Is the Alexa better than the F35?

JMtheDP said:
Steve

That's kinda the point I'm trying to make. If you were cutting the F35 with a camera that did 4K as well as the F35 does, you'd notice a difference (but only viewing at 4K!).
Click to expand...

Yes but it doesn't exist. It have to be CCD 3 strip 4K.

JMtheDP said:
Also @ SJX, I don't think Sony will be coming out with any more upgrades for the F35. I don't know how they planned to get 2K out of it, but 1080P/2K is not really a big difference anyway. In fact 1080P will scale to UHD (which is what most people mean when talking about 4K) better anyway.
Click to expand...

I know sony will not give this update anymore. I wish they had not give up on this 2K update for F35. But it means is possible to output other resolution different from HD with F35. This was my point.
 
I'm not entirely sure it is possible to get other resolutions from the F35 - after all, 2K is actually an aspect ratio difference, rather than a resolution one per se.

What would have been really cool though is a mode to vary how the pixel sets are grouped together, similar to what happens when shooting 2.37 (i think that was the ratio) on the Viper
 
Regarding the 2K for the F35 thing, can someone post the source of that claim again? I only ever heard it as a rumor and never from Sony directly. Given the layout of the sensor, I'm not clear on how they would even do it, without pulling from a larger area on the sensor (more photosites). In short, I'm dubious.

As for the 4K uprez conversation.

Simply put, the image sensor for the F35 was designed to provided the best possible 1080p image possible with a single imager. It was optimized for that resolution. It uses photosites that are not square. The camera internally processes the 12MP of raw photosite data to generate a very sharp yet natural 1080p image. The camera does not give the user access to record the photosite data.

It's optimized for 1080p, and while you can do math to scale it to other sizes, that's the resolution it's intended for and will perform best at.

The Bayer cameras are a bit more confusing because there are endless arugments about exactly what is optimal in debayering.

G R - G R - G R - G R
B G - B G - B G - B G

With a bayer sensor I could decide to use each four photosite block to generate one RGB pixel.

If the light arriving at those 4 photosites is (out of 255 max for demonstration):
R/G/B

100/212/52 89/222/48
94/210/201 110/40/38

If the photosite filter pattern is :
G R
B G

We'll get:
R/G/B
89/126/201

It shouldn't take long to realize that result isn't actually very accurate to the light received for the green and blue values. But it is the sort of data that can confuse a very simple debayering algorithm. To get around this, people do more complicated math that looks more carefully at other surrounding values.

But to complicate matters even more, people frequently want even more resolution from their bayer camera, so instead of debayering a block of four photosites, they try debayering one RGB pixel per single photosite.
Showing you that math and the edge detection checks that become necessary would be very unwieldy to do here.


Suffice to say, you can upscale the F35 in post, and generally the results should look great, but they will be less accurate and may suffer when their are tight patterns. Bayer sensors have similar problems. Sony and Panavision did something very smart and very expensive, but that is difficult to explain and therefore it doesn't market well when everyone else is willing to scream louder with bigger numbers.

"I have the Gruntmaster 4000 with the RX engine 2D torque!"
"Oh yeah? This is the new Gruntmaster MEGA 8000 with the XRD Mega Engine 4D torque and it has Real Time Flibberflam!"

Which one is better?
 
I think pixel count change then is resolution change because is not post crop. But yes many cool things could come for the F35 if sony didn't forget it.
 
Viddovation said:
"I have the Gruntmaster 4000 with the RX engine 2D torque!"
Click to expand...

Someone's been watching Dilbert :)

Basically what has happened with the advent of RED is people have forgotten to separate sensor and output resolution.

It used to be very easy previously, with 3-chip systems probably doing it better than anything else. Sadly, there are issues inherent with those and they would never be viable with sensors the size of 35mm.
 
Viddovation said:
Regarding the 2K for the F35 thing, can someone post the source of that claim again? I only ever heard it as a rumor and never from Sony directly. Given the layout of the sensor, I'm not clear on how they would even do it, without pulling from a larger area on the sensor (more photosites). In short, I'm dubious.

As for the 4K uprez conversation.

Simply put, the image sensor for the F35 was designed to provided the best possible 1080p image possible with a single imager. It was optimized for that resolution. It uses photosites that are not square. The camera internally processes the 12MP of raw photosite data to generate a very sharp yet natural 1080p image. The camera does not give the user access to record the photosite data.

It's optimized for 1080p, and while you can do math to scale it to other sizes, that's the resolution it's intended for and will perform best at.

The Bayer cameras are a bit more confusing because there are endless arugments about exactly what is optimal in debayering.

G R - G R - G R - G R
B G - B G - B G - B G

With a bayer sensor I could decide to use each four photosite block to generate one RGB pixel.

If the light arriving at those 4 photosites is (out of 255 max for demonstration):
R/G/B

100/212/52 89/222/48
94/210/201 110/40/38

If the photosite filter pattern is :
G R
B G

We'll get:
R/G/B
89/126/201

It shouldn't take long to realize that result isn't actually very accurate to the light received for the green and blue values. But it is the sort of data that can confuse a very simple debayering algorithm. To get around this, people do more complicated math that looks more carefully at other surrounding values.

But to complicate matters even more, people frequently want even more resolution from their bayer camera, so instead of debayering a block of four photosites, they try debayering one RGB pixel per single photosite.
Showing you that math and the edge detection checks that become necessary would be very unwieldy to do here.


