I am not sure I see the advantage of a high resolution sensor with limited low ISO DR and amazing high ISO performance!!! The Sony A7R and A7S seem to have the balance right, with the A7 right in between. Same for Nikon D810, D4 and D750.
I think I am right in thinking that larger photosites generally have more read noise and less shot noise. They will look better in mid-tones, but if you push the shadows in post, noise will show up.
Skipper494: All very fine, but in film days we got grain from particles in the film, nothing from die film. As an engineer with quite a lot of experience in sensor design, I can assure you that most noise comes from the proximity of photo sites, or wiring, (even poor software) which is why lower resolution large sensors have less noise and more dynamic range.
Compare the Nikon D700 and the Sony NEX 7, for instance. Consider how good the Fuji S2 Pro was in it's day. Compare a lower resolution 1/1.7 sensor with a higher resolution 1/2.33, such as the Nikon P7000 with the Pan FZ35.
Frankly, too much emphasis is put on high numbers and marketing, instead of good photography results.
Yes, I know what gamma correction does, but it doesn't have the effect you are thinking of in this case - the conversion expands the encoded shadow values, but the quantisation has no effect on the accuracy of the values because the bit depth is much higher than the number of noise limited tonal values - there is no accuracy loss. Besides, it is compressed again by the display anyway. Zero net change.
I think you should read it again....
16 bit editing space gives 65,536 possible values to interpolate however many tones there in the actual image when you apply a contrast change. That's a lot of possible interpolation.
Don't know where gamma comes into this, as Iliah explained. The regular tone curve compression only exists in the raw editor output, not the raw image, and if you mess about with it, a 16 bit editor has plenty of possible values to fill in the gap.
In terms of tonal range there are actually far fewer data points than the bit depth can record. It is massive overkill. Even 10 bit quantisation doesn't reduce the number of tones. The number of data points is limited by signal noise.
Quantisation only matters, a bit, for dynamic range because of quantisation error, but that is usually not much higher than the inherent read noise, so the real loss in DR from quantisation is mostly theoretical. The problem is largely ADC noise at low ISO.
klausious: so I have a question: if i use the same lens 50mm at 1.8 on two different cameras: one is full frame and one is APS-C, beside the fact that the full frame will have a wider angle of view, will the image of the full frame be any better than the APS-C because keeping the same lens and same aperture will keep the aperture diameter the same and so to my understand that the lens will allow the same amount of light pass through the aperture, only that the full frame capture more of the scene and the APS-C doesn't. The comparison in this article really doesn't help me because two camera and two different lens was used, and it was confusing for me. Right now i don't care about the angle of view. I just want to know if with the same lens and same lens setting, putting on two different camera system, which one perform better?
The full frame sensor will have a larger area, so you either get more pixels of the same size, or larger pixels.
Larger pixels have less noise because they record more light and shot noise is = square root of signal.
However smaller pixels means more resolution. If you downsize the image to the same size, then you are combining the signals from the smaller pixels, so the result is the same. The larger sensor has less noise.
In other words, bigger sensor = more light = less noise.
It doesn't matter what lens you use if you use the same exposure for the same scene. It's the total number of photons that you record that matters.
But you don't get more tones. You just get noise. Analogue vs. digital won't make the slightest difference because you have to convert to digital to print it.
If you edit a 16 bit virtual file (in an editor) the editor can 'interpolate' far more tonal graduations than a screen or print can display, and printers naturally 'dither' the output to hide banding.
Noise would be the same whether the image was analogue or digital, except that you still have to convert it to digital to display it or print it.
As it is, there are about 400-500 greyscale tones in a raw 14bit file, which is more than most people can distinguish on a screen with a contrast ratio of 800:1, let alone a print.
Not much point in having more of anything than you can actually see.
martindpr: Most important thing is the type of those test tubes you got there/their efficiency. Cause 2day with our tech and translated in layman's terms, they have a small opening, bad angle towards the rain and those raindrops evaporate too soon too early before reaching the bottom and collecting in the sampler. And all manufacturers use this same bad recipe. In more technical terms, the color filter reduces at least 2/3 of light (practically 95%), the CMOS samples only 1/100 of the light transmitted by the filter and rest is reflected out or lost in the process, and then at least 30-40% is lost in the circuitry. Then the jpeg looses additional 80% of the information and we're left with 1/1000 of the light information of the original scene. Solution? Refractive type of color filter and graphene instead CMOS for no visible nose at any sensor size.
Not necessarily. PRNU noise can reduce tonal range in highlights, read noise reduces DR and tonal range in shadows and thermal and dark current noise cause problems on long exposures. All forms of noise have their different effects.
If the SNR is above about 35dB, it's pretty hard to tell the difference. Post processing can make more than 5dB difference, even dropping the AA filter and reducing sharpening can make more difference.
57even: One big problem with ETTR - channel clipping. The histogram may tell you when the JPEG is clipping, but it usually uses the green channel. If you have a three-colour histogram, it's easier.
Which makes it all a bit haphazard IMO. You can easily blow the red or blue channel without realising it unless you allow for a fudge factor.
I seldom push the highlights up against the stop - you always find parts of the clouds are blown. I normally back off at least 1EV from the clip point, if not more.
You mentioned DR, you didn't say what ISO. Have a look at DXO mark and check the D7200 vs the D750. Exactly the same up to ISO 400, and never more than 1/3 of a stop behind until ISO 6400. I reckon the lens/VR would turn that into a rounding error.
