FS7: Sony FX9 v Canon C500 mkII

No idea what these tests are about but I can say for a fact the RAW on the C500MkII looks noticeably better than the compressed. Will I use it on a paid gig? Rarely.


I'll also add, while I think the FX9 probably has better AF, having used the C500MKII on 25 interviews so far, the Face-detect AF has been fantastic. I haven't shot any zip-wires though.
 
Last edited:
It seems that there are two things.. either
a) Raw doesnt work - aka is not what it is cracked up to be.
or
b) reviewers dont know how to use it.

or c) wide dr profiles have made it easier for shooters to maximize dr from the sensor, meaning log and raw are no longer the only way to get those last two stops.

when Canon released the C300, all the talk was “the right bits in the bucket”. This is just a continuation of that. The fact that Canon has both is fantastic. But Sony has finally joined the party, and given profile which gives incredible dynamic range, without the need for raw or log workflows, but just straight from camera.


Sony previously have hypergammas, but setting them up was a bit of a learning curve (hehe).

And even after setting them up, there was always the fear of having done it wrong, and oh my, will i get cancer or develop spiderman powers if i work with hypergammas? will there be pink super whites when displayed on the television?
 
"Sony previously have hypergammas, but setting them up was a bit of a learning curve "

What issues did you have setting up hypergammas?
 
"Sony previously have hypergammas, but setting them up was a bit of a learning curve "

What issues did you have setting up hypergammas?

The first time i set them up, i had to figure out what they were. Then i had to dial in a few other settings to make it look like what i wanted to get straight from camera.

With a picture profile, it is more striaght forward: standard -click-, neutral -click-, cinetone -click-

But once you get hypergamma and set up and save a profile, it is not difficult. But when you go to a new camera, you have to go through and make sure all the custom settings are the same. Or transfer the scene file from an SD card.

Ymmv.
 
Last edited:
Doesn't sound like a terribly steep learning curve for a video camera to me. In fact I'd take that over mastering an A7's menu structure any day.
 
Doesn't sound like a terribly steep learning curve for a video camera to me. In fact I'd take that over mastering an A7's menu structure any day.

It’s like cars Lenny. If someone mentions how their Tesla doesn’t require motor oil change, i’m not going to go say how easy changing oil is. I’ll just say, wow, that is nice. Cars, Lenny. It’s about cars.


Sorry, the above was a running joke for Paul F to read.

On more serious note, Lenny you are absolutely correct. Cameras paint menus and gamma and all the knees and gamuts and all of that is really fun to dive into. But i do realize that even when i talk to other professional shooters, that they don’t always know what those settings mean. Many do. But many don’t and don’t want to learn. As for me, it can be nice jumping onto an Arri, Red, Sony, or Canon, and actually like the default image settings. There really isn’t some big secret to making an image look good straight from camera. It has puzzled me why, for so many years, manufacturers just lop off the highlights and clip ugly for their default settings. So, while wide dr will never replace personal preferences, it at least takes away the annoyance of traditional standard video profiles, and allows pro and even beginners to benefit from that.
 
Last edited:
It’s like cars Lenny. If someone mentions how their Tesla doesn’t require motor oil change, i’m not going to go say how easy changing oil is. I’ll just say, wow, that is nice. Cars, Lenny. It’s about cars.


Sorry, the above was a running joke for Paul F to read.

We cant keep changing oil on cars tho, since oil is a finite resource and we gotta learn how to do without. An average car in Norway uses 100 litres of oil just in oil change during its lifetime. We will run out so might as well change tech now rather than later.

Been driving electrics for 9 years, its a superior technology for sure once the range issues have been sorted. Norwegian arctic winter? Frickin awesome with electric momentum and absolute control of the tires with anti-spin on the electric engine. Our EV is surpassing fossile fuel cars in hills easily. Theyre spinning due to variable momentum and gearing. Our anti-spin just bosses it up the hills with stud-free tires.

