The Onset of Noise, and its relationship to high-density high-megapixel sensors
Mar 16, 2019 17:37:21 #
Chris T wrote:
One reads a great deal on the Net - vis a vis this... (show quote)
It depends of the source of the noise. Stochastic (random) noise goes as the square root of the number of photons. If you double the number of pixels in a sensor and the size and basic sensor performance are unchanged, you will increase noise by 14%.
Other noise, such as amplifier or digitizer noise may change as well. Typically as the number of pixels increases, both these sources of noise will typically increase.It depends of the source of the noise. Stochastic (random) noise goes as the square root of the number of photons. If you double the number of pixels in a sensor and the size and basic sensor performance are unchanged, you will increase 14%.
Mar 16, 2019 17:39:13 #
Haydon wrote:
You have the wrong data. I bought mine in June of 2012 from Canon. Wikipedia is NOT reliable. The majority of it is user input. Misinformation again.
Here I'll give you a reference from Wiki and this one is correct.
https://en.wikipedia.org/wiki/Canon_EOS-1D_Mark_IV
Here I'll give you a reference from Wiki and this one is correct.
https://en.wikipedia.org/wiki/Canon_EOS-1D_Mark_IV
Oh, great, Haydon … you denounced info provided at a Wiki link …
And, then - proceeded to give me more reliable info - ALSO - at a Wiki link!!!!
The same exact chart - from which I compiled the data for my list - is at the bottom of YOUR link!!!
Mar 16, 2019 17:40:23 #
Chris T wrote:
Larger pixels occur in older cameras with less RES, Bebu … like the 12MP D700, and my own 12MP T3.
It seems to me - they BOTH have a better handle on noise than cameras like the D7100/D7200 models!
It seems to me - they BOTH have a better handle on noise than cameras like the D7100/D7200 models!
The larger pixels don't have less noise. They have less signal to noise ratio.
Mar 16, 2019 17:42:52 #
spaceytracey wrote:
When submitting to Stock agencies, the main rejection reason seems to be "noise". Half the time I cannot even see it so, my post-production is mainly "noise reduction" whether I see any or not. Therefore, I'm interested in any solutions available. Thanks for the link. I've saved it so I can study it later.
Sure, Tracey … enjoy …
Oh, er - you might want to take a brush-up course on Physics, before you start …
And, er - also … Trig II … and maybe - your Photo Sensitometry Course, as well …
Anyway … enjoy the read!!!!
Mar 16, 2019 17:52:52 #
Chris T wrote:
...It seems to me, if you advertise a camera that's capable of delivering an ISO of 640,000 (for example) you should be able to dial up that number - and wind up with a product at that given ISO - w/o excess NOISE!!!...
Advertising is advertising.
The D5 says the highest normal ISO is about 200,000 and the highest extended range (HI5) is 3,000,000.
Advertising hype.
Anything above ISO 1000 has noticeable noise, and I use up to 12800 regularly. I can get useful images at 50,000 with enough postprocessing, 200,000 with enough downsampling. 3,000,000 is wishful thinking, for emergency use only, not to be depended on.
Of course my images are used on the web and in print, where noise is not usually a problem due to the size of the final image.
PS: I don't believe ISO settings are any more accurate than 2 significant figures. No point in writing them out more precisely.
Mar 16, 2019 17:57:23 #
DirtFarmer wrote:
Advertising is advertising. br br The D5 says the... (show quote)
By that you mean ISO 80 is the highest accurate one, Dirt?
Mar 16, 2019 17:58:12 #
johntaylor333 wrote:
It depends of the source of the noise. Stochastic (random) noise goes as the square root of the number of photons. If you double the number of pixels in a sensor and the size and basic sensor performance are unchanged, you will increase noise by 14%...
If you double the number of pixels in a given area then the number of electrons will be halved. The signal to noise will rise by the square root of 2, which is about 1.4, or a 40% increase.
Mar 16, 2019 17:59:03 #
Chris T wrote:
By that you mean ISO 80 is the highest accurate one, Dirt?
What do you mean by accurate?
I get useful images (for my needs) up to 12800.
Mar 16, 2019 18:00:10 #
DirtFarmer wrote:
If you double the number of pixels in a given area then the number of electrons will be halved. The signal to noise will rise by the square root of 2, which is about 1.4, or a 40% increase.
Oh, well, then, Dirt … with THAT formula - you've stated what one can expect from a 100MP Sensor - right?
Mar 16, 2019 18:13:12 #
DirtFarmer wrote:
What do you mean by accurate?
I get useful images (for my needs) up to 12800.
I get useful images (for my needs) up to 12800.
You indicated any numbers beyond two digits - were unreliable!!!
Look back, Dirt!
Mar 16, 2019 18:37:03 #
johntaylor333 wrote:
It depends of the source of the noise. Stochastic... (show quote)
Why, thank you, John … then, this is an explainable phenomenon … so good to see that - thanks so much!
Then, there is actually an exacting formula which explains why one gets more noise as one increases Res!
Mar 16, 2019 18:56:26 #
BebuLamar wrote:
The larger pixels don't have less noise. They have less signal to noise ratio.
Bebu … it seems to me … if you have larger pixels - i.e. D700, D3s, D300s (all 12MP) you will have a greater threshold before the need to deal with noise - which becomes more obvious - using cameras with twice the Res. Triple it to 36MP (D810, K-1 II) and now you're going to see twice as much noise, and - occurring sooner. Add another 8MP (to 42MP) - or another 10MP (to 46MP) you will have three times as much noise to manage. Go all the way up to 50MP … and your threshold, before having to deal with noise - becomes even MORE marginal. i.e. Greater Res - requires more efficient noise reduction techniques. And, this means - you will not only GET larger buckets - you will have to EXCHANGE - even MORE buckets!!!
Mar 16, 2019 19:05:38 #
Chris T wrote:
And the 5D4 is a better camera than the R … and the 6D is a better camera than the 6D2, Dave …
And the beat goes on - la-de-da-de-da ….
And the beat goes on - la-de-da-de-da ….
Is this from personal experience using all these cameras?
Think of the whole system.
Mar 16, 2019 19:34:44 #
spaceytracey wrote:
Correct me please if I'm wrong but, I was taught that shooting @ a high ISO was the main culprit for creating noise.
It is. Just check out what your camera is capable of. Nowdays I do not use the extreme high ISO values on my camera bodies.
If you under expose the image that doesn't help as well.
There may be other solutions to the problem, like faster lenses, adding light, slower shutter speeds.
Mar 16, 2019 20:32:35 #
Chris T wrote:
Sure, Tracey … enjoy …
Oh, er - you might want to take a brush-up course on Physics, before you start …
And, er - also … Trig II … and maybe - your Photo Sensitometry Course, as well …
Anyway … enjoy the read!!!!
Oh, er - you might want to take a brush-up course on Physics, before you start …
And, er - also … Trig II … and maybe - your Photo Sensitometry Course, as well …
Anyway … enjoy the read!!!!
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.