jcolwell wrote:
I guess this makes good sense, as the reason you're likely to use it is to reduce light intensity, to make up for shutter speeds lower that you would use with a "more capable" camera.
The ND filter is an actual physical filter and not some sort of digital facsimile on the XZ1 isn't it ? I know on the Canon G series it is as you can see it slide over the lens element.
millsart wrote:
The ND filter is an actual physical filter and not some sort of digital facsimile on the XZ1 isn't it ? I know on the Canon G series it is as you can see it slide over the lens element.
If it wasn't a physical filter there would be no point in calling it an ND filter, so I don't understand the question about whether it's "applied to the RAW files". A "digital ND filter" would imply simply increasing the shutter speed or stopping down further, or lowering ISO sensitivity, which doesn't achieve what an ND is used for.
millsart wrote:
The ND filter is an actual physical filter and not some sort of digital facsimile on the XZ1 isn't it ? I know on the Canon G series it is as you can see it slide over the lens element.
True. Same as the X100. OTOH, there's nothing like some practical examples.
AhamB wrote:
If it wasn't a physical filter there would be no point in calling it an ND filter, so I don't understand the question about whether it's "applied to the RAW files". A "digital ND filter" would imply simply increasing the shutter speed or stopping down further, or lowering ISO sensitivity, which doesn't achieve what an ND is used for.
I think a "digital ND filter" could be accomplished by attenuating the sensor analogue output before it's converted A-to-D. This implies that there's at least slightly more dynamic range in the sensor analogue output than in its "matched" AtoD converter, but this is certainly possible. A pre-sampling method for "lowering ISO" could provide effectively the same results as a physical ND filter, but only if it's acccomplished before the RAW file is created (i.e. before the sensor analogue output is sampled and converted to digital). Of course, this would also require super signal-to-noise performance.
Thats what I naturally assumed, but if it was a physical filter, then it would make no sense not being applied to a RAW file, as the type of file output a user chooses wouldn't affect a physical filter in the optical path and would be a silly question for the OP to ask given his posting history here.
As such, I thought maybe it actually its a digital version and wondered if thats the case. My Leica X1 for example says it had "image stabilization" when there is neither no sensor based nor lens based stabilization effect.
Camera makers are increasingly calling software effects their hardware names. Sometimes the affect can be quite similar (X1 wasn't lol) but it does make things confusing.
Just you watch, some compact is going to have a "polarizer" option in the menu and rather than a physical filter, its just going to be some software thing that saturates blue skies more. Thus people are going to wonder why their polarizer setting isn't doing anything to cut glare off wet rocks etc
AhamB wrote:
If it wasn't a physical filter there would be no point in calling it an ND filter, so I don't understand the question about whether it's "applied to the RAW files". A "digital ND filter" would imply simply increasing the shutter speed or stopping down further, or lowering ISO sensitivity, which doesn't achieve what an ND is used for.
millsart wrote:
...As such, I thought maybe it actually its a digital version and wondered if thats the case. My Leica X1 for example says it had "image stabilization" when there is neither no sensor based nor lens based stabilization effect.
Exactly. My Fujifilm F31fd has "Picture Stabilization". It simply biases the program modes towards high shutter speeds.
jcolwell wrote:
I think a "digital ND filter" could be accomplished by attenuating the sensor analogue output before it's converted A-to-D. This implies that there's at least slightly more dynamic range in the sensor analogue output than in its "matched" AtoD converter, but this is certainly possible. A pre-sampling method for "lowering ISO" could provide effectively the same results as a physical ND filter, but only if it's acccomplished before the RAW file is created (i.e. before the sensor analogue output is sampled and converted to digital). Of course, this would also require super signal-to-noise performance.
The charge storage ("well") is the bottleneck at the lowest ISO for all sensors I know. It is possible to lower the ISO sensitvity by attenuating the signal electrically between the photodiode and the charge storage. It would have the benefit of getting a very low photon shot noise and a very clean image, since all the photons are converted (as opposed to a physical ND filter where the photons are converted to heat in the filter). On the other hand, it would increase the read noise and higher ISO will suffer. If this method is applied, it would however be no reason not to make this available simply as lower ISO.
Since RAW often has more highlight headroom than JPG, it is reasonable to ask if the ND filter is just eating the headroom. This would be feasible up to about 1 stop, but hardly 3 stops. But again, it would be better to make this available simply as lower ISO, like ISO 50 in Canon DSLRs.
