Other: FX9 vs FX6 - How different is the 'look'...

Well, don't get your hopes up yet. I went ahead and broached the idea of them letting me use the camera in the off-hours for some testing -- and that is a non-starter. Apparently they think that could face some real penalties if they allow a non-government entity to use one of their cameras for a non-government purpose. I have not run into that kind rigid rules before with a government client, but I will not push back on it. It is their call and I'll respect it. The good news is they are just as interested as we are to see how the cameras compare, so we might be able to do something "official" and then they could release the "imagery for public use." The other good news is they have a quite a few e-mount lenses so there won't be any problem putting the same lenses on both cameras. I'll let you know how it plays out later in the month.
 
It was a pretty busy schedule last week, but the last thing we did on Friday was put my FX6 and the client's FX9 side-by-side and shoot some unscientific comparisons with a couple of Sony 90mm f/2.8 lenses. I only had about 30 minutes, so detailed testing wasn't possible. I set both cameras for S-Cinetone and matched every other setting that I could. I will try to go through the footage this week and see if any of it is worthy of sharing. The two biggest things I noticed right away is that the FX9 is more sensitive to light. With matching exposure settings, it was noticeably brighter than the FX6. The other thing I noticed is how much better the face/eye detect auto-focus is on the FX6. The difference is HUGE!! AF on The FX9 was completely unusable and embarrassingly bad with the same settings. But the FX6, was spot-on 99% of the time no matter what we threw at it. I even swapped lenses between the cameras just to make sure we didn't have a bad one, but it made no difference. If AF is important to someone, and I think it should be in 2022, then I would suggest steering clear fo the FX9 for that one reason alone.
 
Left-field, dull "businessy" take on 6-vs-9:

- My FX9 has been free for the past year and a half. Clients are spec'ing it and kindly making the payments for me via kit fee, many times over (really need to check numbers, it might have paid for itself already). This has been especially vital this May/June, when it's been very very spotty to find an FX9 available for rental. Debating picking up a 2nd FX9 body for just this reason if the trend continues (and if I get another contract to secure it, and if I can find one in stock or used)

- Carrying on the point above, were I to assume the risk, I could make the 2nd body available for rental and be assured that its sturdy build would endure the rental kit life.

- The FX6 is largely built in China to... well... not the highest of quality standards. FX6 FB groups are replete with broken mic mounts and headphone jacks and units returned for various internal hardware bugaboos. A rental unit I had last year had a rubber pad on the top handle jettison itself during no rough handling. While it's adequate to use, it definitely feels more fine and delicate, and not at all up to the rigors of potential rando-b-cam-op or rental use.

- I'm still bitter that they didn't put a damned mic jack of any kind on the body :P


All that said, the power-to-weight ratio of the FX6 is just soooo attractive. I'm not tossing my FX9 into the back seat on family outings. The potential of packing 2x FX6's and lenses into a Pelican carry-on for traveling doc/corporate/interview work is immensely appealing... but who the hell can find 2x FX6's these days? :D
 
Doug , Interesting comparison of AF . Thanks for passing on the info. I'm wondering exactly how you tested it though. I assume you used the same settings on both but what was then actual test. I use AF often on the FX9 and have had no problems. Now admittedly over 90% of that in production has been for sit down interviews and often at very open f stops with Sony and/ sigma lenses. Absolutely no problems though with people leaning in and away from the camera. In just testing Its picked up and held cars zooming by etc. I guess the hardest test is someone walking right up to camera. I haven't tested that as much though a friend did have some issues, but I don't know how careful he is. The A7siii I've used in more run & gun situations and its been excellent but maybe that's more like the FX6. Look forward to seeing if you could detect anything about the look. I wonder if the 2 cameras' AF settings should be a little different?
 
Lenny, those are all good questions and concerns that I would be asking if the tables were reversed. I matched every setting I could between the two cameras, including the focus menus. Unfortunately, the FX9 was running firmware V2.1 instead of the newer V3, and there was no time to update it. However, with that said, there is nothing in V3 that affects Face/Eye detect focusing. V3 would give the FX9 subject tracking, but since it was running 2.1, so I didn't waste time testing subject tracking.


