Matching FX9 to FS7
I’m wondering if anyone has had good luck trying to get the FX9 to match the FS7 for multicam work in both Custom and Slog. I have a lot of clients with FS7’s and don’t want to lose that work if I sell my FS7. Ideally I want to be able to create a simple setting or LUT that would convert the FX9 the FS7 both in camera and for post if possible. So far I’ve done a lot of tests and the results have been surprising and more difficult than I expected.
Actually Custom turned out be easier for me to match than Log. Cody Flowers and I spent a bunch of time at Eli Adler’s Clean Slate Productions in San Rafael, CA trying to match charts and photos from old Nat Geos with a vectorscope and we discovered that if we used the same Gammas, then the FS7 in “ F55-Like” Matrix seemed a close basic match to to the FX9 in “Standard” . If the FS7 was basically zeroed out ( except -13 Sat) then the FX9 matched if we raised the Sat Level a bit and moved the phase over just by +4. If we wanted to get fancy we could also make a couple of changes with the Multi Matrix to bring the saturation on a couple of colors in line, but both the visual and vectorscope match was quite good.
The rest of the tests I did at my house with a model and a bamboo screen behind that had a whole bunch of different colored fabrics. It was easier to see mismatches with the fabrics than it was on charts. The custom settings worked well right away but Slog3 was literally a slog.
The first problem was getting the luminance levels of the both cameras in Slog3 to match with similar f stops. If you shot the FS7 at 1000 EI, that would be 1 stop overexposed. But the FX9 at 1000 EI on Low base would be 1/3 stop under, and at High base 1000EI would be 2 stops over. What I ended up doing was running the FX9 at high base 2000 EI but shooting through an N3 filter. Now I had both cameras overexposed 1 stop and shooting at the same f stop. However even here, the gamma curves of Slog3 on the FX9 and the FS7 are not identical, so my first step was always to make minor changes to the shadows, gain and gamma of the FX9 to match. (The custom gammas were much more similar.)
At first I tried loading an s709 (Venice/FX9) LUT into the FS7 camera so both cameras used the same LUT burnt-in. That sounded logical, but they just looked different and while white balancing the FX9 helped, it wasn't good enough. Then I tried the same thing in Post, but while I could easily get a grey or flesh tone to match , trying to get both at once and nail all the colored fabrics didn’t work unless I resorted to secondaries on a number of colors. I didn't want to tell a client they needed to screw with secondaries. Same problems using using LC709A for both cameras. I also didn’t have any luck using the auto match features in Resolve for charts or flesh tones, and didn’t seem to get lucky with Aces either - though I'm not experienced with Aces.
I didn’t particularly like the look of the s709 LUT on the FX9 footage anyway. I’ve heard complaints that its too green though to me it felt too Yellow but I think that’s pretty close to the same thing. Noticing the FS7 was too magenta and the FX9 too green/yellow - I had a brainstorm. I switched the LUTs. FS7 with s709 LUT and FX9 with the LC709A LUT . Just the opposite from what Sony recommends them for. The LC709A moved the FX9 away from the green and the s709 moved the FS7 away from being magenta.
Bingo - not only did I prefer those matchups individually, but I also found that with a few tweaks I could get much better matches between them. Leaving the FS7 w/s709 LUT footage pretty much alone, I went to the FX9/LC709A combo in Resolve and cooled off the color temperature, added some magenta Tint and played with the saturation, but once I did that I had pretty damn good matches on all the colors and the face. No secondaries needed. Even the colors of charts on a vectorscope lined up way closer than before. I'm sure this would also work matching the FS7 to FX9.
Unfortunately in my tests I had to make somewhat different tweaks for different lighting conditions so its not as easy as I hoped it would be, and thus not as easy to sell to a client who might be looking for an additional FS7. However once you know the basic tricks its pretty easy. Maybe with some more tests I could come up with reliable LUTs for a client, but this stuff is very time consuming.
I’m very interested to see if anyone else has had an easier time than me or come up with other solutions. Its hard to do tests and not build in little mistakes that screw up your findings and I am aware that this sounds a little nuts
. I did hear from a friend that he just set up both cameras in Slog3 on an interview and they matched fine using the regular 709 LUT in camera (not LC709A) on presets. I’m a little suspicious, but if the angles were different enough and the shots colors happened to work you might not notice the problems. I'd be happy to hear there's an easier way.
