Diffusion is something I use a lot. I'm a big fan of hard lights as they are so versatile, but they mostly need to be bounced or diffused. I've found that I usually end up using the same 2-3 types of diffusion on shoots; ones that I've tried and tested. That's mostly down to budget. I'm asked what rolls to buy or a selection turn up and I have to decide which ones to crack open. Either way, it's inevitable that you'll go with familiarity.
Somewhat frustrated with my lack of experience when it came to the art of diffusion, I decided it was time to test them all! So I contacted LEE Filters and they sent my friend Kieron, from Manned Camera, and I this lot!
At this point just want to give a shout out to Shane Hurlbut ASC. I ran what I was going to do past him and he gave some very helpful advice. He's incredibly generous with his knowledge. If you want to learn more about cinematography, from a DoP at the top of his game, then take a look at the Inner Circle!
We went with a very simple setup. Shot on Kieron's Arri Amira with Zeiss Super Speeds. One Arri M8 HMI, through each piece of diffusion. This is a little diagram of our measurements. Light was positioned high and to the side as I'm a big fan of Rembrandt lighting. I wanted that clearly defined nose shadow on the cheek, so we could really see how it was effected.
Why the M8? Kieron owned one! I have had some suggest that we should have done this test with tungsten - and I think there are valid reasons for that, namely that tungsten light is more accurate than an HMI. However, the M8 is a superb HMI, and I love using it! We did use a tungsten Arri T2 on the daylight correction stuff, and it was nice to be able to compare tungsten/daylight on the coloured diffs.
One of the long conversations we had, was whether to add fill. Flagging off the M8, so that it just went through the diff would have been the most scientific way to approach these tests. We tried that and found that the shadows were so dark, it was slightly distracting. Adding a touch of fill just lifted the left side of the face and I think makes it easier to see what the diff is doing to those shadows. (We flagged off the light completely for some of the test with CTB diff, so you can see just how dark the shadows got)
However, we found that adding fill with a 2nd light, and then trying to balance that with the amount each diff took out, was going to really slow us down. So we let some light spill onto the floor, which bounced back onto our model.
Now, this presented me with a bit of a problem. Getting a very accurate light meter reading of how much each piece of diff took out. I sacrificed that accuracy for a couple of reasons. One is that you can find that info on the LEE Filters website. The other is that it's unlikely you'll always be using diff in a highly controlled situation, where you only use one light, and that light is completely flagged.
But the T Stop values you see in the test were all done very consistently, pointing straight at the diffusion from the same distance. Overall, we were surprised that less light was generally taken out than we expected. For example Heavy Quiet Frost and Soft Frost was supposed to take out between 3-4 stops (according to LEE Filters). We measured one! Even factoring in a little spill, that's way off. Got me thinking that the type of light used, distance from the diff/subject etc are all variables that will effect this. So unlikely we're ever going to match what LEE measured. Speaking with David Morphy at Cirro Lite recently when looking at the new Dedolight LedRama, he came up with the theory that we had made a large light source by using the diffusion rolled out. Anyway - long story short - we were more interested in how each diff made our model look, over T Stops.
Still, very interested to hear others experiences with this in the comments!
Incidentally, the Amira was exposed every time using the grey card and false colour, so that exposure remained consistent. This worked very well and our models face remained at the same levels throughout, without the need for any correction in post.
So with that all sorted, we went about the tests. Our model Andreea Paduraru sat for 2 full days turning her head from side to side, as we went through every diffusion LEE Filters had sent us.
I found it really fascinating to do this and instantly realised I was learning something hugely valuable, that was going to transform the way I light. Going on my next few jobs after doing these tests, I found myself using the video with the director to make sure we were both happy with the look.
My hope is that DP's, whether they are just starting out or have years of experience, can use this video to assist them when choosing the right diffusion.