Now lets talk cameras ;)
 
I do understand your point James. Particularly if you are handed a camera you don't know then hypegammas could look like a pretty oblique and arbitrary system . In fact they are an oddly screwy system. From my perspective as a a guy who has owned Sony cameras since the EX-1 introduced Cine Gammas ( very similar) and who's a bit of a geek though, the challenge of learning my camera's color system has been a point of pride. So i've been using my own matrix settings since then which always have gotten around the notorious "Sony look" and I actually didn't like Canon color because i didn't know how to manipulate it. I've also always read everything Alister Chapman has written about each new camera and I certainly was mystified by hyper gammas until I read that .

Actually I confess to being a little annoyed that s-cinetone works so well right out fo the box . Won't us geeks have bragging rights anymore?
I can console myself though as I've been working out how to match FS7's and FX9's and that does take a little bit of brain work.

Lenny Levy
 
I do understand your point James. Particularly if you are handed a camera you don't know then hypegammas could look like a pretty oblique and arbitrary system . In fact they are an oddly screwy system. From my perspective as a a guy who has owned Sony cameras since the EX-1 introduced Cine Gammas ( very similar) and who's a bit of a geek though, the challenge of learning my camera's color system has been a point of pride. So i've been using my own matrix settings since then which always have gotten around the notorious "Sony look" and I actually didn't like Canon color because i didn't know how to manipulate it. I've also always read everything Alister Chapman has written about each new camera and I certainly was mystified by hyper gammas until I read that .

Actually I confess to being a little annoyed that s-cinetone works so well right out fo the box . Won't us geeks have bragging rights anymore?
I can console myself though as I've been working out how to match FS7's and FX9's and that does take a little bit of brain work.

Lenny Levy

Haha! Totally agree. Sometimes, i miss the days when it was expected that a camera looked bad straight out of box.

Your ability to match the Fs7 amd Fx9 will be so valuable. Would be interesting to see the two cameras with matching profiles.
 
Last edited:
Actually I confess to being a little annoyed that s-cinetone works so well right out fo the box . Won't us geeks have bragging rights anymore?

The "venice" look is easy to achieve on any FS7 with HG8 and few other paint menu tweaks. No big deal. Even the FX9 can be made to look better than the "out of the box" look with just a few more teaks. Anyone who uses any camera with out of the box settings is either ignorant, lazy, or just doesn't care about image quality.
 
Everyone (?) agrees that the best camera line in the world, the ARRI Alexa/Amira produces the best images with the least work, straight out of camera. For example, select ProRes and rec709 (there's a few flavors, including low contrast (AKA wide DR)), expose correctly and the result will be accurate/gorgeous including holding up well in challenging lighting conditions (mixed etc.). In 2020, we've got rec2020 (expanded colorspace) and rec2100 for HDR displays (rec2100 adds specific gamma definitions, e.g. HLG and PQ). Being able to shoot properly lit SDR and HDR and get awesome color straight out of camera (SOOC) is a worthy goal. I've been saying this since buying a C300 II years ago. In the past I found Sony's hyper complex menus and settings as a bandaid to sub-par color science, especially compared to Canon. However, I found the C300 II wasn't able to match what I could get out of a custom color & gamma tuned 1DX II. The 1DX II was tuned to match what I observed from ARRI: correct and pleasing skin tones in as many conditions as possible from testing and tuning custom color and gamma profiles.

This is where I realized that raw and log are a bit of a scam: with a good sensor and at least 10-bits, a wide DR closer-to-709 gamma was sufficient to achieve max sensor DR without having to deal with wide color spaces and log gammas. Even with only 8-bits and a lower DR sensor, I was getting (and still get- have the 1DX II & EOS R and C300 II was sold) great SOOC color with very little need to fix things in post.

Many of us have seen <insert camera here> Alexa LUTs or in-camera settings for the ARRI or Canon look, however while these methods can sometimes match Alexa/Canon in specific settings, they don't generalize and work for all cases. The same is true for matching the S-Cinetone/Venice look. The color science is non-linear, and a simple fixed transform, even a 3D LUT isn't going to work in the general case.

