Another nice C300 short and Review

[Charles Papert]

Haha, thanks Barry. Can you tell this issue has been percolating in me for some time (started with the 35mm adaptors four or five years ago)...

I spent so many years having to take focus deadly seriously. Getting called in for a very somber meeting with the DP and director because of focus issues in a critical take that weren't noted (and that's me as the operator, folks, not the focus puller), working with AC's to help them with focus cues, putting extra effort into hitting my marks on Steadicam or maintaining equal distance to actors (chalking out circles around the subject for 360 degree shots). I've seen plenty of people get fired off jobs because of focus problems. I've done the firing myself. It sucks but there it is. As a DP, now I sit behind large high end monitors staring intently at two cameras and calling in focus notes to the AC's via walkie. Sigh...

So anyway--it's just a bit painful to watch the keys to the kingdom (i.e. large sensors and the resulting depth of field) be handed out to the "masses", and be treated so cavalierly by so many. One of the problems is that few people getting into the camera department these days want to be AC's--everyone's a DP (or a director) and the tremendous amount of skill required to nail focus with few rehearsals and marks is not being given as much weight as I think it should.

Of that Steadicam sequence, the first three shots are largely soft. The last two are fine. I'm not going to go through and count, but there are many others. The closeup at 1:44, for instance, is continuously out.

Not trying to single out these guys, obviously they moved quickly and shot a film intended to be a test. I just see more of this kind of thing than not and it's hard to not make a little stink about it.

[Brian Drysdale]

Yes, very much a case if you want to shoot having motion picture images, learn the skills to get those images, they're not grabbed, there's a precision and a process you need to nail. Any 1st ACs I know are deadly serious about setting shots up and getting marks etc. It can be difficult enough for the less experienced on 2/3" or Super 16, never mind Super 35.

[Justin Molush]

Would love to get my hands on the prores file to see how it handles some intense color pushes. Downloading the .MXF files now to give them a whirl...

[Ken Diewert]

Originally Posted by Charles Papert

I spent so many years having to take focus deadly seriously. Getting called in for a very somber meeting with the DP and director because of focus issues in a critical take that weren't noted (and that's me as the operator, folks, not the focus puller), working with AC's to help them with focus cues, putting extra effort into hitting my marks on Steadicam or maintaining equal distance to actors (chalking out circles around the subject for 360 degree shots). I've seen plenty of people get fired off jobs because of focus problems. I've done the firing myself. It sucks but there it is. As a DP, now I sit behind large high end monitors staring intently at two cameras and calling in focus notes to the AC's via walkie. Sigh...

Charles,

After just spending a couple of weeks pulling focus on an Epic for the first time, I can reeeaally appreciate your comments. In the pre-production meetings I was told we'd have the Canon mount..."auto, rack focus"...blah, blah, blah, "who needs focus pullers",...

2 days before shooting starts, I'm told we need to pull focus because we have no Canon mount. I had to 'make' a speed crank. We're shooting at t2.8 at 50- 85mm from distances of 6 to 15 feet, and I'm getting glares from the director when I take an extra few seconds to double check focus, and rehearse a couple of racks.

I'll take 1080p sharp over 5k soft anyday.

[David Heath]

Originally Posted by Oleg Kalyan

Мy colleagues from Russia noticed some aliasing signs. They are in places where straight lines are in focus. Compression artifacts? Comments?

I doubt they're aliases - was the material interlaced? Looks a little like differences between the two fields of a frame....? Assuming it's progressive, looks more like what you'd get with a diagonal line and horizontal scanning, regardless of camera system.

To be honest, I wouldn't really like to speculate much on material that's not direct from camera, no additional processing, compression, etc. And at this stage I'd really like to see some decent scientific tests, zone plates ideally. That's the only way to really see just how good or bad the aliasing etc is. On "real" pictures there are just too many variables to be quantitative.

From what's been released, it's most likely that a standard 1080 zone plate is going to look pretty good. What I'd like to see is how much aliasing much finer detail gives, say about 1500-2000 lines. Obviously the camera won't - can't - resolve it properly, but will it just be flat grey (as it should) or will it be coloured aliasing?

[Mikko Topponen]

Originally Posted by Oleg Kalyan

Мy colleagues from Russia noticed some aliasing signs. They are in places where straight lines are in focus. Compression artifacts? Comments?

Did your colleagues deinterlace the image? Or maybe the creators of the video did that. It's mind boggling how many people accidentally deinterlace their own progressive images causing these huge stairsteppings and halving of resolution. I see it monthly at work. Take material into FCP don't check what it's doing = BANG, deinterlaced. Take material to Premiere, don't check what is happening = BANG, deinterlaced.

Especially FCP is quite prone to it if the material has come out of mpegstreamclip or others that flag the files as interlaced.

[Oleg Kalyan]

Mikko, I would love to see the same piece MXF!

[Barry Goyette]

Originally Posted by Mikko Topponen

Did your colleagues deinterlace the image?

I really don't think that's it. There are a few rare cases of aliasing in the video, visible in the shot referenced, (you see it in his zipper as well) and in the close up of the main characters eye upon returning to the tram. In each case it's a sharp, high contrast line at an oblique angle. These are all visible in the 1080p download. I've not seen anything like it in the mxf's, and I kinda doubt you will ...but any angled sharp edge like that will alias on any sensor unless it's blurred. Non-issue, in my opinion. The fact that both moire and aliasing are controlled to a great degree is really good news.

[Alister Chapman]

I noticed some stair-stepping on strong verticals and horizontals at the Visual Impact C300 open house that didn't appear in the near identical F3 and Alexa shots. BUT I have no idea how the camera was set, nor do I know how the monitors in use behave with enough conviction to say I saw an artefact. Whatever it was it was pretty minor, but it's certainly something I'd like to investigate further.

[Kevin Dooley]

Originally Posted by Barry Goyette

but any angled sharp edge like that will alias on any sensor unless it's blurred.

I think this is the key point - moire and aliasing have been in every digital sensor that I have ever seen. With DSLRs and the massive (and usually sloppy) down-rezzing from sensor to recorded video the problem blew up and it became a big consideration when shooting. However, people have been turning sharpness down on cameras forever to minimize this (and for other reasons)...

[Alister Chapman]

This is where a well designed optical low pass filter comes in to play. It should reduce the resolution of the incoming light to somewhere around half the luma pixel count to prevent stair stepping and aliasing.