Michael Cioni prepping for a 4K world

[PatrickFaith]

... In images and cameras this is rarely the case, but we do try. The optical low pass filter (OLPF) tries to limit the incoming signal so that aliases never occur. But OLPFs are quite "soft" in that as well as reducing the frequencies that cause aliasing, they soften the lower frequencies we wish to keep as well. Getting a balance between maintaining over-all MTF and reduction of aliasing to a point where it's rarely, if-ever provoked in the real world is a tricky balancing act.

Graeme

To me the benefit of 4k is it gives me the sense of "touch", in that I can see textures and then my minds implies "touch". This gives a whole other "sense" to the movie experience. My problem so far, is textures, like the repeating thread patterns in a dress, need to be filtered out (either by lens OLPF, or sometimes by "softening" filters) because of the spatial frequencing sampling errors. Also where i'm doing 2.5D calculations on textures where i'm doing spatial tracking on the pixels between frames, the lens OLPF is no where low enough shooting at 4k with the lenses i work with(and it's a low end glass, high end is even worse). This also gives tiny artifacts, where people say they "don't" like the extra resolution - my feel is it's not the extra resolution of 4k they find objectionable, but the sampling errors. I think as we move to full 4k projection, balancing OLPF to sensor resolution will be something the DP will have to think about - if they want that feeling of touch in their movie. Also btw, i think "touch" is a stronger sense then vision (at least equal) ... so it can have a major impact on the movie experience.

 

[Meryem Ersoz]

Michael has a way of saying the most provocative things without them coming across as provocative at all. It is truly a gift. I don't know many folks who can talk about 4K as "the beginning of the collapse of broadcast television" and have it come across as a thrilling new development, rather than a threat to the status quo.

Things I want to hear more about....

1) 4K is the new 3D...very intrigued by this statement. I've been thinking that since first witnessing 4K projection at the first Red Studios event. I can't wait until NAB to see how the bright minds, who have been imagining this future, are now shaping it. I truly believe this NAB is going to rock the world by unveiling many of the possibilities which haven't yet been realized for these 4K+ cameras we have all been shooting. We've been living in a world where we are shoe-horning 4K into lesser systems - what will the delivery world look like when we have accessible and affordable delivery systems which actually can mirror our acquisition systems?

2) death of standardized aspect ratios - this is an intriguing point because it does open up quite a few creative possibilities and opportunities to re-imagine cinematic delivery. I can completely imagine a RAW file-based REDRay system where aspect ratio can be completely customizable and easily re-jiggered to fit delivery, without having to wrestle with multiple output formats or media.

Really cool stuff - I feel like this presentation was something of a sneak peek at REDRay without ever saying the word, because it's clear that the think tank is busy re-imagining new delivery models based on the feature sets of the EPIC/R1/SCARLET acquisition systems.

Very exciting....can't wait for the veils to be lifted.

 

[M Most]

Totally understand that it was the filmmakers choice to essentially only end up with a 3k worth of resolution in a 4k image...

That's not what they did. Cropping does not affect resolution. Resolution is usually measured in line pairs per millimeter. That has nothing to do with the overall size of the image, it has to do with the information contained in it. So if you have a 5K image that can reproduce close to 4K resolution, cropping it won't change that parameter. You'll simply have a smaller image with the same resolution. This also explains why it is always preferable to either crop or pad rather than scale to fit images into standard containers, such as the 2048x1080 container of a DCP - you don't want to scale to 1998x1080 for a 1.85:1 presentation, you crop to that to avoid resolution changes due to scaling. Similarly, if you're making a DCP out of material that is 1920x1080, you pad it rather than scale it to fill out the required container size.

 

[Graeme Nattress]

llp/mm would be a relative measurement of resolution where as a pixel dimension such as 5k would be an absolute measure. To say that cropping doesn't effect resolution is to say it doesn't effect relative resolution (because you're reducing both line pairs and mm when you crop) but it does reduce absolute resolution. It also reduces resolution if you consider projected size of image being a constant because then you've kept mm constant and reduced lp.

The reason you don't scale 2048x1080 to 1998x1080 is because such small reductions in scale can produce easily visible artifacts from the downsampling filters and that cropping produces a cleaner image.

Graeme

 

[M Most]

llp/mm would be a relative measurement of resolution where as a pixel dimension such as 5k would be an absolute measure. To say that cropping doesn't effect resolution is to say it doesn't effect relative resolution (because you're reducing both line pairs and mm when you crop) but it does reduce absolute resolution.

True, but my comment was specifically directed at Brad's comment regarding "3K resolution" from a 5K original due to cropping, which is to me a comment on relative resolution and NOT absolute resolution, since the ultimate delivered frame was intended to be 4K in the first place. It sounded to me like Brad was thinking that reducing the image size would affect the relative resolution, and in point of fact, it would not.

As fr the scaling artifacts, it's kind of a matter of semantics. I know you understand these things, and (at least I hope) you know I understand them. We're both "kind of" saying the same thing, which is that cropping does not directly affect the information that's retained because there's no real processing being done, whereas any operation involving a transform - be it a color transform or a scaling transform - will affect the remaining information very directly.