I began asking the question: why don't camera manufacturers make the default factory setting look amazing, and work for the general case with fantastic results, as ARRI has been doing? Without going into 'conspiracy theories' related to the CamCartel, perhaps they just didn't have the tech/skill to create generalized and correct color science that can hang with ARRI? A big part of this is exposure invariant color. Meaning no color shifts into the shadow or highlights (typically magenta/green are the main problem however Canon has issues with yellow/blue too (note the complementary colors)). This is why the exposure ramp tests on skintones are so useful: makes clear what the color science is doing, and shows why the better cameras look better than the lesser cameras even when exposed perfectly. For example highlights and shadows have correct colors without subtle or not so subtle color shifts, which ultimately affects the overall image and perceived quality.

Below the ARRI level, Canon's WideDR with 709 got close, and CLog isn't all that flat, so with minor gamma curve adjustments can work OK. However the color of Canon's CLog is not as good as other possible settings on e.g. the EOS R. It's possible to nearly match CLog DR with custom curves via a custom picture style. This then allows for using much better color settings as well as allowing the color problems to be fixed somewhat for skin tones (using Canon's Picture Style editor). Why doesn't Canon ship this kind of color as the default? Everyone who's commented on these improvements has reacted positively to the real world results. While it took a bit of time to figure out, it wasn't that hard in the grand scheme of things. Are they holding back on purpose, waiting to deliver better color for a future camera or ?

I've watched Sony's color improve since I owned an FS700, A7S I, and A7S II. The A7 III is pretty good, and the new S-Cinetone's performance on the FX9, especially compared against raw + log (at +/- 6!!) makes it clear that we don't need raw or log for what most of us shoot. From these tests we can see the Venice is a little better, and the Alexa still holds a comfortable lead (the gap has closed quite a bit).

At the end of the day, most creators want the tools to work without thinking about the tools as much as the creative product. For my live shoots with the EOS R, I use my custom color profile along with AutoWB and AutoISO (indoors/night). Outdoors min ISO + vND and in most cases AutoWB works well. For narrative, I lock things down for more control, as extra post is a given for story/emotion. Phones will get AI Color before dedicated cameras due to their massive advantages in computing power (now including AI hardware), meaning really good skintones no matter what the conditions. See also Apple's Deep Fusion for both detail enhancement as well as color using AI: https://venturebeat.com/2019/10/03/...-on-ai-sharpens-photos-like-hdr-fixes-colors/

Who wouldn't want AI Color in a dedicated camera and/or great color SOOC with the default settings? Perhaps some enjoy tinkering, which is fine and can be fun, however many of us just want the camera to be accurate as the eye sees so we can focus on the content/product with the least hassles in dealing with menus/settings or having to fix things in post.
 
Anyone who makes any camera that needs tweaking out of the box is either ignorant, lazy, or just doesnt care about image quality.

Wrong. Do we all agree on what looks good? Of course not. I've seen your videos and there's no way that even the two of us would agree on what looks best, let alone thousands of other camera owners around the globe. Please understand that is is not meant to be an insult, just a statement of fact that we all have our own tastes. it is much better to begin with a neutral image and then paint to taste. I wouldn't want it any other way. I don't want an anonymous engineer in Japan deciding for me what looks good because I guarantee I won't like it. Anyone who can't be bothered to spend an hour tweaking their camera once when they first get it is ignorant, lazy, or doesn't care about image quality. Tweak once -- and then enjoy images YOU like for the life of the camera.
 
Last edited:
This is where I realized that raw and log are a bit of a scam: with a good sensor and at least 10-bits, a wide DR closer-to-709 gamma was sufficient to achieve max sensor DR without having to deal with wide color spaces and log gammas. Even with only 8-bits and a lower DR sensor, I was getting (edit) great SOOC color with very little need to fix things in post.

