OnLandscape magazine has had a series on neutral density graduated filters, with one part on colour accuracy. Tim Parkin, editor and author, happens to mention that his very old Lee resin filters had changed colour over the years and now had a strong colour cast. I had suspected this myself with my own set of Lee resin filters. A quick test confirms this. Here are some shots I took, the captions tell the story.
I took several shots of my landing, which is directly underneath a roof window. It was a light overcast day so the natural light was slightly blue. In some later shots (all taken within twenty minutes of each other and the light didn’t change) I included a neutral grey card as a colour reference. I imported the raw files into lightroom and used the shot in which no filter was in place but which included the grey card to get a reference colour temperature and tint that would render the grey card neutral in that shot. I then applied this same colour temperature (and tint) to all the other shots. Any colour change between shots should now only be due to the effect of the filters.
I tried shots as follows:
0.9 grad with the boundary half way down the shot
0.9 grad through the clear part of the filter
0.9 grad through the dark part of the filter
0.9 half way down
0.9 through clear part
0.9 through dark part
…and with the grey card in place…
0.9 through dark part
You can see quite easily that the dark part of the filter gives a distinct colour cast to the shot. Shooting only through the clear part has little effect on the colour. With the filter half way down you can see the change of colour. Shooting entirely through the dark part gives a cast to the whole picture. To be precise, with the grey card in place and shooting entirely through the dark part, the RGB readout from Lightroom changes from approximately 40:40:40 to 45:41:40 .
I’ve had this set of three filters (0.3, 0.6 and 0.9 hard graduated) for well over ten years. I’ve used the 0.9 almost exclusively and the others have barely been out of the box. So I repeated the tests with the 0.6 grad and this showed very little colour cast.
0.6 through dark part
So I’m left to conclude that over the years there has been some degradation in the 0.9 filter due in part to age but also in part to exposure to either atmosphere or light, which the 0.6 filter hasn’t experienced. Either way it’s something to watch out for if you have ageing Lee filters. Until I can afford a new set I’m going to switch to using the 0.6 from now on.