coderemover

Well, that's a bummer. I wanted to record a full concert, about 1+ hour long, piece by piece. In many articles they stated that H2S can record 4k/60p for about 240 minutes at temperature 25°C. I recorded in air conditioned room, temperature likely less than 25°C, probably between 21 or 23°C. Yet to my surprise, about 15 minutes into recording the camera displayed a yellow warning about getting close to the thermal limit and in another 10-15 minutes the warning changed to red. Then it didn't allow me to start the next recording until it cooled down. So there is a huge discrepancy between what was promised and what I observed. I guess maybe I did something wrong with the setup. Besides doing the obvious like buying an additional accessory fan, are there any things in the settings I could make to reduce the heat production, without reducing quality to FHD? My settings were: 4k/60p, compression h265, LongGOP, color 4:2:0, bitrate 100 Mbps, recording to SD card.

AF with face detection alone, when not zooming, has worked perfectly fine for the type of video I make (piano concert). It never lost focus and actually even dealt correctly with two people playing on the same piano, which was a bit challenging, as they were not facing the camera, and they were quite close to each other. However, it tracked the selected person just fine and didn't get confused by another one slightly behind. It also tracked them fine after the performance, when they walked to the door. BTW: Eye detection still needs some work in the movie mode, though. Sometimes it detects eyes where there are no eyes, like e.g. it often mistakes *ears* for eyes. But that's not a problem for me, as from that distance to the stage face detection alone is good enough. The problems seem to happen only in combination of AF + zooming. AF alone is fine. Unfortunately I have only one lens so far, so cannot test whether it is a lens problem. I have applied all firmware updates, including the lens. I'll recheck when I get another lens (I plan to buy XF 50-140 f/2.8).

Ok, I heard about parfocal lens but they are... extremely expensive. Are there any affordable options for enthusiasts on X-mount? I thought that AF system could simulate parfocalness to some degree. It has all the information required to keep focus - e.g. the camera knows the movement of the zoom ring and also gets continuous information from AF sensors, so I really don't get why it could not simulate a parfocal lens, especially when zoom movement is slow and smooth. The AF system has somehow no trouble in following a person walking towards the camera quite accurately so I was expecting it could also correct defocus caused by zoom in a similar way. Edit: I just checked how "parfocal" my lens is and to my big surprise it is almost parfocal. In AF-S mode I could zoom over the whole range and the focus stayed mostly correct - I observed no wobbling and the focus shift is really minor. It works well both for distant and close objects. Here someone tested the other fujii lenses and found them to be mostly parfocal as well: https://www.thewanderinglensman.com/2017/06/do-you-need-to-refocus-after-zooming-on.html?m=1 However if I do the same thing in AF-C mode, it looks much worse. As if the camera tried to actively change the focus back and forth around the correct focus point erratically, making the image blurry for some time. So IMHO the problem is AF, not the lens. And this blur caused by AF-C is much, much worse than the focus inaccuracy caused by the lens varifocalness. This is a good news as now I have a really simple workaround - I'll just engage AF-lock before zooming and unlock it after I stop zooming. Anyway it would be good if Fujifilm added an option to lock AF automatically when zooming, or better fix the algorithms so they don't try to overcompensate / hunt for focus. Edit 2: Digging into it more now I can see that if I change the zoom quickly between 18 to 50, the focus is lost temporarily even if AF is turned off. So I guess what's happening is that my lens compensates for varifocalness electronically. Huh, maybe those two systems (camera AF and lens focus compensation) fight with each other while zooming? Anyway, if someone could test how this effect happens with their Fuji X 16-55/2.8 lens, I'd really appreciate. Maybe it is a Sigma+Fuji incompatibility.

I've just got my new X-H2s. Yesterday I went to a concert to make some videos. I have a Sigma 18-50/2.8 lens and I noticed that if I zoom in/out, even slowly, the focus is immediately lost and the camera really struggles hard to regain the focus for a while. The face detection works correctly and I can see the object detected all the time. Did anyone here notice it? Is it a problem with that particular lens only, or is it a problem with AF system algorithms which haven't been optimized to compensate zoom automatically?