07-22-2012, 04:25 PM
[quote name='Brightcolours' timestamp='1342962438' post='19532']
Lets introduce an imaginary sensitivity rating called Meh.
Suppose the LX5 with its really small photo sensitive diodes has a sensitivity of 50 Meh.
And lets suppose the A57 with is quite a lot bigger photo sensitive elements has a sensitivity of 250 Meh.
The difference is made by the much bigger surface of the "pixels" from the A57,which simply can "catch" more photons during the same exposure time.
Now we want photographers to make photos with each camera. Suppose they have "comparable" lensing, with a similar field of view, and the same f-value (f4 for instance). With the A57 with a sensor of 250 Meh sensitivity, a "correct" exposure would for instance last 1/250th of a second.
When we make the same photo with the LX5, it would then take 1/50th of a second, just because the sensor is so much less sensitive due to its smaller photo sensitive elements.
As you can imagine, this would be a bit troublesome for photographers. With every camera, with the vast variety of sensors available, exposure times would be very hard to get a grasp on... All would need different exposure times with the same familiar lens settings (field of view, f-stop).
That is where "ISO" comes in. It is a standard to equalize all different sensitivities, and make exposure times predictable, over all the different cameras and sensors. Camera makers are free to amplify the signal to whatever factor they need, to reach a certain ISO setting.
For ISO 100, for instance A57 would need a 1.5x amplification factor (just a made up number) to get its 250 Meh sensor to reach ISO 100. The LX5 would then need a 7.5x amplification to reach ISO 100.
We as camera users do not see how much the signal is amplified, we just get easily handled ISO "standard" settings, a nice abstraction layer to keep things simple and usable.
Stop thinking of "ISO" as a sensitivity rating, it is not (with DIGITAL). It is an exposure index rating.
[/quote]
Ok, i see.
So coming back to original question, lets see how this pans out:
Sony A57: shot at F8, 1/250, ISO 1600
Nikon D5100: shot at F8, 1/250, ISO 1600
You're saying that the exposure will be same. Ok.
You're saying that A57 receives less light so if it has to maintain correct exposure, it will have to increase the sensor gain by 1/2 stop. So because it has amplified the sensor signal or whatever you call it, at 1600 ISO, it will have more noise by 1/2 stop, compared to D5100 which has a less amplified signal at that ISO.
Which means, for noise to be equal, we have to shoot at ISO 1200 on A57, but that will underexpose our picture by 1/2 stop.
Wonder that's what Klaus meant when he replied to my question....
BC, am i getting it right now?
By the way, to my eyes at least, the noise difference looks something like 1 stop. I see lots of noise in A57 files. Is that how it's supposed to be in SLT vis a vis DSLR? I have red something like 1/2 stop.
Lets introduce an imaginary sensitivity rating called Meh.
Suppose the LX5 with its really small photo sensitive diodes has a sensitivity of 50 Meh.
And lets suppose the A57 with is quite a lot bigger photo sensitive elements has a sensitivity of 250 Meh.
The difference is made by the much bigger surface of the "pixels" from the A57,which simply can "catch" more photons during the same exposure time.
Now we want photographers to make photos with each camera. Suppose they have "comparable" lensing, with a similar field of view, and the same f-value (f4 for instance). With the A57 with a sensor of 250 Meh sensitivity, a "correct" exposure would for instance last 1/250th of a second.
When we make the same photo with the LX5, it would then take 1/50th of a second, just because the sensor is so much less sensitive due to its smaller photo sensitive elements.
As you can imagine, this would be a bit troublesome for photographers. With every camera, with the vast variety of sensors available, exposure times would be very hard to get a grasp on... All would need different exposure times with the same familiar lens settings (field of view, f-stop).
That is where "ISO" comes in. It is a standard to equalize all different sensitivities, and make exposure times predictable, over all the different cameras and sensors. Camera makers are free to amplify the signal to whatever factor they need, to reach a certain ISO setting.
For ISO 100, for instance A57 would need a 1.5x amplification factor (just a made up number) to get its 250 Meh sensor to reach ISO 100. The LX5 would then need a 7.5x amplification to reach ISO 100.
We as camera users do not see how much the signal is amplified, we just get easily handled ISO "standard" settings, a nice abstraction layer to keep things simple and usable.
Stop thinking of "ISO" as a sensitivity rating, it is not (with DIGITAL). It is an exposure index rating.
[/quote]
Ok, i see.
So coming back to original question, lets see how this pans out:
Sony A57: shot at F8, 1/250, ISO 1600
Nikon D5100: shot at F8, 1/250, ISO 1600
You're saying that the exposure will be same. Ok.
You're saying that A57 receives less light so if it has to maintain correct exposure, it will have to increase the sensor gain by 1/2 stop. So because it has amplified the sensor signal or whatever you call it, at 1600 ISO, it will have more noise by 1/2 stop, compared to D5100 which has a less amplified signal at that ISO.
Which means, for noise to be equal, we have to shoot at ISO 1200 on A57, but that will underexpose our picture by 1/2 stop.
Wonder that's what Klaus meant when he replied to my question....
BC, am i getting it right now?
By the way, to my eyes at least, the noise difference looks something like 1 stop. I see lots of noise in A57 files. Is that how it's supposed to be in SLT vis a vis DSLR? I have red something like 1/2 stop.