I would agree that getting your exposure correct is very important to the final image quality. But…there are times where getting your exposure right requires having a high ISO for a shot and then you’ll still get more noise especially in the shadows. That’s where fewer MPs with larger light buckets (pixels) that can hold more light and achieve better signal-to-noise can have an advantage. With that said though, it’s amazing how far newer high MP sensors have come, along with other supporting hardware and software, to make the noise level difference pretty small.
@@glennn.3464 agreed, even though you are raising the ISO as long as your exposure is correct you should be fine. The problem is lotta people think just because they have oodles of dynamic range that they can underexpose overexpose and just fix it in post with tons of noise.
The ISO setting mainly adjusts the camera's amplifier rather than the sensor itself, akin to how an amplifier is needed for the stylus of a magnetic cartridge playing an LP record. An aspect not addressed is that signal-to-noise ratios are significantly higher in high-density sensors, such as Canon's 45MP. Since a 45MP sensor is about 50% smaller than a 24MP sensor, it captures a smaller area of light than the larger 24MP sensor, leading to greater noise in the 45MP sensor. In the future, camera sensors might be developed where each photocell can autonomously adjust its output gain, enhancing the dynamic range to more closely resemble human vision. I shoot the R5 and R5ii in RAW for action sports with average settings of 5000 ISO, f/2.8, and shutter at 1/1000 sec., depending on the venue.
No, sensors don't all have the same sensitivity to light. Only a fraction of the incoming light is converted into signal useful to create an image. This is called Quantum Efficiency (QE), it's the proportion of incoming photons that are actually converted into electrons. A typical full frame sensor has a QE between 50% and 60%. The full frame camera with the best QE today is the Canon R3, at 68%. There are several micro four third cameras with a QE above 81%. A 24 megapixels MFT camera has really small pixels - 4 times smaller than the pixels of a 24 megapixels FF camera, and yet they have much better QE. For the vast majority of situations, signal to noise ratio (SNR) depends mostly on shot noise, and in these situations better QE implies better SNR. Just within full frame cameras in the same brand, between the canon R3 and the canon R6, the R6 has larger pixels, but much worse QE - therefore worse SNR. In very low light situations - when you need ISO 25600 and above - read noise become significant, compared to shot noise in shadows. But there, again, smaller pixels often have lower read noise than larger pixels. So, regardless of the amount of light in the scene, pixel size isn't a very good predictor of SNR. As a relevant anecdote, the Sony A7s III is hailed by many photographers as the best low light camera of all times. Surprisingly, it has a 48 megapixels sensor, small pixels binned by 4. It also a very low QE of 51% (that might be close to the worst QE of all cameras released in the past 5 years).
I would agree that getting your exposure correct is very important to the final image quality. But…there are times where getting your exposure right requires having a high ISO for a shot and then you’ll still get more noise especially in the shadows. That’s where fewer MPs with larger light buckets (pixels) that can hold more light and achieve better signal-to-noise can have an advantage. With that said though, it’s amazing how far newer high MP sensors have come, along with other supporting hardware and software, to make the noise level difference pretty small.
@@glennn.3464 agreed, even though you are raising the ISO as long as your exposure is correct you should be fine. The problem is lotta people think just because they have oodles of dynamic range that they can underexpose overexpose and just fix it in post with tons of noise.
The ISO setting mainly adjusts the camera's amplifier rather than the sensor itself, akin to how an amplifier is needed for the stylus of a magnetic cartridge playing an LP record. An aspect not addressed is that signal-to-noise ratios are significantly higher in high-density sensors, such as Canon's 45MP. Since a 45MP sensor is about 50% smaller than a 24MP sensor, it captures a smaller area of light than the larger 24MP sensor, leading to greater noise in the 45MP sensor. In the future, camera sensors might be developed where each photocell can autonomously adjust its output gain, enhancing the dynamic range to more closely resemble human vision. I shoot the R5 and R5ii in RAW for action sports with average settings of 5000 ISO, f/2.8, and shutter at 1/1000 sec., depending on the venue.
@@JGZphotography you nailed it that’s correct
Words matter
@@jimbo27 true
Thank you another great topic!
@@Jennifer-n1z 👍
Yet again well said.
@@castieldiallo2945 Thank you
Yes I’m a new canon owner and subscriber! From uk Punjabi ! I’m a wedding and event photographer and the question is ? Is 24 mp enough for weddings?
Sure as long as you get the exposure right and don’t crop heavily you should be fine.
Do all sensors have the same sensitivity to light. If so, bigger pixels better low level light. If not then the specific sensor would matter.
@@MichaelAnderson-l3k no because the pixel size and how many makes a difference
No, sensors don't all have the same sensitivity to light. Only a fraction of the incoming light is converted into signal useful to create an image. This is called Quantum Efficiency (QE), it's the proportion of incoming photons that are actually converted into electrons. A typical full frame sensor has a QE between 50% and 60%. The full frame camera with the best QE today is the Canon R3, at 68%. There are several micro four third cameras with a QE above 81%. A 24 megapixels MFT camera has really small pixels - 4 times smaller than the pixels of a 24 megapixels FF camera, and yet they have much better QE. For the vast majority of situations, signal to noise ratio (SNR) depends mostly on shot noise, and in these situations better QE implies better SNR. Just within full frame cameras in the same brand, between the canon R3 and the canon R6, the R6 has larger pixels, but much worse QE - therefore worse SNR.
In very low light situations - when you need ISO 25600 and above - read noise become significant, compared to shot noise in shadows. But there, again, smaller pixels often have lower read noise than larger pixels. So, regardless of the amount of light in the scene, pixel size isn't a very good predictor of SNR.
As a relevant anecdote, the Sony A7s III is hailed by many photographers as the best low light camera of all times. Surprisingly, it has a 48 megapixels sensor, small pixels binned by 4. It also a very low QE of 51% (that might be close to the worst QE of all cameras released in the past 5 years).
@@comeraczy2483 correct