The quick Face/Eye tests that we did involved walking towards camera, talking head, turning head, sudden movements, leaning in, etc. The FX6 beat the FX9 100% of the time, but as you suggest, for a normal sit down interview, the differences might not matter in those circumstances. But for a documentary, b-roll, reality show, etc. there's no way I'd trust the FX9. No way. It slipped off the face too many times and went to the background. One thing I really appreciated about the FX6 is that as a face coming towards the camera exited the frame as it passed, focusing stopped completely at the last focal point, thus making it easy to make a cut or dissolve to the next shot. But the FX9 immediately racked focus to the background the instant it couldn't find the face anymore. I'd hate to edit that mess.


I have a SDI field monitor that can do a quad split so we could watch the face/eye tracking boxes on both cameras simultaneously. If I ever have the chance to do some better testing I would want to record that split-screen view because it revealed exactly what the two cameras were seeing in real time. I think if you just look at the raw footage later it won't be nearly as impactful as seeing the camera's viewfinders and focus boxes working in real time "live" while the footage was being recorded. Trust me when I say it was a huge and surprising difference between the cameras.
 
Last edited:
Thanks Doug . Interesting . I imagine the A7siii would be closer to the FX9 on some of those. Though I am, surprised that the FX9 pulled focus away to the background after the subject left the shot . That sounds like a setting to me. Its very easy to screw up matching settings when testing . I've learned that the hard way.
I wonder if the 6K sensor is harder to get rapid AF on .
 
I agree that it is easy to screw up matching the settings when comparing cameras, however, in this case I am confident that did not happen. When the performance of the two camera was so surprisingly different, not only did I swap the lenses, I also went to the AF menu page on both cameras and compared every setting. This was very easy to do with my quad-split monitor because the two menu pages were side-by-side on the same screen. Even a monkey could of checked them. :-)

BTW, these are my personal default settings on the FX6 and the ones we used on both cameras:

AF TRANSITION SPEED ...................7(FAST)
AF SUBJECT **** SENSITIVITY........5(RESPONSIVE)
FOCUS AREA .....................................FLEXIBLE SPOT
FOCUS AREA (AF-S) ..........................FLEXIBLE SPOT
FACE/EYE DETECTION AF ................FACE/EYE PRIORITY AF
PUSH AF MODE .................................AF
TOUCH FUNCTION IN MF ..................TRACKING AF (menu not availalble on FX9)
AF ASSIST ...........................................ON
 
Last edited:
For better or worse, here are my test shots from last week with the FX9 and FX6. I wish I'd had more time, but 30 minutes was all I had. Read the description on Youtube if you want more details about the setup. I uploaded a 4K file, but as of this minute, the 4K version isn't showing.

 
Surprised nobody's commented yet so I'll kick off. First, thanks Doug for bothering to do this test and for presenting it in such a useful way. Hopefully nobody will complain about your methods ...

Many of the shots confirm my views after shooting with both on the same job. The X6 often, and I mean often, looks and is sharper than the 9. I have never bothered to check it as thoroughly as you but it was something i "felt" was so. Now I see it most definitely is. I'm wondering, does the better AF of the 6 more than make up for the larger sensor of the 9? There can be no doubt that with spot on focus the 9's sensor gives a very slightly more detailed image - but only if you're pixel peeping hard with similar shots side by side.

For me, this test was really useful and confirmed my decision to buy the 6 for its better features and rent the 9 when needed. In terms of "look", my experience is the difference takes the briefest of moments in post make them indistinguishable.

Again, thanks for doing and sharing the test.
 
FWIW: based on the settings you chose for the FX9's AF, the behavior you described is exactly what I'd expect.

- Face Priority vs. Face Only... if in Face Only, it won't snap to backgrounds
- AF Subject Sensitivity - the more sensitive, the less "sticky" it behaves on a single subject. Your sensitivity was cranked all the way up (happily shifts subjects). Level 1 is "Locked On". Also helps if you "register" the face in the AF box (clicking the handle joystick when the face is highlighted, orange line appears below "registered" face

Also, it seems the FX9 wasn't on the latest firmware, because in V3.01 the touch AF works spiffily.

