Mark Romero 2

The new AI features in resolve take advantage of the Tensor Cores in the RTX series of GPU's. I believe that every (standard) nVidia carf in the RTX series has some tensor cores. I believe that the AMD cards don't have the technology to take advantage of the current AI features in resolve, but I could be wrong about that, so please double check on this first. Also I have no clue about the Quadro cards. (I know that you were looking at nVidia cards already so I just wanted to mention that I think you are on the right track going with an RTX GPU).

Thanks for posting. Personally, the S1H looks "best" on my monitor. Maybe a tad too much magenta in the skintones for my taste, but I think it could easily be tamed. In general, I like just a VERY slight skew toward magenta in the blues, which the S1H seems to have. Of course, this is completely subjective.

Thanks!!!

Ahh... didn't realize that the AF in video was useless with the MC-21. Guess most people were complimenting its stills performance. One other thing... how do you find the DR performance of the BMPCC4K compared to your a7 III? Are you shooting in BRAW with the 4K and SLOG 2 with the a7 III?

I assume you have already tried using an MC-21 or Metabones adapter with your canon glass on your a7 III, right? I just ask because a lot of people say that their L glass focuses BETTER on their a7 III with an adapter than it did naively on their Canon bodies. Maybe they had really old canon bodies where the AF was dated. I don't know. Could be pilot error.

Just so I understand though, the X-T3 currently downsamples in-camera to 4k though (from the 5.26K 60p), right??? Will the X-T4 still downsample in camera to 4K from 5.26K? Or will it actually have a 5.26K file?

Let's see... I'll take 5.26K for 60p, Alex. (Fuji Jeopardy)

Yeah, thanks for pointing that out. if you go into the full menu it shows more details about the codec. I have it set up in the quick menu (Q) and it shows less details. For anyone who might be interested, if you are in the menu system and click the DISP button, it gives a little pop-up blurb that describes the settings in a bit of detail. Can be helpful for those who aren't really sure what the settings do.

Neither HLG nor VLOG is available as 4K 60p 10-bit internal. Internally, you can record 1080p 60p 10-bit at 100Mbs (don't know if that is 4:2:2 or 4:2:0 though).

Yeah, I think you are correct about that. I guess I must have misunderstood. Not sure if they were using 8-bit internally, or 10-bit HDMI out.

Yeah, but my credit card is weak...

OK, how do you like the monopod? I assume during the ceremonies you just stand it up somewhere that it won't get kicked over by a guest and let it roll.

Yes, as of this writing, it is still cropped. Don't know if further on down the road, with any more firmware updates, it will offer full frame readout or not. I know a few other S1 owners and they have said that they find there is a bit too much noise in VLOG to use 4K 60p so they tend to use HLG instead if shooting 4K 60p. I would assume that is due to the crop instead of using 6K downsampled in camera to 4K, which is what a full frame readout would give. I don't want to swear by it since I don't own one, but I believe the 4K 60p in the S1H is cropped as well.

Is it easy to mix V LOG L (from a GH5) with V LOG from an S1 in Resolve? I noticed that for Resolve, out of the box, there is an ACES input transformation for VLOG, but don't know if that works the same for VLOG-L or not. Also, the color space transform node has a VLOG input, but no input for VLOG-L. I gotta assume that people shooting in VLOG-L on a GH5 are able to use ACES and CST in Resolve though, right?

Darn... I might need to get a GH5 now as a B cam to my S1.

Just to clarify, the S1 (with the film makers upgrade) shoots at 23.976 fps, as opposed to "true 24.0 fps." I would assume that before the upgrade it still shot at 23.976 fps. Resolution is 3840 X 2160, so it is UHD and not DCI 4K. At least, that is as far as I can tell. The S1 is VERY complex (for me) and some of the settings seem to affect / interact with other settings. So maybe there is a way to get DCI 4K and true 24.0fps that I am not familiar with yet. This is for the S1. I don't know what the S1H does.