Suffice to say, you can upscale the F35 in post, and generally the results should look great, but they will be less accurate and may suffer when their are tight patterns. Bayer sensors have similar problems. Sony and Panavision did something very smart and very expensive, but that is difficult to explain and therefore it doesn't market well when everyone else is willing to scream louder with bigger numbers.

"I have the Gruntmaster 4000 with the RX engine 2D torque!"
"Oh yeah? This is the new Gruntmaster MEGA 8000 with the XRD Mega Engine 4D torque and it has Real Time Flibberflam!"

Which one is better?
Click to expand...

http://www.fdtimes.com/2010/06/07/sony-f35-upgrade/

Film and Digital Times reported the F35 2k upgrade 6/7/10. I trust that they probably got it straight from Sony.
 
Getting back to Alexa (kinda) and F35, there was a comparative series of tests shot at various exposures between an Arri D-21, the F35, and a Red over at cinematography's website.
(Link http://www.cinematography.net/Hannover/Index.htm)
No commentary that I could find, but I've enjoyed downloading the dpx files and grading them.
I've been able to get the F35 footage to match the D-21 plates so well that even with a wipe transition they look close.
The Red - not so much.
I understand these tests aren't exhaustive, but with all that color information built into the F35 pipeline, you would think it would be quite a chameleon at imitating other looks.
Here's a D-21 plate followed by an F35 plate.
D-21_Native.jpgF35_Native.jpg

At first glance, it would appear the F35 has no chance at adopting the soft warm tones of the D-21.
But with a little work, here's the F35 again. Unfortunately, a car parked in front of the dark window outside for the F35 shot.

F35_Matched.jpg

And for a little highlight recovery, and a smidge more warmth for her skin, here's the F35 again.

F35_Matched_Recovered.jpg
The irony to me is that it seems the film DOP crowd don't like looking at cameras as just one step in a long process, but as something more self contained - where they want to lock in a look from the camera itself. Is this true?
The difference between the stills photographer mindset and motion picture photographer mindset has always been apparent, but it puzzles me why people want to lock a decision down, unless they don't trust the people in post.
I started as a stills (film) photographer and custom printer, where every picture was "made" in the lab. Dodging and burning detail into paper wasn't a luxury, it was the only honest way of getting all that detail out of the neg. Ansel Adams once explained that he wouldn't give up on a negative until he'd spent at least 40 hours working in the darkroom with it. So I find it strange that a camera's inherent look would dominate a DOP's choice - when the camera's capacity to be stretched into the widest varieties of looks would offer more opportunity.

Just thought I'd share this little experiment.
 
Sandbox said:
The irony to me is that it seems the film DOP crowd don't like looking at cameras as just one step in a long process, but as something more self contained - where they want to lock in a look from the camera itself. Is this true?
The difference between the stills photographer mindset and motion picture photographer mindset has always been apparent, but it puzzles me why people want to lock a decision down, unless they don't trust the people in post.
I started as a stills (film) photographer and custom printer, where every picture was "made" in the lab. Dodging and burning detail into paper wasn't a luxury, it was the only honest way of getting all that detail out of the neg. Ansel Adams once explained that he wouldn't give up on a negative until he'd spent at least 40 hours working in the darkroom with it. So I find it strange that a camera's inherent look would dominate a DOP's choice - when the camera's capacity to be stretched into the widest varieties of looks would offer more opportunity.
Click to expand...

The traditional still photographer both did the field work himself/herself and also did the darkroom processing themself. A modern still photographer is likely to both do the field work and do the digital processing on the image themself. By contrast, Directors of Photography for motion picture virtually never do the editing or color timing of their footage in post on small projects, let alone big projects.

So yes, when a single individual is responisble for both field and post work, they're likely to treat it all as one process, with every decision contributing to the whole, and with no urgency to make a decision in the field that can just as easily be made later.

However, when a creative craftsman (or craftswoman) is only responsible for one stage of the process, and they have a very specific vision, they may very well take steps to prevent other from later undoing what they see as their hard work and vision. Because the editor and the color timer are also creative people, who are likely to see cool ideas and ways of telling the story, and it might not be the same coherent vision the DP when they shot the footage. And that's not even to address the fact that directors, producers, and even studio folk sometimes diegn to tell the people in post to make changes at their whim.

So, it's not exactly uncommon for a DP to wrap production, have a drink at the wrap party, come to the screening two years later, and the film doesn't resemble what they thought they had worked on for six weeks.

And, best as I can tell, that's why you see a mentality in DPs to lock looks in with the camera. Because they see the story a certain way, and they don't want anyone down the pipeline screwing with their vision.
 
Makes sense, Viddo.
Then, when film-centric DOPs compare digital cameras, do you think they sit with colorists as part of the process, or do they remain isolated? Probably too general a question, but when I read articles where a DOP chose to shoot D-21 instead of F35, I have to wonder why, when I can see the F35 would be able to match it - while also offering other options.
I know I'm getting into the weeds, but is it safe to presume that the F35 would be more flexible at imitating various looks than the current front-running cameras do? 12-bit 444 vs RAW off a chip that is inherently a 422 sensor?
 
I would go with an 4:3 capable Alexa given the chance, but there are many projects I would choose the F35 over the Alexa as well- I just shot using a bunch of strobe lights and the F35 was magic because of the CCD, I would have to have used an Alexa with rotary shutter otherwise and that camera is WAAAYYY heavy for handheld.
 
You must log in or register to reply here.
Back
Top