I use this in doc situations plenty and don't have many terrible things to say about the AF... not always perfect, but definitely usable when the settings suit the scenario. I keep the AF adjustments on the quick-access 2 and 3 buttons to vary them throughout the day
 
stewhem said:
he X6 often, and I mean often, looks and is sharper than the 9. I have never bothered to check it as thoroughly as you but it was something i "felt" was so. Now I see it most definitely is. I'm wondering, does the better AF of the 6 more than make up for the larger sensor of the 9? There can be no doubt that with spot on focus the 9's sensor gives a very slightly more detailed image - but only if you're pixel peeping hard with similar shots side by side.For me, this test was really useful and confirmed my decision to buy the 6 for its better features and rent the 9 when needed. In terms of "look", my experience is the difference takes the briefest of moments in post make them indistinguishable.
Click to expand...

I agree completely and it pretty much sums up my observations.

Here are some random thoughts about the two cameras.

Frankly, I'm surprised how well the two cameras matched with almost zero effort. I truly expected the image quality of the FX9 to look noticeably sharper, crisper, and cleaner because of the 6K sensor. I don't see any advantage to having a higher-resolution sensor in these tests. That surprised me. Back when Sony loaned me a FX9 in November 2019 for a few days I shot a ton of test footage (series of 10 videos on YouTube) and I liked the camera, but ultimately the conclusion I reached by the time I sent it back is that it wasn't much more than a FS7m2 with a FF sensor. But as time went along I started to wonder if I had missed somthing and it seems I let the FX9 became better in my mind that it really is. This testing was a reality check. Once again I have learned that the FX9 is nothing special and I wouldn't trade my FX6 for an FX9 if someone offered. Not only does the FX6 have far better AF (something that has become very important to me now that it is so good) it also has 120 fps 4K with Picture Cache; cheaper media options; smaller size and weight; lower power consumption; RAW without an XDCA; etc.

I don't see any advantages to the FX9 except that it comes with a viewfinder loupe and a ton more paint menus. However, the lack of a viewfinder on the FX6 was easily solved by adding a Zacuto Gratical, which is better than the viewfinder on the FX9 anyway. Another issue with the FX9 is that now that it has subject-tracking, how are you going to tap on the LCD to tell the camera where to focus if you're shooing with the loupe over the LCD? So, even if I had an FX9 I'd still need to add a Zacuto EVF in order to use touch-tracking AF.

So, as I see it, that really only leaves the greater number of paint menus as the big advantage of the FX9 over the FX6. I gotta say that S-Cinetone really looks like crap. Is anyone impressed by the images in these tests? I sure am not. If this was a real shoot all that footage would have to be seriously graded.

Testing the two cameras with S-Cinetone as the "base" setting does a disservice to the FX9. I have no doubt that I could create a far better WYSIWYG look on the FX9 that would look much better than anything the FX6 is capable of because if it's lack of paint menus. On the FX6 you have to choose between S-Cinetone or S-LOG3 because those are basiclaly the only viable choices. But, on the other hand, the FX9 will allow you to dial in a much better look becuase it has 5x more paint menus and tons of options not found on the FX6.. This an advantage if someone needs WYSIWYG. In my opinion, under no circumstances is the FX6 suitable for a WYSIWYG workflow. Footage from that camera always needs grading to look good. Anyone who thinks the FX6 looks good right out of the camera probably enjoys Wonder Bread and store-brand vanilla ice cream. :-)

The bottom line is that I'm even happier with my FX6 after these tests, but still frustrated I can't paint it to look better.
 
mcbob said:
FWIW: based on the settings you chose for the FX9's AF, the behavior you described is exactly what I'd expect.

- Face Priority vs. Face Only... if in Face Only, it won't snap to backgrounds
- AF Subject Sensitivity - the more sensitive, the less "sticky" it behaves on a single subject. Your sensitivity was cranked all the way up (happily shifts subjects). Level 1 is "Locked On". Also helps if you "register" the face in the AF box (clicking the handle joystick when the face is highlighted, orange line appears below "registered" face

Also, it seems the FX9 wasn't on the latest firmware, because in V3.01 the touch AF works spiffily.