 

[Graeme Nattress]

Oh, I follow you Mike, but we have to be quite careful with our explanations as we're not in a fully interactive environment where we can agree easily on terminology and assess comprehension, or lack thereof as people's eyes glaze over....

Graeme

 

[Ian Vertovec]

As long as Michael is looking in on this discussion: I'd like to know how both Pablo rooms were used simultaneously on the project. I assume each room mainly did every other reel, which makes sense on a lengthy project that had to get done in a set amount of time. I'm also surprised nobody has made much of a fuss of the fact that they opted not to use R3D files and just debayer everything out to uncompressed 4K/5K DPX log files and go from there.

Did the client ask you to pull along a second simultaneous timeline of the original dailies color decisions, just as a reference? I'm curious how and when that was used in the workflow, in terms of the final color correction. I'm also curious when you chose to color-correct in 2K vs. 4K on the available projectors, and how much of an effect that had on the decision-making process.

During color, one room was always coloring one reel, while the other was updating vfx for the other 8. Essentially every shot was a vfx due to the volume of stabs and splits. When we did our final output, we used both systems to export simultaneously.

I feel like we've discussed raw vs. rgb grading quite exhaustedly here on reduser. I think most people agree that raw just isn't an option for jobs that involve vfx and multiple venders. Besides, if you create a properly balanced REDlogfilm there really aren't any drawbacks. AND you still have the option to go back to the raw on a case by case basis.

For every job we load an offline reference as a guide. I want to see what the director has been staring at for several weeks. I usually don't look at it before I start working however. In Walter Murch's "In the Blink of an Eye" he mentions how he doesn't like to visit set ever. He felt that experience would containinate his perception of the footage. The need to use a shot because of the difficulty in production, etc. Now if a filmmaker asks me to match the offline, I will do that. However, that is still drawing a box around the creative dialog in some ways.

As far as projectors, we used the 2K Christe for the color correction, and the 4K Barco to QC the 4K DCPs. I'm not happy with the black level on ANY 4K projector at this time. So since we really wanted the shadows to be right we used the 2k. In the future I hope to be working with 4k pictures at all times (even 2K files look stunning on the 4K projector)

 

[Brad Allen]

True, but my comment was specifically directed at Brad's comment regarding "3K resolution" from a 5K original due to cropping, which is to me a comment on relative resolution and NOT absolute resolution, since the ultimate delivered frame was intended to be 4K in the first place. It sounded to me like Brad was thinking that reducing the image size would affect the relative resolution, and in point of fact, it would not.

As fr the scaling artifacts, it's kind of a matter of semantics. I know you understand these things, and (at least I hope) you know I understand them. We're both "kind of" saying the same thing, which is that cropping does not directly affect the information that's retained because there's no real processing being done, whereas any operation involving a transform - be it a color transform or a scaling transform - will affect the remaining information very directly.

Hi Mike - thanks for your comments. I think the confusion is based on my wording in the original post you referred to. I was meaning to refer to absolute resolution but didn't really make this clear.

Essentially what I meant was that by cropping a 5k image to 4k - say 4096x2160 pixels - we will only end up with about 3.2-3.5k worth of resolution due to the OLPF and debayer process also affecting that original footage. Thus the actual measured resolution is less than 4k even though there is 4k worth of pixels on screen.

On the other hand. If the full 5k frame were used but downscaled into that 4k frame (4096x2160) then we would come much closer to hitting a measured resolution of 4k (assuming a good downscaling algorithm is used).

I think talking about relative resolution would be confusing to most Redusers here. Otherwise we would say that a 5k sensor and a 1k crop of a 5k sensor have the same resolution. In a relative sense they do - but only if no scaling is going on to bring them both back to an absolute measurement of pixels.

 

[Maxi Claudio]

Cioni, referred in the video of converting footage to 720p h.264 and using itunes "podcast" to distribute footage to people who need to see it. Anyone know how this is done? Can you put up many clips and have it private only to the people you want to see it? It sounds like a nice way to send quick footage to people to watch on an ipad.

 

[Martin Stevens]

Cioni, referred in the video of converting footage to 720p h.264 and using itunes "podcast" to distribute footage to people who need to see it. Anyone know how this is done? Can you put up many clips and have it private only to the people you want to see it? It sounds like a nice way to send quick footage to people to watch on an ipad.

I thought they used the Pix System for this on TGWDT.

See..... http://www.pixsystem.com/index.php

 

[Maxi Claudio]

actually, I didn't mean how they trans-code to 720p. But how are they uploading the H.264 through itunes as a podcast, and limit who can access the footage? I haven't found any info yet on google or youtube.

 

[Marc Wielage]

I feel like we've discussed raw vs. rgb grading quite exhaustedly here on reduser. I think most people agree that raw just isn't an option for jobs that involve vfx and multiple venders. Besides, if you create a properly balanced REDlogfilm there really aren't any drawbacks. AND you still have the option to go back to the raw on a case by case basis.

