F35 vs. Blackmagic 4K

Hi James,
Yes to me the native iso on the bmc4k looks like 200. It's the only setting I would use for that camera in a professional environment.
The pattern noise at 800 iso horrible, completely unusable.
At 400 iso, it's passable if you need need extra light, but fixed pattern noise creeps up in the shadows.
 
Interesting, I had no reason to disbelieve my mate (a gaffer) but I was somewhat shocked as ISO 200 is a shockingly low figure for the market - only the other day I spoke to a new DoP who said he often shoots "800 ISO, at T1.3, on the EPIC and it's barely enough".... seems lighting is a dying art.

Fixed pattern noise is what put me off the original RED ONE, so I can't think I'm going to love the BMC4K. Again, comparing it to a $250K camera like the F35 seems almost cruel, but I think the people who buy the Blackmagic genuinely think there is no reason you could possibly want or need an ALEXA etc....
 
Regarding pixels and photosites I think John's article is a little biased and if I didn't know any better I'd say he's an anti-bayer advocate.

Sure a true 444 strip sensor is going to be the only one to produce 4k for each RGB photosite, but there's really no need to completely discount sub-sampled chroma techniques that so many cameras use and can produce beautiful images with.

:)
 
If only there was some nice medium ground out there....

I agree that Bayer sensors are a minor miracle and do a much better job than you'd think they can, but they definitely do not render 100% of their claimed resolution.... ever.

I would never argue with anyone that a RED M-X is a lower resolution camera than, say, the F35. I just take issue when RED etc.. try and tell me their "4K" is 4x the resolution of my 1080P. That's a flat out lie.
 
To me this test http://www.dvxuser.com/V6/showthrea...-Where&p=1986432515&viewfull=1#post1986432515 also proves that sharpness and contrast detail from Bayer sensors can be marginally improved if and when they are done in post production rather than in real-time due to limits and trade offs of real-time debayering and encoding in camera.

One huge benefit of RGB striping that just dawned on me is the simplicity and ease at which it can be processed in camera, no debayering mumbo jumbo complexities which compromise quality.
 
starcentral said:
Regarding pixels and photosites I think John's article is a little biased and if I didn't know any better I'd say he's an anti-bayer advocate.

Sure a true 444 strip sensor is going to be the only one to produce 4k for each RGB photosite, but there's really no need to completely discount sub-sampled chroma techniques that so many cameras use and can produce beautiful images with.

:)
Click to expand...

I think John's basic argument boils down to quality over quantity. When Panavision teamed up with Sony to develop the Genesis(F35) the express goal was to mimic, as closely as possible, motion picture film - and that meant having full red, green and blue information for every pixel (just as the layered colour dyes on film effectively does).

His point is that a bayer pattern sensor (or any kind of sensor pattern for that matter) that's interpolating (i.e. guessing) colour information is inherently compromised, which is true. And although you can certainly get more-than-acceptable images from sensors that are interpolating data, it still doesn't mean that more is really more.

I think the F35's capacity to still wow people owes a lot to this fact, and I think the fact that most people seem to find 4K from the C500 a little underwhelming, but find its 2k image glorious, speaks to it as well.

Personally, I think bayer patterns are perfectly fine (excluding occasional moire issues) but they clearly work best when they use their oversampling to deliver a full colour resolution (4:4:4) image - even though those images come at a 'lower' resolution.
 
Well I think that's also been ARRI's point for a long time, 4k is not really 4k. It really just ends up being really good "HD" and if it were shot on a good sensor design like RGB stripe then you are fine with 5760 x 2160 pixels like on the F35.

For me, I don't care about resolution.. what I want is those damn highlights and skin tones from my F55 like I see on the F35 or Alexa! :crybaby:
 
starcentral said:
For me, I don't care about resolution.. what I want is those damn highlights and skin tones from my F55 like I see on the F35 or Alexa! :crybaby:
Click to expand...

Yeah, and what's important to remember about color is that it is somewhat independent of resolution.
The comparison of Film Scientist's two oranges is clear whether the resolution is at 1/4, showing on Youtube, or at full-rez playing to a 60" Plasma. I think that's a valuable point. The F35 brings something to the table that can be appreciated by a variety of viewing options.

Even when talking about gains from downsampling 4K to 2K to gain color information, it seems moot if that 4K color isn't being sampled sufficiently (losing subtlety due to debayering).

Thanks again to Film Scientist for such a simple, yet illuminating, comparison. It opened my eyes. And it's an effective tool for communication, too.
I can walk up to the average pedestrian, show them them the two samples, and they choose the F35 every time!!!
How cool is that?! :)
 
I just tried to blow the TIFF of the F35 focus test up to 4k in photoshop.
bicubic sharper + sharpen filter
It does hold very well!
 
You must log in or register to reply here.
Back
Top