I use this in doc situations plenty and don't have many terrible things to say about the AF... not always perfect, but definitely usable when the settings suit the scenario. I keep the AF adjustments on the quick-access 2 and 3 buttons to vary them throughout the day
Click to expand...

It's great to hear your opinions, but I'm going to push back on some of this.

The point of the tests wasn't to demonstrate how smoothly AF the FX9 could be programmed by choosing different menu choices. The point was to use the same settings on both cameras and see how they compared. The settings I used on both cameras are my normal default settings that I use every day on the FX6. Even if those aren't the settings YOU would choose, how can you explain such a huge difference in performance? You say you were not surprised by the FX9's behavior, but then you must have been very surpised by the FX6's behaviour, right? They both used the SAME settings and one outperformed the other by a huge margin. Are you not surprised there is such a huge difference between them? That is the point of the tests.

You are correct that with Face Only the camera might not snap to the background so easily, but how do you explain the FX9 constantly losing the face (even while it is still in frame) and snapping back and forth to the background. It should hold focus on that face regardless of whether face-priority or face-only has been chosen. The FX9 AF performs terribly and would totally unacceptable on a real shoot. But notice that the FX6 has no trouble staying focused on the face using the exact same settings.

Your comments about the using the joystick to "lock on" the face are irrelevant since there was only one face. Just one face. Registering a face only matters when there are multiple faces. The camera should be smart enough to stay focused on one face as it moves around the frame when there is only one face. The FX6 way out performed the FX9 on shots #2 and #3. Just admit it. The evidence is incontrovertible and there's no need for the jury to retire.

I wish I had recorded the face/eye boxes moving around on the monitor while the video was being shot because it really showed how much better the FX6 tracked. The FX9 lost tracking even more than what you can tell just by watching the video clips.

Yes, it would have been nice to have firmware 3.01 on the FX9, but there is nothing in the latest firmware that would have caused the camera to perform differently in these tests. There is nothing in the latest firmware that improves Face/Eye tracking.

I'm glad to hear that touch-tracking works great on your FX9. Touch-tracking works fantastic on the FX6 too, but I didn't make that part of these tests because the FX9 I had was running V2.1 and doesn't have touch-tracking. So I purposely stayed away from doing any comparisons of that mode. But on paper, both cameras should perform exactly the same for face tracking. That is what I tested, and we see that there is a huge difference.

If you disagree with my tests, I suggest you get your hands on an FX6 and do some tests of your own that show something different.
 
Actually I WAS surprised by the FX6 performing like that with those settings... since I'm used to the FX9 :D. Really fascinating, especially that it never shifted to the background after the subject disappeared, despite being in Face Priority vs. Face Only

Completely understood, it's a quick n' dirty test, and I'm taking no substantive issue with your process nor your main conclusions (in limited use I also think the FX6 has better AF)... but I do disagree with the assertion that the FX9's AF is unusable. Yes, in those situations with those settings, it's not usable... but I've been using it plenty on different settings in different scenarios with un-small clients and (with the appropriate settings) it hasn't freaked out like that or ruined my call-backs. It's not "set it and forget it" like the A7S3 or apparently FX6, and does take situational adjustment... but it's definitely usable.

As to why it was jumping around in the single-dancing-face scenario... again, Face Priority vs. Face Only. On both FX9 and C500.2 (both of which I have way more experience with than FX6), they love to look for and lock on to specular highlights and high-contrast areas in the background. If I'm AF'ing a face with either, I'm definitely setting them to Face Only. I'm not enough of an engineer to know why that is... only possibly guessing slightly older AF systems designs.

Thanks for the side-by-side, and next time I get both cameras in the same room I'm happy to try and bodge together some additional comparison


...but I'd still own the FX9 because it hires out all the damned time
 
mcbob said:
but I do disagree with the assertion that the FX9's AF is unusable. Yes, in those situations with those settings, it's not usable... but I've been using it plenty on different settings in different scenarios with un-small clients and (with the appropriate settings) it hasn't freaked out like that or ruined my call-backs.
Click to expand...

Yeah, I might have been a little over the top in my comments. If someone can't shoot great video with the FX9 I'd tell them to look in the mirror to find the problem. :-)
 
mcbob said:
...but I'd still own the FX9 because it hires out all the damned time
Click to expand...

