Discussion in 'News - Discussion' started by CSBC, Sep 22, 2016.
I started this sub-thread in a genuine way wanting to know what this apparent metric signified but nobody has rushed in with an explanation so I think I have to agree with Andrew that it is probably nonsense
I do apologize I put across what I was trying to explain badly using "noise ceiling" phrase.
Every sensor has a point where above a certain ISO the raw image will have some noise which then gets worse as the ISO climbs.
DxOMark do tests for this which they refer to as "low-light ISO"
Now if you look at their test say for a Nikon D3s (12MP) its low-light ISO is 3253 where as the new Sony A7R II (42MP) is 3434
It shows the technology progress in capturing more photons on new sensor and improving signal processing.
This progress is still on going and is there in the 1" sensor area as well.
The Nikon 1 S1 (10MP) has a low-light ISO of 397 but the newer Sony DSC RX100 V (20MP) has low-light ISO of 586
Panasonic have actually crack the problem back in 2013 but have been freezing development in a new sensor.
From what I have read it would give at least a 1 stop improvement so the low-light ISO would jump from 586 to 1172
Of course Panasonic are in the MFT business at the moment LOL
Thanks for your response you have convinced me that any references the metric "Noise ceiling" is best just ignored
While we're on this subject, let's just be clear on one thing - chip fabrication is a numbers game. The more slices you can take out of a wafer the higher the profit (or at least the lower the loss). That's why designers concentrate their efforts on producing the smallest chip that will do the job, thus maximising the return against the very high gross costs, Full frame and larger sensors are an abberation in this context because not only do you get fewer slices from the wafer, you also increase the wastage. The best return comes from improving the smallest sensors and so that's what all the reports say they're doing. 50MP FF sensors provide a marketing benefit but probably add far less to the bottom line than an improved 1/2.3 sensor that sells to several phone makers as well as going into compact cameras.
Trouble is 1/2.3" sensor don't yield the performance that even consumers are happy with.
People see images everywhere and want to try and capture that sort of work. They generally find that cameras based around 1/2.3" don't return the quality seen in those images. So they might seek out a better camera especially if they going on a special trip.
Then you find that the old stay dSLR is a bit bulk when you want to shoot with reasonable zoom or focal length.
1" sensor appear to be above the quality level of APS sensors were 10 years ago.
That is why I do think the Nikon 1 system has a place for consumers.
That's a sweeping statement. Have you any evidence to back it up? I use a camera that has a 1/2.3 sensor and I'm very pleased with it. My wife uses a different make of camera that has a 1/2.3 sensor and she's very pleased with it. My sister uses yet another make of camera with a 1/2.3 sensor and she's very happy with the pictures she gets. So who are these people who aren't happy?
Agree. I had a Canon G10, the model that was widely criticised for having "too many pixels" and being noisy as a consequence but it performed very well in reasonable light. Main reason for replacing it was the handling not the performance.
I think if you enlarged a image from even a modern 1/2.3" and had a reasonable close look you see the diffraction. The lens struggles to get a clean sharp image into the sensor. The sensor is ONLY 27.9mm square (4.5x6.2mm)
As for the G10 it uses a 1/1.7" sensor which is 6x8mm or 48mm square. LOL
If 1/2.3" camera could do the job then there would be no FF, APS, MFT etc
To produce a none interpolated print 12x8" at 300ppi you need 8.6MP from the camera
But if the 8.6MP is generated from a sensor that is so small it forces the lens to struggle then the image will probably be soft.
1" sensors strike a balance between being small enough so a lens focal length doesn't have to be very high to give the shooter reach. Thus keeping weight, size & cost down for the shooter. The sensor is big enough to give low-light shooting good ISO and for DOF.
So if you are printing at 12x8 300ppi and one image is from a 1"sensor camera with good lens it will hold its own against FF camera with a good lens. Might be tricky to spot which one is which.
It might even work at 12x18" But I am pretty sure you would spot the issues between a 1/2.3" and FF sensor.
So Andrew what do you do with your images from the 1/2.3" cameras? Do you print large or enter into camera club competitions?
Of course photography is a art form so the technical performance can take a back seat if the image creates the right emotion or reaction from the view. But if a loss of detail could spoil a shoot.
Andrew can I ask which 1/2.3" cameras you use then please?
Yes, but with the exception in the other thread of someone wanting a 36x24" from a camera phone, most "consumers" don't go round printing 12x8", they print 6x4" or look at the results on a monitor.
It seems to me you don't understand the interaction between the pixel size and the effect of diffraction. The Airy Diameter for a typical lens used with a 1/2.3 sensor is 2.7 µm at f4 up to 21.3 µm at f16. In general, a lens for this format will be used in the range f4 to f8. The received wisdom is that the Airy Disk can cover between 4 and 9 pixels before diffraction limits resolution. All this is subject to other conditions. In essence this implies that a reasonably designed lens will provide an acceptable image on a 1/2.3 sensor in normal conditions which will be roughly equivalent to the image supplied by any other size of sensor with the same pixel count and an equivalent lens. This is a cropped image from a Sony HX90 with the full frame inset at top left...
Precisely, Pete. And don't forget the redundancies for reference & control purposes. But as is pointed out later on this Thread, each site or pixel, call it what you will, is discrete. Which is exactly the opposite of that posted in bit of text that didn't seem right and rang an alarm bell in my head. Perhaps Prof Newman could step in here and help us out. Oly
Depends, Andrew, I guess on how much noise you are prepared to tolerate. It becomes very subjective from one user / viewer to another. Depending on subject matter probably, I am inclined to tolerate some noise if it's crispish and doesn't colour fringe and appears filmic whereas others I know run screaming from a room at the merest hint of noise in an unimportant part of an image. Cheers, Oly
Don't think sensor making is particularly expensive anymore. Is not silicon old tech anyway?
Paul, who says consumers are unhappy with small sensors in phones, tablets, pads and laptops? Are they not buying them in ever greater numbers to the detriment of camera sales? And where do most 'consumers' see their photos? I bet it's not in A4 prints, let alone A3+ and larger.
I've seen one One in action out there in the real world and another One being carried. IIRC, Nikon have stopped selling it in some markets due to dismal sales. They should either have joined the Four Thirds syndicate or teamed up with their sensor provider to make APS compacts. Probably too late for the first but they could still fall in behind Sony, now the latter seem determined to go full frame with as much as possible. Cheers, Oly
I think it is expensive and the cost ncreases with chip size, My impression is that the manufacturers were really surprised at the barrier to making 24x36mm chips. My perception is that they took quite some time to decide to run with dedicated APS-C lens lines. The remarkable cost of doubling the chip area to get part-way to the old medium format has stuck. They can bang out the small chips for phones etc. OK but the large ones are a real problem.
As I pointed out the plus to a 1" sensor is the lower focal length for reach you can use. APS and MFT require longer lenses
Well what can we say about smartphones, tablets & pad?
They are light, small & compact. So on that logic the Nikon 1 system has the edge. Because its lens catalogue is going to have smaller lenses.
I don't think you can shoot wildlife in action or other action related images on a smartphone or tablet. This becomes clear to those users of such devices.
Others would disagree with you: http://www.discoverwildlife.com/wildlife-nature-photography/how-photograph-wildlife-smartphone
No. That would currently be 1/2.3, as illustrated above.
Narrow definition of wildlife! Can't see the results much threatening the sale of 600mm F4 lenses.
Very true but I was challenging another of those sweeping statements. I wish people wouldn't make up pretend rules for photography - just get out and take the pictures.
Separate names with a comma.