no_connection

Can it output 720p50? or at least 1080p50? Can it use synchro scan for shutter speed? We are looking for a camera that can be used live but also be used for recording.

https://www.pugetsystems.com/recommended/Recommended-Systems-for-DaVinci-Resolve-187/Hardware-Recommendations

It's more of a pointer to "there is a lot more to think about than just grabbing some high core cpu and a graphics card" type of thing. CPU performance between now is probably only about 30% better than 5 years ago clock for clock, but we have more cores and higher clock speed nowdays, with some exceptions. And lower power draw. There is a lot of different ways to approach it depending on what you need to have done, if it's rendering maybe buying a used server with decent CPU and shove it in a different room and let it work while you do other things. If it 's scrubbing then maybe that needs some proxy tweaks and clever tricks. Maybe a server that transcodes it before you start edit. H264 with quick decode option and low keyframe number can be very snappy with pretty thin files if you struggle with IO performance. https://fxhome.com/forum/discussion/42415/transcoding-to-fast-decode-avc-for-timeline-edit-performance If you are low on RAM that can be a problem, and to make it even worse RAM price is way over the top so a new system ain't going to be cheap. For GPU Resolve is going to have to play nice with it and I can't make any big pointers there since I have not had it work perfectly in any workstation. Even OS might be a thing, https://www.pugetsystems.com/labs/articles/DaVinci-Resolve-14-Performance-Windows-vs-Linux-1126/

I have seen so many bad unsharp mask tools that I lost hope, turns out the one in gimp is pretty decent, don't have premiere so I have no idea how that looks but I will give it the benefit of the doubt instead and hope it's somewhat close to what a free utility can give. I guess you could run all frames through a batch to get what you want from gimp but the amount of data would get very high very fast.

"Unsharp mask" is the worst of the worst way to "sharpen" an image. If you want something useful look up wavelet sharpen, you can find builds of gimp with it working out of the box. If you needed to use unsharp mask you could upscale and tweak it then downscale again to make it a bit less ugly.

I don't know what he means in the video, to me the EOS R does NOT get sharper with the sharpness setting, sure it adds a halo and increase the local contrast making letters stand out more, but I don't see any indication of "size of blur" getting smaller/more detailed. Sure it kinda looks sharper and does make it easier to show some detail but not really fixing the softness. You can see the outline of the Chinese? letters around the 4:08 mark on the battery grip box get more defined while on the X-T3 (without halos mind you) you can see they have detail inside of them. So I guess my take from it is, yeah it does not lack sharpness as in halos and edge contrast since I really don't like that. But it sure does lack detail/resolution compared to what other offers without being overly sharpened (in this case X-T3). Now if this is the megapixel race of 4k or not I'm not sure, but cramming too much content all the way to nyquist do mess up a lot of scaling algorithms and do become a hassle at times, with audio you want the resolution to be way higher than the content so I don't see the problem doing the same for video if you know what you are doing. I'm kinda balancing between "yeah it's soft and lack details but" and "I want that 1:1 pixel detail because reasons" So is the real resolution about 3k vs 4k(example)? and does it matter at the end of the day? If you had quick turnaround and uploaded camera file in 4k to youtube, do the 1080 being scaled by youtube look better than similar X-T3 shot? How does it look like on a phone?

Could you post some comparisons (in png not tif please) since you have a few of the contenders here?

I agree, would have been nice if the video was actually about bokeh and not just amount of blur, although that was maybe interesting. It does show that ppl care more about balance and composition than fast lenses. ie if lots of blur don't add anything to the image, don't use it. Or use it creatively if you want to put the viewer at unease for some reason.

You fight some compression artifacts but it sharpens up pretty nicely. Would probably add more selective sharpening to avid trees being too dominant.

I have a Cosmicar 25mm f1.4 c-mount television lens that have developed some sort of speck in the middle on two elements inside, makes an interesting "hole" in the bokeh. I don't have anything to use it on at the moment tho =/

"it's still an onboard mic, it's for reference, it's nothing you could use for production" *hears his lav mic crackle and pop loudly* For tests it's very unclear what profile or settings they used. "it's nice from afar but far from nice"

I replayed it several times and it looks wierd cause of the clipped bright sky background being suddenly blurred out. And not iris change, my mistake.

It does show the auto iris being very ugly, and autofocus hunting and being very ugly. The slider shot looks so bad.

Just because a cop didn't stop you driving too fast does not make it right to drive too fast. Even if the cop takes money under the table to look the other way. Many artists seems fine with ppl using them, and would probably not be as big if there where not, you just need one of the not so happy artists tho to get really hurt by the system.

@androidlad That is good to hear cause that would have been a dealbreaker. I can understand iris being somewhat harder to get smooth.

Probably got the idea from this

Looks like auto ISO is useless in video of that is important to you.

Don't forget the "agree not to sue" clause when you did that. At least they got roasted for it.

That made me laugh out loud, on the inside. This is not a new camera right? I remember looking at something Olympus but it had some serious things limited in firmware for no good reason so gave up on it.

Nice thanks. You probably picked this up already but it's 400 Megabyte not megabit. So a lot of room for data transfer. Rurmor has it that they will put best possible coded in there that they can, and if that is true it will have internal prores raw, cause it can. Not that that prores raw is be all end all but it's not bad ether. On the contrast based AF. It was stated somewhere to have deep learning built in, so in theory it could rival your ability to manual focus. If that ability is used like we want or just for tracking and identifying faces who knows. It's not like pdaf don't come with drawbacks ether as seen by several interesting lines on some cameras in certain light. And that you have to fill them in all the time. For the back display, Maybe they thought it would interfere with the HDMI port. Or as I have been using the flipoutscreen as an armrest to hold the camera it might be too fragile for such a large camera system and they did not want to risk users using it in that way. Why don't anyone make a cheap small HDMI monitor that is powered off usb just for this use? Don't need much to replace flip out screen the few times you need it. Or maybe there is an app for that :P

Now that would be sweet, have only seen it mentioned for S1R so far but good info is far in between all the noise. I hope price going to be somewhat decent for the S1. S1R gonna be way too much just like Z7.

They put pixel shift on the wrong camera. S1 is the one that would benefit the most. And I highly doubt it would eat any profit from the S1R ether, just more ppl going for S1 in the first place.

Button layout and ergonomics looks great! and finally a dedicated back focus button instead of reassigning ae-l button. I like the top LCD, none of that OLED crap! And the big lens for EVF looks very good. It better not have a glaring flaw to break it all.

It barely does 25 fps so I don't think that will work all that well. The m43 version max out at 28 fps so not that impressive from ether of them. I see no indication that those bits are in any other form than linear so you need the bits when you shove them around for whatever profile you want to burn into the encode. And 12bit might be pushing it as it is.