I really should show more love for the FX9. When Sony loaned me one for three days in 2019 I decided to go shoot out a Veterans Day parade in Orlando with it. Just me and the camera on my shoulder walking around for a couple of hours. It was fun to shoot and I have already earned enough money from stock footage from that parade to more than pay for an FX9 body. I sold three shots from the parade just today. Yeah, I should show more loyalty to the FX9! :-)

https://youtu.be/hdVBareJalk
 
Doug Jensen said:
I agree that it is easy to screw up matching the settings when comparing cameras, however, in this case I am confident that did not happen. When the performance of the two camera was so surprisingly different, not only did I swap the lenses, I also went to the AF menu page on both cameras and compared every setting. This was very easy to do with my quad-split monitor because the two menu pages were side-by-side on the same screen. Even a monkey could of checked them. :-)

BTW, these are my personal default settings on the FX6 and the ones we used on both cameras:

AF TRANSITION SPEED ...................7(FAST)
AF SUBJECT **** SENSITIVITY........5(RESPONSIVE)
FOCUS AREA .....................................FLEXIBLE SPOT
FOCUS AREA (AF-S) ..........................FLEXIBLE SPOT
FACE/EYE DETECTION AF ................FACE/EYE PRIORITY AF
PUSH AF MODE .................................AF
TOUCH FUNCTION IN MF ..................TRACKING AF (menu not availalble on FX9)
AF ASSIST ...........................................ON
Click to expand...

But surely you would change these AF settings for an interview , they are to be blunt , regarding speed and sensitively , totally wrong for an interview ,single face scenario , it would be the 1 and 1 , face only , the fx9 will not jump around to the background . Not really a test if the setting are complexity wrong ?? you could set a lens at f22 and be at 50mm and say it doesn't have any shallow DoF ??
 
Thanks for doing this quick comparison, Doug. I appreciate you taking the time and opportunity to do so.

From a quick first look I'd say that if the clips weren't labeled I would have no way of knowing which was the FX6 and which the FX9 (aside from the focusing problems of the latter). Admittedly, this is on a 14" MBP screen rather than my usual monitor, and the clips have gone through YouTube's compression, but I'm even more skeptical now of the FX9's special "je ne sais quoi" sauce. This is not to say it isn't a great camera with an amazing image—rather, the bar for all Sony cameras (and most cameras, generally) is high enough at this point that the returns in image quality are diminishing or nonexistent until you get to the very top tier (e.e., the Venice).

I am still curious to see how they compare when shooting slog3 and then creating the look in post.
 
Donny 123 said:
But surely you would change these AF settings for an interview , they are to be blunt , regarding speed and sensitively , totally wrong for an interview ,single face scenario , it would be the 1 and 1 , face only , the fx9 will not jump around to the background . Not really a test if the setting are complexity wrong ?? you could set a lens at f22 and be at 50mm and say it doesn't have any shallow DoF ??
Click to expand...

First of all, stop calling me Shirley.

Did you see an interview in my test shots? I agree that different settings might be more appropriate for an interview setup, but I didn't shoot an interview in these test, did I?

I repeat: "The point of the tests wasn't to demonstrate how smoothly AF the FX9 could be programmed by choosing different menu choices. The point was to use the same settings on both cameras and see how they compared." Period.

If I'd had a whole day to kill maybe I could have tested more settings, but I only had about 30 minutes. If you think the testing is flawed, you are welcome to do your own tests with the settings you think are the right settings. BTW, I've shot interviews with my FX6 using these settings. No problem. Once it locks onto the person's face none of the other settings matter at all. An interview is the absolute easiest thing you could ever ask AF track.

To be honest, testing the focusing performance wasn't even on my mind before we started. The original idea was just to compare the picture quality of the two cameras, but once I saw how different the focusing performed it was easy to get side tracked onto that aspect of the cameras. I never expected there to be such a difference between the two cameras.

The other surprising thing is how close the image quality is between the cameras. I expected the FX9 to look superior, even without pixel peeping, but it does not. They almost look identical, but I prefer the slightly sharper FX6.
 
You must log in or register to reply here.
Back
Top