Hi all - it seems there are plenty of us who shoot video with Sony mirrorless cameras - and nowhere to go to get some answers
Thought we might try a thread for this propose (similar to the active one on the Canon forum.)
I know it has taken me a while to get up to speed, choose a video editor (Davinci Resolve for me), figure out what the hell log is and how to expose it, etc. etc.
I picked up an FX3 a couple of months ago and it's been a really fun camera to use. Feels very immediate and straightforward for video, making the experience pretty seamless and enjoyable.
Been working on figuring out how to use CineEI on it (as opposed to the usual, "flexible ISO" method) with mixed results, but I'm slowly getting the hang of it. It's also been entertaining to figure out color correction and grading techniques in Davinci Resolve.
Nothing substantial to show on YouTube but I've posted some clips on Instagram.
One thing about slog3 I've noticed is that the footage tends to look less contrasty in camera with the built in Sony Rec709 LUT than when you transform to Rec709 in Resolve. I manage to compensate for it in Resolve but it would be nice to have more accurate 1-1 preview in camera. Wonder if it's something to do with the Sony LUT or the physical camera display.
For what it's worth, it's not something I've encountered when shooting c-log2 on my R5II. Their display LUT and screen are pretty accurate and things match in Resolve right away. Even then, the video shooting experience on the FX3 is a lot more enjoyable than with the R5II, which is now relegated to mostly photo duties.
n8rv wrote:
One thing about slog3 I've noticed is that the footage tends to look less contrasty in camera with the built in Sony Rec709 LUT than when you transform to Rec709 in Resolve. I manage to compensate for it in Resolve but it would be nice to have more accurate 1-1 preview in camera. Wonder if it's something to do with the Sony LUT or the physical camera display.
For what it's worth, it's not something I've encountered when shooting c-log2 on my R5II. Their display LUT and screen are pretty accurate and things match in Resolve right away. Even then, the video shooting experience on the FX3 is a lot more enjoyable than with the R5II, which is now relegated to mostly photo duties. ...Show more →
I have found for S-LOG3 footage on my Alpha 1 that overexposing by +1.5 or +1.7 and then using the DaVinci Resolve SLog3SGamut3.CinetoLC-709 LUT yields far superior results than using i.e., Sony SLOG3 to Rec709.
And yes - learning about shooting and editing video has been great fun for me after shooting only stills for decades ;-)
You want to go from S-LOG3 to Davinci Wide Gamut and then Davinci Wide Gamut out to Rec 709 Gamma 2.2 or Gamma 2.4 depending on your deliverable or the Gamma to match your calibrated monitor and preferably doing this via node CSTs as opposed to setting up color management in Davinci’s project settings. If you are starting out FIRST AND FOREMOST learn a proper color management workflow as opposed to using a LUT for this purpose. There are some good LUTS out there like Cullen Kelly’s or the Phantom LUTS but starting out I would not use them.
Take a deep dive into Davinci’s color page and learn a proper color managed workflow from Cullen Kelly, Darren Mostyn, Casey Farris or Quazi’s YouTube sites and then only later decide whether or not you want to use a LUT as a starting point for look development. Nothing wrong with using LUTS (if they are good ones) but look at them as ONLY one piece in a pretty big puzzle to getting the most out of S-LOG 3 or any other log footage. What you do in between your incoming and outgoing CST nodes and NOT the LUT is what matters the most in terms of getting the most out of and grading log footage. I use LUTS, nothing wrong with them but they will really hinder your color grading skills if you rely on them to much and particularly if you rely on them when you are starting out.
I don't really use LUTs in Resolve anymore. At first I was, but quickly moved on to CST nodes, then color management. Over the past number of months I've been using that method, as outlined by some of the people mentioned above and so far it's worked well for me.
The only discrepancy I'm having is specifically with the FX3's monitor. Either the built-in default LUT is off or the screen itself is just not very accurate. Following the same method with C-log2 files from the R5II gives me pretty accurate match to what I see on that camera's screen.
Interesting note about using CST nodes as the best option. Can you elaborate on why that's better than color managed? Or is there a video that goes into that? I can understand how fully relying on nodes gives you a little more flexibility, but otherwise, I've always assumed you can get the same results.
n8rv wrote:
I don't really use LUTs in Resolve anymore. At first I was, but quickly moved on to CST nodes, then color management. Over the past number of months I've been using that method, as outlined by some of the people mentioned above and so far it's worked well for me.
The only discrepancy I'm having is specifically with the FX3's monitor. Either the built-in default LUT is off or the screen itself is just not very accurate. Following the same method with C-log2 files from the R5II gives me pretty accurate match to what I see on that camera's screen.
Interesting note about using CST nodes as the best option. Can you elaborate on why that's better than color managed? Or is there a video that goes into that? I can understand how fully relying on nodes gives you a little more flexibility, but otherwise, I've always assumed you can get the same results. ...Show more →
One is not technically better than the other and doing the same thing but you have a lot more control and flexibility managing color via CST nodes as opposed to doing so in the project setting. A bit beyond the scope of this discussion but you can even nest CST's within CST's as some color effects or adjustments similar to DCTL's may produce better results or are designed to work in a different working color space than Davinci Wide Gammut but would need to be converted back to a DWG to Rec 709 CST output transform prior to export. A much more common example would be if you had a timeline that had input from multiple cameras who's input transforms would need to be managed separately at the clip level. If you plan on going beyond the absolute basics in Davinci probably best to stick with CST nodes to manage color. Not complicated at all, set up your node tree with input and output CST's and then save as a Power Grade to quickly apply for future use.
Casey Faris is an absolute Fusion Guru and you can learn a ton from his site outside of Fusion topics but if you are looking to upgrade your color grading skills IMO Cullen Kelly and Darren Mostyn are probably two of the best YouTube resources to "up your game".
