Pool Shark. New footage.

[vladnik]

interesting read ....
http://www.luminous-landscape.com/es...d_sensor.shtml
we should think big...
instead downrez from 2.5k to 1080p ...
go 2.5k to 4k ...

[randyman]

we should think big...
instead downrez from 2.5k to 1080p ...
go 2.5k to 4k ...

Not to go OT, but that's something I have in mind; by shooting anamorphic, even without any firmware mod on the BMC, the resulting image will be way over the 4K mark and more than enough for theatrical projection. Very exciting.

[Andrew]

we should think big...
instead downrez from 2.5k to 1080p ...
go 2.5k to 4k ...

that's something I have in mind; by shooting anamorphic, even without any firmware mod on the BMC, the resulting image will be way over the 4K mark and more than enough for theatrical projection. Very exciting.

I'm not following what you guys are saying. Why up-res from 2.5k to 4k? I don't see a reason.

[vladnik]

first : raw images have more resolution than any compressed codec (post production debayering is better than any quick'n'dirty in camera ) ...
second: camera without aa filter can match in resolution camera with 30% to 40% more pixels ....
so why waste all that potential on 1080....

[AndrewDeme]

I'm not following what you guys are saying. Why up-res from 2.5k to 4k? I don't see a reason.

In a nutshell, the higher the resolution the higher the frequency of moire (will take finer patterns to set it off), till you get to the point that the frequency is so high that the moire is no longer visible. Is one of the reasons Nikon have their D800E option as it is a 36 MP sensor.

[Andrew]

Because demosaicing a bayer pattern image won't give you full res rgb. You want to over sample on capture for your delivery resolution. 2.5k to 1080p is a pretty decent oversampling.

I don't think there would be much (if any really) benefit to upres from 2.5k to 4k. If you want max detail finish at 2.5k. It probably won't be resolving a true 2.5k, but at least you won't be losing anything.

Sorry for getting OT.

[Andrew]

AndrewDeme, I agree but that that's an example of a native high res sensor. I can't think of a reason to upres a 2.5k image to 4k.

[AndrewDeme]

AndrewDeme, I agree but that that's an example of a native high res sensor. I can't think of a reason to upres a 2.5k image to 4k.

Totally agree, was just commenting on the relationship between resolution and the frequency of Moire patterns.

I for one would much prefer no AA filter as there are a number of easy ways to solve the problem. In a run and gun style camera sure it makes more sense, but for the BMC I will be putting more thought into the art I produce so am happy to take this into account and if for some dumb reason I screw it up, then post production here I come.

If it was me, I would have used a bit of visual effects and set that mesh chair on fire, smoke and all.....would have fixed the moire and left the viewers with another question in there minds about just what was going on at the time Casey was shooting pool. Even could have the VFX lighting from the flames playing across her as she is walking around the table and bending over taking each shot.

Take that you pixel peepers (;->>

[randyman]

Again, sorry to go OT, but just to explain what I was talking about… I wasn't meaning to up-res just for the sake of up-resing.

If I shoot at the BMC's RAW resolution (2432x1366) using a 2X anamorphic lens, I wind up with an image that's 4864x1366. Yes, the horizontal dimension has been artificially up-resed, and that aspect ratio is ridiculously wide (as the 2X lens was really designed to be shot 4:3 to start with.) But if I chop off the sides to get 4096x1366 for a 3:1 aspect ratio, it's pretty amazing to me that by using an anamorphic lens, the BMC can deliver something that would stand up to 4K projection.

In reality, it's far more likely that I'll crop the sides down to a more traditional 2.35:1 (Panavision) and down-res the result to fit the standard 1920-wide HD frame. Sorry again for the OT.

[Andrew]

Sounds good Randy and I see what you're saying now. Looking forward to seeing your anamorphic stuff! I love a super wide aspect ratio.

[razz16mm]

Totally agree, was just commenting on the relationship between resolution and the frequency of Moire patterns.

I for one would much prefer no AA filter as there are a number of easy ways to solve the problem. In a run and gun style camera sure it makes more sense, but for the BMC I will be putting more thought into the art I produce so am happy to take this into account and if for some dumb reason I screw it up, then post production here I come.

If it was me, I would have used a bit of visual effects and set that mesh chair on fire, smoke and all.....would have fixed the moire and left the viewers with another question in there minds about just what was going on at the time Casey was shooting pool. Even could have the VFX lighting from the flames playing across her as she is walking around the table and bending over taking each shot.

Take that you pixel peepers (;->>

The moire of the chair is not the camera. the color moire on the speaker grills and the top of the carpeted speaker in the fore ground are though. For still photogs, getting by without an AA filter is workable, but for motion, unless you want to do a lot of post work individually spot correcting every frame, or setting up correction masks, you will end up processing the whole image to a lower resolution than a well matched AA filter would cost you to get rid of the moire. When aliasing artifacts occur they affect lower than Nyquist resolution details too.

[AndrewDeme]

The moire of the chair is not the camera. the color moire on the speaker grills and the top of the carpeted speaker in the fore ground are though. For still photogs, getting by without an AA filter is workable, but for motion, unless you want to do a lot of post work individually spot correcting every frame, or setting up correction masks, you will end up processing the whole image to a lower resolution than a well matched AA filter would cost you to get rid of the moire. When aliasing artifacts occur they affect lower than Nyquist resolution details too.

How about I say it a different way.....am pretty happy with what I am seeing in the footage shot so far. If BMD still have ability to make ongoing improvements then I will be over the moon.

[razz16mm]

How about I say it a different way.....am pretty happy with what I am seeing in the footage shot so far. If BMD still have ability to make ongoing improvements then I will be over the moon.

I like the quality of images coming from the camera too. Aliasing performance is just another element that must be considered when picking shots.

[Andrew]

but for motion, unless you want to do a lot of post work individually spot correcting every frame, or setting up correction masks, you will end up processing the whole image to a lower resolution than a well matched AA filter would cost you to get rid of the moire.

How effective can post moire and aliasing removal be? I totally get what you're saying, but maybe spot removal won't be so bad when you pair it with Resolve's amazing 3D tracker, assuming it's only an occasional problem. I wouldn't want to spot remove it from every other shot.

[RyGuy]

Uhh... at 0:23 upper left corner you can see INTO the lightbulb. I haven't seen anything like that except for one Alexa test video. Obviously not all lights are created equally and there are different watt bulbs, but I would NEVER dream about seeing that on my 5d.

I looked everywhere for that test video on vimeo, but couldn't find it. It's basically a side-by-side of a 7d, Alexa, and one other camera (F3?), shooting a black blanket with different objects on it creating a super high contrast shot. It goes from a deep shadows, all the way up to a turned on light bulb. The two less expensive cameras had a hard time on both ends, but the Alexa could see deep into the blacks, as well as the actual ignited coil of the lightbulb. Mind blown. Think I might be seeing that again...