I absolutely agree with this. I was having discussion with a friend today about this very thing. As the cameras get better and better I see less and less need for LOG or RAW. I'll accept having to shoot RAW for high speed slow-mo on my F55 because that's the only way I can do 4K at 120 fps. But I'll rarely, if ever, be shooting V-LOG on my new Panasonic S1H. I can get a fantastic look on-board the camera and then teak it with some mild grading in post and have a great looking image that couldn't be improved on shooting LOG. Remember a few years ago when 35mm adapters from Letus and other companies were all the rage -- until we got DSLRs that could shoot video and then S35 sensors on video cameras -- then those adapters went the way of the dodo bird. Well, I see this whole LOG thing as just another stop-gap measure a lot of us will look back on and say, "crap, remember when we had to shoot LOG and then run everything through Resolve to get the best image? Now we just do 90% of the grading in-camera and then do some minor touch-up in post". Those days are coming, mark my words. LOG is going to look like a silly old-fashioned thing if your aren't working at the highest levels of Hollywood-style cinema.
The emperor has no clothes.
 
Hey Doug- you've got a lot of great experience and knowledge to share; sometimes we agree, sometimes not, and I appreciate your feedback. We can perhaps best help those less experienced by providing accurate/useful info while not insulting anyone (or at least make a joke out of it so won't be offensive (tricky though as jokes can backfire, especially with just the written word)).

Totally agree the SOOC default color is a high priority moving forward and more manufacturers will have to step up their game to compete. Kudos to Sony for the FX9 and the idea behind the Venice and S-Cinetone re: default color. While S-Cinetone has a bit of style to it, it's not that far off neutral. FX9/Venice SLog3 look's fairly neutral, though I'm not sure which LUT each example I've watched used. Looks like Sony has some issues with a default LUT to not have e.g. green/magenta problems (Canon has similar issues with default LUTs for their cameras). Seems strange that these manufacturers can't get their LUTs right!? Once again, ARRI has great SOOC color, and their LUTs work perfectly (even works (OK) on C300 II CLog 2 with wide gamut color + Production matrix (probably C500 II as well)).
 
Wrong. Do we all agree on what looks good? Of course not. I've seen your videos and there's no way that even the two of us would agree on what looks best, let alone thousands of other camera owners around the globe. Please understand that is is not meant to be an insult, just a statement of fact that we all have our own tastes. it is much better to begin with a neutral image and then paint to taste. I wouldn't want it any other way. I don't want an anonymous engineer in Japan deciding for me what looks good because I guarantee I won't like it. Anyone who can't be bothered to spend an hour tweaking their camera once when they first get it is ignorant, lazy, or doesn't care about image quality. Tweak once -- and then enjoy images YOU like for the life of the camera.

Oh most of my work is all on the wonk :)

But any canikon raw stills camera looks good to me - I would expect cameras to adhere to some standards - grey being a multiple of 1-1-1 skin on the skinline etc. Leaving the user to mess that up to taste.

As it happens this is my most representative stuff. Nikon and phantom seeming to deliver the simples and most normal colour.
https://vimeo.com/323633787
 
Last edited:
Oh most of my work is all on the wonk :)

But any canikon raw stills camera looks good to me - I would expect cameras to adhere to some standards - grey being a multiple of 1-1-1 skin on the skinline etc. Leaving the user to mess that up to taste.

As it happens this is my most representative stuff. Nikon and phantom seeming to deliver the simples and most normal colour.
https://vimeo.com/323633787

Looks pretty good to me!
 
With the simplicity and integration of 3D LUTs into basically every piece of post software out there, it's beyond me why so many people still dance between SOOC and LOG workflows. It just seems like an unnecessary layer of complication to me.

Learn a single Log profile, shoot that profile, drop in your Rec709 LUT and grade to taste. It's so little effort, and you maintain both consistency in how you handle/exposure the camera, and consistency in having the full latitude the camera can offer.
 
With the simplicity and integration of 3D LUTs into basically every piece of post software out there, it's beyond me why so many people still dance between SOOC and LOG workflows. It just seems like an unnecessary layer of complication to me.

Learn a single Log profile, shoot that profile, drop in your Rec709 LUT and grade to taste. It's so little effort, and you maintain both consistency in how you handle/exposure the camera, and consistency in having the full latitude the camera can offer.

Because there are situations when you don’t have time for that or where you cannot be sure that the guy who grabs your card knows what to do. Yes, that still exists.
 
Back
Top