Very helpful, Ian. I appreciate the feedback. Congrats on a good job.

As far as projectors, we used the 2K Christie for the color correction, and the 4K Barco to QC the 4K DCPs. I'm not happy with the black level on ANY 4K projector at this time.

The caveat there is: what are we seeing in theaters if they're using a 4K projector? Are the Sony projectors any better?

Now that I think about it - I think I knew that this happened. It just seems bizarre in my way of thinking. If it was my choice I think I would be aiming to have my cinema release as the standard against which all others are measured. Max resolution, best colours, least compression, best sound and optimum framing.

As long as the DP and director approve it for the specific venue, it doesn't matter. Just going from a cinema-sized display to a consumer display is a compromise; it's just a question of whether it works artistically and technically.

 

[Jim Woo]

Here it is:

http://www.youtube.com/watch?v=Qztrb9arZl4

 

[Hrvoje Simic]

Well presented.

Two suggestions:

1) Although resolution benefit mentioned does provide an additional re-framing benefit, suggestion of general re-framing ability can lead to misconceptions for less experienced. For example, re-framing to the extent of using two different shots from one take is more as a "salvage" feature than a creativity feature and a lot of DoPs hate it because it messes up their image, creating two shots which weren't composed or lighted for.

It may look as an interesting and cool creative feature to more post-centric logic, or a director missing a shot, but simple "zoom-in" doesn't qualify as a substitute for a different lens and coloring doesn't substitute different lighting. One thing is a well experienced director or a DoP in a specific and prepared situation, but using this in general as a just another creative feature just because it can be done can easily lead to young or inexperienced DoPs and directors getting sloppy, and relying on the post guy to fill in the gaps. Which doesn't lead to increase, but decrease of quality.

2) Although 10bit DPX has been considered as "good enough", in case of 12 and 16 bit RAW master it is a loss of "thickness" of the digital negative. Both aesthetically and technically speaking. A significant loss of thickness. With less precise creative options and less room for creative work. And salvaging of highs and lows.
Seeing 4K/5K raw as a transformation to a completely new concept, requires skipping the 10 bit middle part. IMO. If the change is happening by increasing spacial resolution which image acquisition allows, it makes sense to utilize the full color depth, the tonal resolution as well. When possible.

 

[Roberto Lequeux]

Well presented.

Two suggestions:

1) Although resolution benefit mentioned does provide an additional re-framing benefit, suggestion of general re-framing ability can lead to misconceptions for less experienced. For example, re-framing to the extent of using two different shots from one take is more as a "salvage" feature than a creativity feature and a lot of DoPs hate it because it messes up their image, creating two shots which weren't composed or lighted for.

It may look as an interesting and cool creative feature to more post-centric logic, or a director missing a shot, but simple "zoom-in" doesn't qualify as a substitute for a different lens and coloring doesn't substitute different lighting. One thing is a well experienced director or a DoP in a specific and prepared situation, but using this in general as a just another creative feature just because it can be done, can easily lead to young or inexperienced DoP's and directors getting sloppy and relying on the post guy to fill in the gaps. Which doesn't lead to increase, but decrease of quality.

2) Although 10 DPX has been considered as "good enough", in case of 12 and 16 bit RAW master it is a loss of "thickness" of the digital negative. Both aesthetically and technically speaking. A significant loss of thickness. With less precise creative options and less room for creative work. And salvaging of highs and lows.
Seeing 4K/5K raw as a transformation to a completely new concept, requires skipping the 10 bit middle part. IMO. If the change is happening by increasing spacial resolution which image acquisition allows, it makes sense to utilize the full color depth as well. Where possible.

I am down with keeping two more bits, hopefully all the way to the screen.

 

[Marc Wielage]

I am down with keeping two more bits, hopefully all the way to the screen.

How much more time and money would you be willing to spend to do this, particularly with a feature project?

Talk to your post house, and see if they can deal with 12-bit or 16-bit images and what this will cost you. I've been around post a long time, and I can't recall ever seeing a display device that could do 12 bits, let alone 16. Even an honest 10-bit display is hard to come by, and I'm fairly certain that 99.9% of the digital cinemas in the world right now are maxed out at 10 bits.

Although resolution benefit mentioned does provide an additional re-framing benefit, suggestion of general re-framing ability can lead to misconceptions for less experienced. For example, re-framing to the extent of using two different shots from one take is more as a "salvage" feature than a creativity feature and a lot of DoPs hate it because it messes up their image, creating two shots which weren't composed or lighted for.

I think Michael was pretty specific in saying that the extra framing was generally used for minor headroom issues and small push-ins. Nobody's expecting to shoot one shot and get a wide shot, a medium shot, and a close-up from the same frame. But having the ability to tilt up another 1/2" can mean all the difference in the world, particularly with handheld material or minor framing problems.

 

[Will Keir]

Did the Jan 27th interview come online yet?

Here it is:

http://www.youtube.com/watch?v=Qztrb9arZl4