Sorry I reacted that way imag.
And thank you Rick. I guess that just begs the question that I was trying to get at earlier, at what point (and why) are we overexposing the sensor? It's not only the amount of photons falling on it, it's also how we're driving it.
This is how I understand it, correct me where I'm wrong. A CCD or CMOS sensor is like having only one speed of film (native) If you vary from that by going to a higher or lower ISO its sensitivity actually does not change. Rather it is like pushing or pulling your film when it is processed. That is why you get "noise" like grain in film.
But I would argue that the iso rating of a sensor (or even of a filmstock) is not really as fixed as it appears (in fact there were several filmstocks made that could be processed at differing iso ratings). And while one can not do variable processing on a frame in a roll of film one can do variable (analog) processing on individual readouts from a sensor. One can even feed that sensor a slightly different voltage to change its sensitivity. Right?
If I change the ISO settings on my AF100 or my GH1 or even my old D40, obviously the gain of the system is changed, amplifying the signal from the sensor. What's not clear to me is where the amplification takes place. Complicating matters is that I think the lowest ISO setting is certainly not always the native ISO. To get to the lower ISO it seems to be applying negative amplification (negative gain?). And yet the lowest setting doesn't clip where the native ISO does, it's able to handle more photons. Nor is there this insane loss of dynamic range that everyone is warning me about. So what's happening?
As I understand it, changing the ISO setting can be responded to in two ways (or a combination). The signal from the sensor can be amplified (more or less) prior to AD conversion and / or the digital version can be adjusted numerically after AD conversion. There are noise suppression approaches for both. But of course if the signal coming from the sensor is clipped, there's clearly nothing we can do. I believe I stated as much already:
I believe most DSLRs have an amplifier before the ADC. I don't know if this camera is the same. But if it is, is the amplification amount "theoretically" adjustable without changing any hardware? Is it a firmware script that sends the voltage to the sensor? I have a hunch that it is, and that this is how we get lower ISO settings on cameras that have native settings of say 400. If so, then we have a non-destructive way to introduce negative gain without negatively affecting the image. Capisce?
I think the word "gain" is throwing everybody. How about Electronic ND? Or Sensor ND? I should probably copyright both terms, because the more I think about it, the more I can't believe it hasn't been done already.
I'm 100% open to being dead wrong about HOW to implement it: I'm not really that smart about engineering. But I do think that Electronic ND is small potatoes compared to the engineering hurdles that this camera has overcome thus far in coming to market.
FWIW I'm not dissing the camera, I'm still probably gonna buy one, even though I already have an AF100. I also promise you I will and do use Real ND filters whenever possible, as well as rock the vari-NDs when appropriate. Scout's Honor. Also, I don't think I'm smarter than the engineers who designed the camera, but I'm not sure that anyone has asked if they could do this.
I'm simply trying to respond to the OP's request to Sony in light of the reality that it can't happen the way he wishes. I'm trying to posit another alternative that could accomplish what he's asking for and could be implemented at this stage of the production cycle.
And thank you Rick. I guess that just begs the question that I was trying to get at earlier, at what point (and why) are we overexposing the sensor? It's not only the amount of photons falling on it, it's also how we're driving it.
This is how I understand it, correct me where I'm wrong. A CCD or CMOS sensor is like having only one speed of film (native) If you vary from that by going to a higher or lower ISO its sensitivity actually does not change. Rather it is like pushing or pulling your film when it is processed. That is why you get "noise" like grain in film.
But I would argue that the iso rating of a sensor (or even of a filmstock) is not really as fixed as it appears (in fact there were several filmstocks made that could be processed at differing iso ratings). And while one can not do variable processing on a frame in a roll of film one can do variable (analog) processing on individual readouts from a sensor. One can even feed that sensor a slightly different voltage to change its sensitivity. Right?
If I change the ISO settings on my AF100 or my GH1 or even my old D40, obviously the gain of the system is changed, amplifying the signal from the sensor. What's not clear to me is where the amplification takes place. Complicating matters is that I think the lowest ISO setting is certainly not always the native ISO. To get to the lower ISO it seems to be applying negative amplification (negative gain?). And yet the lowest setting doesn't clip where the native ISO does, it's able to handle more photons. Nor is there this insane loss of dynamic range that everyone is warning me about. So what's happening?
As I understand it, changing the ISO setting can be responded to in two ways (or a combination). The signal from the sensor can be amplified (more or less) prior to AD conversion and / or the digital version can be adjusted numerically after AD conversion. There are noise suppression approaches for both. But of course if the signal coming from the sensor is clipped, there's clearly nothing we can do. I believe I stated as much already:
I believe most DSLRs have an amplifier before the ADC. I don't know if this camera is the same. But if it is, is the amplification amount "theoretically" adjustable without changing any hardware? Is it a firmware script that sends the voltage to the sensor? I have a hunch that it is, and that this is how we get lower ISO settings on cameras that have native settings of say 400. If so, then we have a non-destructive way to introduce negative gain without negatively affecting the image. Capisce?
I think the word "gain" is throwing everybody. How about Electronic ND? Or Sensor ND? I should probably copyright both terms, because the more I think about it, the more I can't believe it hasn't been done already.
I'm 100% open to being dead wrong about HOW to implement it: I'm not really that smart about engineering. But I do think that Electronic ND is small potatoes compared to the engineering hurdles that this camera has overcome thus far in coming to market.
FWIW I'm not dissing the camera, I'm still probably gonna buy one, even though I already have an AF100. I also promise you I will and do use Real ND filters whenever possible, as well as rock the vari-NDs when appropriate. Scout's Honor. Also, I don't think I'm smarter than the engineers who designed the camera, but I'm not sure that anyone has asked if they could do this.
I'm simply trying to respond to the OP's request to Sony in light of the reality that it can't happen the way he wishes. I'm trying to posit another alternative that could accomplish what he's asking for and could be implemented at this stage of the production cycle.