Lenny Levy
I’m wondering if anyone has had good luck trying to get the FX9 to match the FS7 for multicam work in both Custom and Slog. I have a lot of clients with FS7’s and don’t want to lose that work if I sell my FS7. Ideally I want to be able to create a simple setting or LUT that would convert the FX9 the FS7 both in camera and for post if possible. So far I’ve done a lot of tests and the results have been surprising and more difficult than I expected.
Actually Custom turned out be easier for me to match than Log. Cody Flowers and I spent a bunch of time at Eli Adler’s Clean Slate Productions in San Rafael, CA trying to match charts and photos from old Nat Geos with a vectorscope and we discovered that if we used the same Gammas, then the FS7 in “ F55-Like” Matrix seemed a close basic match to to the FX9 in “Standard” . If the FS7 was basically zeroed out ( except -13 Sat) then the FX9 matched if we raised the Sat Level a bit and moved the phase over just by +4. If we wanted to get fancy we could also make a couple of changes with the Multi Matrix to bring the saturation on a couple of colors in line, but both the visual and vectorscope match was quite good.
The rest of the tests I did at my house with a model and a bamboo screen behind that had a whole bunch of different colored fabrics. It was easier to see mismatches with the fabrics than it was on charts. The custom settings worked well right away but Slog3 was literally a slog.
The first problem was getting the luminance levels of the both cameras in Slog3 to match with similar f stops. If you shot the FS7 at 1000 EI, that would be 1 stop overexposed. But the FX9 at 1000 EI on Low base would be 1/3 stop under, and at High base 1000EI would be 2 stops over. What I ended up doing was running the FX9 at high base 2000 EI but shooting through an N3 filter. Now I had both cameras overexposed 1 stop and shooting at the same f stop. However even here, the gamma curves of Slog3 on the FX9 and the FS7 are not identical, so my first step was always to make minor changes to the shadows, gain and gamma of the FX9 to match. (The custom gammas were much more similar.)
At first I tried loading an s709 (Venice/FX9) LUT into the FS7 camera so both cameras used the same LUT burnt-in. That sounded logical, but they just looked different and while white balancing the FX9 helped, it wasn't good enough. Then I tried the same thing in Post, but while I could easily get a grey or flesh tone to match , trying to get both at once and nail all the colored fabrics didn’t work unless I resorted to secondaries on a number of colors. I didn't want to tell a client they needed to screw with secondaries. Same problems using using LC709A for both cameras. I also didn’t have any luck using the auto match features in Resolve for charts or flesh tones, and didn’t seem to get lucky with Aces either - though I'm not experienced with Aces.
I didn’t particularly like the look of the s709 LUT on the FX9 footage anyway. I’ve heard complaints that its too green though to me it felt too Yellow but I think that’s pretty close to the same thing. Noticing the FS7 was too magenta and the FX9 too green/yellow - I had a brainstorm. I switched the LUTs. FS7 with s709 LUT and FX9 with the LC709A LUT . Just the opposite from what Sony recommends them for. The LC709A moved the FX9 away from the green and the s709 moved the FS7 away from being magenta.
Bingo - not only did I prefer those matchups individually, but I also found that with a few tweaks I could get much better matches between them. Leaving the FS7 w/s709 LUT footage pretty much alone, I went to the FX9/LC709A combo in Resolve and cooled off the color temperature, added some magenta Tint and played with the saturation, but once I did that I had pretty damn good matches on all the colors and the face. No secondaries needed. Even the colors of charts on a vectorscope lined up way closer than before. I'm sure this would also work matching the FS7 to FX9.
Unfortunately in my tests I had to make somewhat different tweaks for different lighting conditions so its not as easy as I hoped it would be, and thus not as easy to sell to a client who might be looking for an additional FS7. However once you know the basic tricks its pretty easy. Maybe with some more tests I could come up with reliable LUTs for a client, but this stuff is very time consuming.
I’m very interested to see if anyone else has had an easier time than me or come up with other solutions. Its hard to do tests and not build in little mistakes that screw up your findings and I am aware that this sounds a little nuts
. I did hear from a friend that he just set up both cameras in Slog3 on an interview and they matched fine using the regular 709 LUT in camera (not LC709A) on presets. I’m a little suspicious, but if the angles were different enough and the shots colors happened to work you might not notice the problems. I'd be happy to hear there's an easier way.Lenny Levy
Last edited:
