I posted this question not so long ago on couple of FaceBook
photography forums that which one produces most noise in the image? And options
were,
A-
50
B-
100
C-
125
Majority of the people as expected choses Option “C” which
makes total sense when we are speaking generally. But reality is different. Now
I am not trying to be guru of photography here or trying to ditch any PRO which
most Pro’s think I do. My intention is pure which is to share what is right
when it comes to photography. Nothing more. Nothing less.
ISO 50 is the right answer. How? Let me explain in detail. Note that description about why ISO 50 is the right answer is going to be long so if you like to learn something new, you may continue else you may skip to the next post or check your whatsapp.
As I have mentioned multiple times in past on various post that ISO has nothing to do with the exposure. Neither it is a part of the exposure triangle, already shown in past the real exposure which everybody laughs at me till Tony Northup came up with his video and proven the same thing which I said well before his video. ISO is basically an amplifier (gain) inside the camera whose primary job is to amplify the signal. Simple is that. Nothing more.
Another interesting fact is that EVERY camera sensor which is ever made by any camera company is always comes with ONE ISO value which is ISO 100 in general. Cameras like Nikon D700, D3s are having ISO values starting with 200. Fujifilm camera also have base ISO 200 in many of their cameras. D810/850 is having starting ISO value 64. So whatever ISO value you see on the camera is just its amplification gain like ISO 400, 800 all the way up to 102,400. But in reality, camera sensor is having ONLY ONE ISO value which is either 64, 100 or 200. And the “extended range of the ISO” which is generally LO1, LO2 LO3 (ISO 25, 32, 50) ~ H1, H2 H3 etc (ISO 204,800 and beyond depending on the camera) is basically nothing but SOFTWARE SIMULATION. Physically these values don’t exist. Camera simulate them for you to FIX the exposure.
Now when you are shooting low light, it means that you are not able to capture enough light signal (AKA “luminosity level” in PS language) on the sensor. So, in result, you going to increase/amplify the signal level by increasing the ISO. What happen next is, when you amplify the signal level, it amplifies EVERYTHING. Everything means, signal + noise. Even at ISO 100 you always have noise present in the captured image, though its less in amount but it’s always there. So, when you have less light, sensor of the camera not able to capture enough light ultimately. Now camera has to do something to fill that missing light. This is where ISO kicks in.
ISO 50 is the right answer. How? Let me explain in detail. Note that description about why ISO 50 is the right answer is going to be long so if you like to learn something new, you may continue else you may skip to the next post or check your whatsapp.
As I have mentioned multiple times in past on various post that ISO has nothing to do with the exposure. Neither it is a part of the exposure triangle, already shown in past the real exposure which everybody laughs at me till Tony Northup came up with his video and proven the same thing which I said well before his video. ISO is basically an amplifier (gain) inside the camera whose primary job is to amplify the signal. Simple is that. Nothing more.
Another interesting fact is that EVERY camera sensor which is ever made by any camera company is always comes with ONE ISO value which is ISO 100 in general. Cameras like Nikon D700, D3s are having ISO values starting with 200. Fujifilm camera also have base ISO 200 in many of their cameras. D810/850 is having starting ISO value 64. So whatever ISO value you see on the camera is just its amplification gain like ISO 400, 800 all the way up to 102,400. But in reality, camera sensor is having ONLY ONE ISO value which is either 64, 100 or 200. And the “extended range of the ISO” which is generally LO1, LO2 LO3 (ISO 25, 32, 50) ~ H1, H2 H3 etc (ISO 204,800 and beyond depending on the camera) is basically nothing but SOFTWARE SIMULATION. Physically these values don’t exist. Camera simulate them for you to FIX the exposure.
Now when you are shooting low light, it means that you are not able to capture enough light signal (AKA “luminosity level” in PS language) on the sensor. So, in result, you going to increase/amplify the signal level by increasing the ISO. What happen next is, when you amplify the signal level, it amplifies EVERYTHING. Everything means, signal + noise. Even at ISO 100 you always have noise present in the captured image, though its less in amount but it’s always there. So, when you have less light, sensor of the camera not able to capture enough light ultimately. Now camera has to do something to fill that missing light. This is where ISO kicks in.
When you increase the ISO in the camera ultimately it increases
the signal level in the “CAPTURED” image (before it dumps it into the memory
card) because your original image signal/value (AKA luminosity level) is not
strong enough to be distinguish as an image. So basically, camera TRIES to fill
up the GAPS which is not captured by the sensor (due to low light) using predictable
algorithms which is embedded inside the camera SNR firmware. The more you
increase the ISO value, the more camera SNR tries to FIX the things inside the
image. And the more fixation is going to be done, the more damage in reality is
going to happen on the final image.
As I have made the base for you guys with above explanation, now coming back to the main topic how come ISO 50 is having more noise then 100 or 125. When you switch the ISO to LO1 (AKA simulated ISO) range (for example ISO 32 or 50), camera makes you believes that your camera ISO value is less. In reality these values don’t exist. So, using your camera base ISO value (let’s say ISO 100), camera processor simulated that ISO 100 as an ISO 50 and then it pushes the exposure to the dark side of the histogram which doesn’t exist in reality. What happening is, Camera SNR firmware tries to generate those exposure value which is not present in the scene and camera processor pushes those generated signal values to get register on the camera memory card which ultimately appear as a noise. The more LO ISO value you are shooting with, the more noise you will be adding in your image eventually.
Bottom line is this, simulated ISO range always generates more noise than amplified values. How? Amplified values are basically fixing the missing parts. Simulated values don’t exist at all and they are entirely created from top to bottom this is why extended ISO range produces more noise. Same phenomenon happens when you shot at based ISO with -5EV under expose and then tries to bring back the exposure in post processing. It doesn’t matter exposure recovered inside the camera or in post processing. The theory of “fixing” the exposure is the same. Watch this video @6:15 and see by yourself. And still if you want to go more deep into the physics behind all this, check out this article which will sure make you scratch your head.
Hope it clears the confusion. Happy learning guys.
As I have made the base for you guys with above explanation, now coming back to the main topic how come ISO 50 is having more noise then 100 or 125. When you switch the ISO to LO1 (AKA simulated ISO) range (for example ISO 32 or 50), camera makes you believes that your camera ISO value is less. In reality these values don’t exist. So, using your camera base ISO value (let’s say ISO 100), camera processor simulated that ISO 100 as an ISO 50 and then it pushes the exposure to the dark side of the histogram which doesn’t exist in reality. What happening is, Camera SNR firmware tries to generate those exposure value which is not present in the scene and camera processor pushes those generated signal values to get register on the camera memory card which ultimately appear as a noise. The more LO ISO value you are shooting with, the more noise you will be adding in your image eventually.
Bottom line is this, simulated ISO range always generates more noise than amplified values. How? Amplified values are basically fixing the missing parts. Simulated values don’t exist at all and they are entirely created from top to bottom this is why extended ISO range produces more noise. Same phenomenon happens when you shot at based ISO with -5EV under expose and then tries to bring back the exposure in post processing. It doesn’t matter exposure recovered inside the camera or in post processing. The theory of “fixing” the exposure is the same. Watch this video @6:15 and see by yourself. And still if you want to go more deep into the physics behind all this, check out this article which will sure make you scratch your head.
Hope it clears the confusion. Happy learning guys.
Babar Swaleheen
No comments:
Post a Comment