Hi,
This is my unsophisticated explanation of
why the Wratten filter reduces grain – but I’m probably wrong.
First, it acts as a neutral density filter
by blocking some of the yellow light thus preventing the blue channel from
being otherwise over-exposed. That, in turn, allows the blue channel to contain
image detail. (In my sample you can see how over-exposed the unfiltered blue
channel is)
Second, because it acts as a two stop neutral
density filter effectively for the blue channel only – it requires an
increase in exposure to compensate. That in turn means that the R and G
channels benefit because now they receive more exposure (the Wratten filter does
not have much of a neutral density affect on the R & G channels just the B
channel). More exposure typically reduces grain (is that what is meant by more
signal to noise ratio?)
Best, gordon pritchard
Ray’s original
explanation…….
I repeated Gordon's experiment with a few more controls. I did the
experiment with the following setup:
Canon 5D
ISO 1600
RAW file
Full manual control
Increased 2 stops to compensate for the blue filter
White Balance done in Adobe Camera RAW
The result was that the noise in the Blue channel was less with the blue
filter.
So, what was wrong with my Signal to Noise Analysis?
While signal to noise ratio is very important in many situations it is not the
critical element in this situation. Unfortunately the eye does not look
at both the signal and then compare it to the noise. It just looks at the
absolute value of the noise. The absolute value of the noise does
increase by the same amount as the signal.
I did a further experiment that showed that if you decrease the ISO setting by
two stops (1600 to 400) and do white balance in ACR, you can achieve the
same resulting decrease in noise. Of course you will have to increase
your exposure by two stops just like you did with the filter.
I would also point out that the noise in the red channel does increase slightly
when the blue filter is used. However, this increase is not as great as
the blue channel without the filter. You can see this effect in Gordon's
examples.
The decrease in ISO setting does not cause the red channel noise increase.
Hope this satisfies everyone.
Stefan and Gordon, I was wrong and will admit it.
I believe that I have explained the effect correctly this time.
BTW when I borrowed the last blue filter in stock from my local photo dealer,
he told me he no longer stocked these correction filters since you can do the
same thing in Photoshop...I am not sure if I should ask him to read this thread
: )
Ray