Getting to know the new camera
Added 2018-12-01 15:12:17 +0000 UTCThe production workflow on the new camera is very different and I'm having to make some adjustments to get started with it.
Yesterday I changed my shooting setup to have that monitor above my work surface as my reference display. In the past I've used my right monitor(you can see it in quite a few videos) but the new camera outputs YCbCr over it's HDMI port. My Nikon must have output RGB because my DVI monitors has no problems with it. But none of my usual monitors could handle YCbCr. So I had to mount the only HDMI monitor I have where I can see it. It's not a great monitor, so I don't want to replace on of my three with it.
When I shoot the camera is usually behind or beside me. With the Nikon lenses on this camera though, it's going to be pushed back even more. The APS-C lenses on a MFT mount end up being basically double their focal length. So the shots are all tighter now. That moving the camera back coupled with the screen on the camera not moving at all made this setup mandatory.
I'm going to try shooting some real footage I'll actually use in a video today. But I may have to stop and go back and retake them as I'm working. The video output from the camera directly is very different from what it looks like after you've processed it. Here's a before and after of the Gameboy shot from the test video: https://imgur.com/a/fGRAszb . You see on the camera what the first image looks like because that's what you get out of it. So it can be harder to get the exposure just right visually and I have been tending to lean too high I think. After trying different grading a approaches on a clip for about an hour I finally had an idea to take advantage of a feature the camera has and it helped a lot. I learned how to export a "LUT" which is basically a settings blob for grading footage automatically. I exported my settings from the gameboy shot since that looked great. The camera can apply the LUTs in realtime while filming so you can preview what it will look like when you're done working with it. Here is a comparison of the shot I spent an hour grading vs. that same shot done with the gameboy LUT so I could visually see the result better: https://imgur.com/a/6vHIoVd
So I'm making good progress. Today I released this video: https://www.youtube.com/watch?v=mr61p6SFPy8 which has been a "backup" video for me since July that I never used. It was made using the old mic setup and I figured now was my last chance to use it without it being too wildly different from my current production standards. I'll release that Nikon adapter video Tuesday and hopefully this time next week will be my first public 4k video! I'll put it up here as soon as it's ready for you all of course though!
Comments
Yeah, my script became out dated after LTT released that CPU video so I know I'm going to get a bunch of comments about customs CPUs like that. Custom built chips like the ASIC coin miners are something I wouldn't really count. And I don't think build to order CPUs are ever going to hit the general consumer market. I think economy of scale just won't work. Windows RT should have been an example of how an ARM based consumer desktop environment could work. But it failed completely due to software incompatibility, which admittedly was definitely a bit intentional on Microsoft's part. Microsoft has been teasing a new ARM based Windows 10 with x86 compatibility. It would be interesting, but I'm not going to hold my breath that it's going to shake up the market. I think the concept of a "home computer" is going to fade away more likely than x86 be replaced as people buy more powerful phones and tablets instead. Or maybe Chromebook like devices that are basically a modern day terminal where the underlying technology could be completely changed without anyone noticing or caring. I see the enterprise space as being totally different. Servers with 20+ cores have been common place for a while and it wasn't until recently that consumer CPUs got more than 4 without breaking the bank. I could see massive changes happening at the drop of a hat there if someone comes out with something that can replace an x86 CPU with better $/performance. I think it helps that in server environments most of the software you'll be running is bleeding edge. So it's much more likely for it to get support for a niche compiler for a new type of CPU.
Tech Tangents
2018-12-07 18:23:48 +0000 UTCYou make a good point about architectures. X86 clearly won, but now basically every company can build their custom processors using Aarch without sacrificing compatibility. I think we are at the beginning of a “new architecture war”: Amazon achieved a stunning feat with their own server grade ARM processor, outperforming a comparable Xeon using a fraction of the power (=cooling) and cost. To me at least this is a great outlook :)
Daniel Helgenberger
2018-12-07 16:31:41 +0000 UTC