Yes, the S1 can output 4K 60p (in aps-c crop) 10-bit 4:2:2 to an HDMI recorder. The S1 has some nice features when paired with the native Lumix 24-105 lens, but it has some drawbacks too, particularly on a gimbal. The native lenses are pretty expensive and pretty big and heavy. I use the S1 and 24-105 f/4 on my Weebill S and it is a bit hard for me to keep under control. The 24-105 is so big that you have to slide the whole rig back real far on the weebill s and that can limit how far up you can tilt the screen (I tend to shoot close to waist level on a gimbal, so it might be less of a bother to others who hold the camera closer to eye level). Oh, and that is with the 24-105 at the 24mm end. If you zoom in, then the barrel extends a LOT and you will have to rebalance and move the camera further back on the gimbal, if that is even possible. The S1 has two zebra memory positions, but unlike the S1H, you can't show BOTH zebras at the same time. On the other hand, some good things are that: The 25-105 is about as parfocal as can be, as far as I can tell and have seen on the internets The 24-105 has minimal focus breathing. While we all know about Panasonic DFD autofcus, doing a focus pull using the touch screen is REALLY smooth. It isn't fast, but it very gently eases in to the new focus point. It probably eases out from the first focus point, too. Yes, it is far slower than my Sony cameras, but it is much, much smoother and is pretty similar to a really good manual focus pull. (Much better than I can do manually.) And speaking of manual focus pulls, the S1 allows you to set the 24-105 to a linear focus response when turning the focus ring (so much better than the typical focus by wire response), and... You can adjust the amount of degrees of rotation in the focus ring from minimum focus distance to infinity. I think you can have a minimum of 90-degrees and a max of at least 360-degrees. There is a setting that says "Maximum" but got no clue what that means. The other good news is that the Sigma MC-21 seems to work pretty well with a lot of Canon glass (at least for AF-S, there is no AF-C at this time with the SA-21 adapter and ANY lens, either sigma or canon), and your typical manual vintage FF lenses work well on the S1. That's a good thing because there isn't a whole lot of L Mount glass available now. For example, the only native 50mm lens is an f/1.4 costs around $2,300 and there isn't anything on any lens roadmap for a native 50mm f/1.8 for a couple hundred dollars. There is an 85mm f/1.8 on Panasonic's lens roadmap, but I wouldn't be surprised if it is big, heavy and expensive.

Just out of curiosity, you do use SOME form of stabilization with your current X-T3s, right? (Or are you using X-H1's???)

Thanks!!!

Thanks. Does that hold true if the final codec is h.265 (as opposed to h.264)? Also, does the "preset" being used (fast, slow, very slow) not affect the final file size?

Well... with the X-T4 release, maybe pick up another X-T3 body at a good price used and get a gimbal??? As you can probably tell, I have no idea of what your workflow is like on site, so I my suggestion is probably as useless as can be. That hasn't ever stopped me in the past

Yeah, exactly. The X-T3 is the "better" camera for most situations when compared to the a7 III. Just in those cases that I personally find myself in, where dynamic range is paramount, then the a7 III would have been better. So I figured just get the S1 (which is, BTW, a really cool camera, but man is it HEAVY!!! Makes my old D750 and DSLR lenses seem light.)

While I really liked my D750, I think they should have just spent the effort making an F Mount to Z Mount adapter with a built in AF screw drive motor so that people could adapter their older Nikon AF-D lenses on the Z6 / Z7 . Otherwise, they should have just made the D780 as a mirrorless F Mount body so that the user would get access to all the AF points while using the viewfinder., and still use all their old lenses natively. As it is with the D780, one only gets the more narrow spread of AF points when using the viewfinder. And since the sensor in the D780 has all those PDAF points on it already for live view (LCD shooting), seems like a wasted opportunity.

Not trying to take anything away from the dynamic range of the X-T3 (and even the X-T2), but doesn't the chart you posted reference the a7S II and not the a7 III? Do you have a chart that features the DR of the a7 III? And while I trust the lab results of the Cinema 5D, they should also be pointed out that they are still lab results. One of the factors that affect dynamic range performance in the lab results is the amount of noise in the shadows. In the real world, SLOG 2 and SLOG 3 are almost always over-exposed by 1.5 stops or so, and because of the way those codecs handle highlights, they can be brought down in post and still retain detail (yes, there is potential banding). I don't know if that is the same for F-LOG or not. When I was thinking of buying an X-T3 a while back, people generally advised NOT to overexpose F Log by more than half a stop. I've searched around and haven't ever seen a "real world" comparison of DR between F Log on the X-T3 with S LOG 2 of the Sony a7 III where different over- / under-exposure values are used to try and squeeze as much DR as possible out of their codecs. I've had a play with some F Log footage from an X-T3 that was sent to me by various owners and to be honest, the DR looked to be slightly less than what I could get out of my a6500 shooting in S Log 2. Could be operator error though so hardly scientific. And for anyone who cares, I bought a Panasonic S1 for the dynamic range in 10-bit 4:2:2 30fps, so I am not just trying to "defend" Sony. Also, I hope - and expect - that the X-T4 will be an amazing camera.

Yeah! Absolutely can't argue with that.