Compare it to a 16MP FF camera, like the D4, and the bigger sensor is more than 1 stop worse at low ISO where I need it most, and doesn't equalise till ISO 800, but doesn't reach a stop better till ISO6400.
Since I shoot around 95% of my stuff at ISO100-800, this doesn't seem all that relevant in DR terms. Sure, I know the SNR is better, approximately 3.5 dB, but only at higher ISO, and tonal/colour range is no better at low ISO. Plus there is more resolution to play with on the D7200. That high read noise clobbers tonal range/colour response at low ISO as well.
Big pixel DSLRs are specialised gear, and certainly not the primary choice for daylight shooting.
One big problem with ETTR - channel clipping. The histogram may tell you when the JPEG is clipping, but it usually uses the green channel. If you have a three-colour histogram, it's easier.
So how come the D7200 has more dynamic range than the A7mk2 (let alone the 5Dmk3).
TakePictures: Fuji take their time and bring out one gem after the other. True 35-mm thinking in the APS-C world is rare these days. It requires some rethinking in terms of DOF, but fortunately the story is the same regarding light gathering. To me this lens is an optimal tradeoff between those two factors (given that the lens needs to stay reasonably compact).
The trouble with DOF equivalence statements is that some lenses work really well at maximum aperture, and some have low contrast and bokeh fringing and are almost unusable.
I regularly have to stop fast lenses down to F2.8 to get rid of these issues, which rather changes the equation.
Similarly, 'equivalence' ignores the stopped down situation, where larger sensors need stopping down more for the same DOF, with consequent issues for shutter speed and ISO.
In fact, it really only matters if you like portraits where only one eyelash is in focus, which not something any serious photographer would routinely do.
SnakePlissken: Someone needs to tell Harris that Nikon, like Canon, are finished and will be in liquidation very soon and he should sell all his equipment and go mirrorless. Does this pro not read DPReview message boards or what? Get with the times Harris!
If you think it's fun to carry that lot around, good luck to you. Not arguing that different people have different styles, but every kilo of dead weight I can shift without losing image quality is a good thing, and I don't really notice the speed difference in single shot mode. CSC cameras are improving all the time, and the best have very rapid AF.
57even: If Samsung want sales in the UK they had better sort out the dealer and supply issue, and provide a decent after sales service. Sorry, but I am not one to buy expensive gear off the Internet, I want to buy from someone I can talk to if there is a problem.
At the moment, they are only available in big white-goods stores, with a very limited array of accessories. Not a viable option for serious users.
I don't need a camera store to tell me what camera to buy, I want them to provide a professional service in terms of supply and service.
You can deal with the manufacturer of course, if the product is under warranty, but I assure you it all goes much better if the dealer conducts the negotiations for you. The fact that they can back up your version of events helps a lot.
Virvatulet: Many seem to support the idea that one could arbitrarily dictate what others can do in public places. Using a public space inherently conveys an accord on limited privacy; the very prospect to be heard, seen and remembered by others.
As long as no physical change is forced on one's personal proximity, there really is nothing what one should do about other's freedom to choose how they use the immediate free space of theirs.
Passively collecting stray photons with a camera can not be objectively seen as excessive behaviour. It's also vital to remember to separate photographing from publishing (legislation variant). What can be legally seen shall be legally photographed (a priori), regardless of the circumstances affecting lawful publishing.
The ladies in the picture exercised their right to choose how they want to be seen and remembered when they use public space. Not willing to give even a split-second moment of their day to the memories of a gentleman randomly crossing their path.
I don't always ask people, but nor would I embarrass them. If one of their friends saw them on my gallery, there would be nothing compromising about the image in any way.
Pervs wandering around beaches looking for topless bathers and posting on youtube are making life a real chore for the rest of us.
The real advantage of specialist dealers is that they have a hotline to the manufacturer for accessories and supplies. If something is not on Amazon I am sunk. Big box dealers just shrug and say 'we don't stock that item'.
The NX1 is aspiring to be a pro-level camera, and should have pro-level backup and support. That's all I am saying. One key feature of such support is a competent regional repair centre and a pro dealer network.
I would insist on that if putting £3000 on the table for a camera system.
There is a fine line between offense and reportage. It is also very dependent on the society you are in. In some countries, images are seen as intrusive, in others they are common place.
For me, good manners usually dictates what I shoot and publish. If someone objects, I don't push the issue, neither do I publish embarrassing images of people. I once unintentionally caught a girl when her skirt blew up around her waist, but I would never dream of publishing it. In fact I deleted the image.
Other people would have it all over youtube in a minute.
Social media has dragged public photography into the privacy debate. You may not like it, but expect bans in more countries as a result.
I missed that bit, thanks for the heads up.
My experience of street shooting with a Fuji XE2 vs a Nikon D800 is that no-one notices me when I point a Fuji at them. Wave a Nikon in their face and I get reactions much like those Spanish ladies.
I concur with the sciatica problem as well. I am in my 50's and many years of carrying a DSLR kit around has taken a toll.
Frankly, not that it matters, but I can't see a single shot here that could not have been made with a mirrorless camera.
Clearly he is going to use his kit hard, and needs something that can take a battering. I'm not criticising his choice at all, but for most of us part-timers this is overkill.