Tried my first test of the time lapse today on the a1 ii just because we have these rolling storms coming through. Lesson learned is I would lock focus and then put in mf because when I lost focused a couple of times it got jittery. Wind probably contributed too.
I think starting out using LUTs is a great idea - to get your feet wet with shooting LOG - rather than getting bogged down with node trees, etc., and giving up because it’s so daunting.
And yes - it is daunting starting out ;-)
Tried my first test of the time lapse today on the a1 ii just because we have these rolling storms coming through. Lesson learned is I would lock focus and then put in mf because when I lost focused a couple of times it got jittery. Wind probably contributed too.
Also, I give myself a few seconds of buffer at the beginning and end of the recording to account for any jitter from pressing the Record button. That or triggering remotely / via app. But I rarely do that. I prefer to keep the camera on airplane mode to preserve as much battery as possible when recording.
Tom RC wrote:
One is not technically better than the other and doing the same thing but you have a lot more control and flexibility managing color via CST nodes as opposed to doing so in the project setting. A bit beyond the scope of this discussion but you can even nest CST's within CST's as some color effects or adjustments similar to DCTL's may produce better results or are designed to work in a different working color space than Davinci Wide Gammut but would need to be converted back to a DWG to Rec 709 CST output transform prior to export. A much more common example would be if you had a timeline that had input from multiple cameras who's input transforms would need to be managed separately at the clip level. If you plan on going beyond the absolute basics in Davinci probably best to stick with CST nodes to manage color. Not complicated at all, set up your node tree with input and output CST's and then save as a Power Grade to quickly apply for future use.
Casey Faris is an absolute Fusion Guru and you can learn a ton from his site outside of Fusion topics but if you are looking to upgrade your color grading skills IMO Cullen Kelly and Darren Mostyn are probably two of the best YouTube resources to "up your game"....Show more →
That makes sense. A good thing to keep in mind.
And yes, I follow all three on YouTube and it’s really accelerated my getting familiar with Resolve / Fusion. Highly recommend.
For anyone wondering about A1 vs A7r5 for video [aside from the potential rolling shutter issues with the A7r5] I have found the A1 to be so much better.
The files come out with this beauty that I could not achieve with the A7r5, no matter how many months I put into trying (again - this is with Resolve.)
I guess it's a combination of the faster sensor on the A1 (though I do not shoot fast things, so no swinging the camera around here), A1's somewhat better dynamic range (although the tests show them to be very close in DR), and some other magic in there
So, for me, the A7r5 did not make a good backup to the A1 for video shooting, because the files were just not as nice, especially in the highlights.
ronno wrote:
I think starting out using LUTs is a great idea - to get your feet wet with shooting LOG - rather than getting bogged down with node trees, etc., and giving up because it’s so daunting.
And yes - it is daunting starting out ;-)
Watch some YouTube videos for a few days or a week and you will find that setting up node trees or using CST nodes for color management as opposed to using a LUT to do this is anything but daunting or complicated. And you want to use separate nodes for each of your basic adjustments like exposure, color & contrast so that you can easily go back and tweek your adjustments after the fact as needed. Doesn't mean don’t use LUTS as good ones are great to use but you are leaving a ton on the table if you just use a LUT and call it a day. Even a few nodes upstream of the LUT on your node tree for basic adjustment can do a tremendous amount to enhance the look of even the best LUTS.
Node trees and using CST nodes for color management sounds like scary and daunting concepts but are very similar to the concept of layers in Photoshop and quite simple to set up. No more complicated than learning the basics of Lightroom or Photoshop. IMO the color page within Davinci is what really separates it from the other video editing programs and more than worth the effort to learn some of these basics and IMO best to learn when you are just starting out. Head over to Darren Mostyn or Cullen Kelly's YouTube sites and you will have it down pat in no time!!
@Tom : re ” you are leaving a ton on the table if you just use a LUT and call it a day.”
Tom, I agree with you.
Above I wrote LUTs are a good way to get your newbie toes wet - I did not claim that using a LUT and then “calling it a day” is the best practice.
I really like this site. I have been through several of his videos and will be going through more. "The only nodes video you will ever need" was 30 min well spent. Thanks for posting this information.
Honestly, at just $200 extra, it's not a terrible alternative for anyone considering an A7IV but may be more interested in video, while still having a great photo camera included. In theory it should have better DR and sharpness than the FX3, plus the AI bells and whistles. And the EVF is cool. But yeah, the rolling shutter and heavy crops at higher FPS really keep it at entry/mid-tier.
What I'm guessing is, it's kind of their video-centric A7C variant, but instead of calling it A7CS, they just decided to slap on the FX branding. I'm sure it'll sell well.
I dont think you can count it as a C model with its weight- 700g .
n8rv wrote:
Honestly, at just $200 extra, it's not a terrible alternative for anyone considering an A7IV but may be more interested in video, while still having a great photo camera included. In theory it should have better DR and sharpness than the FX3, plus the AI bells and whistles. And the EVF is cool. But yeah, the rolling shutter and heavy crops at higher FPS really keep it at entry/mid-tier.
What I'm guessing is, it's kind of their video-centric A7C variant, but instead of calling it A7CS, they just decided to slap on the FX branding. I'm sure it'll sell well....Show more →
Among other tests that have also shown the A7IV sensor has better DR than the A7SIII/FX3. Maybe marginal, maybe not.
I’m not enough of an expert to explain exactly why, but I’m assuming the larger sensor is better tuned to resolution and DR than speed (hence the rolling shutter).
And I’m not saying it’s officially a C series camera but it does seem to fall into a similar tier, based on price and specs.
