Red vs Film

[Wesley Scoggins]

Except for the argument that image quality and latitude is better on film. Oh, and the workflow is better established.

Same for many black and white stocks when compared to Color less than a century ago.

At this moment in time, film still has advantages, but they can't last forever.

While the image quality of a raw stock may be excellent, film has a far worse generational degradation than digital, and many other limitations.

Regardless, workflow kinks are quickly being sorted out and the industry is quickly adjusting to meet the needs of digital cinema cameras. For how far the workflow has come in just a decade (compared to film having a full century of refinement to it's workflow), I think that workflow issues is a moot point with how short of a time it's been around and how far it's come already.

Keep in mind that film stocks are changing and evolving, too.

Well film is reaching the limit of what it can accomplish. Stocks ARE still evolving, but film has largely reached it's limit of what it can accomplish, while digital has only just started developing as a real cinema tool.

Digital technology has accomplished what has taken film development over a century to get to. There are a realistic limit to how big film can be made and still be economical, we won't be seeing anyone using anything bigger than IMAX on a film anytime soon.

While digitally we're less than a year from 28K.

Film stock is evolving, but it took Kodak 5 years to get from Vision2 to Vision3, for only a minimum visual advantage (it's advancement true, but slight).

It took Red 2 years to get from Prototype to commercial product, and it's taken them another 3 years to get from 4K to 28K.

Look at how far we've come in 10 years in digital cinema, and think of how far we'll be in another 10 years (heck, another 2, we're far closer to having more efficient workflows very very soon.)

[Gavin Greenwalt]

Except for the argument that image quality and latitude is better on film. Oh, and the workflow is better established.

Keep in mind that film stocks are changing and evolving, too.

Digital has its benefits, but for the utmost in image quality and latitude, film is still king.

Two arguments which will only be true for another few years. Digital is improving far far quicker than film. And the ease of workflow with film is only a function of it having come first. It's intrinsically more difficult to work with, just more widely and better understood.

[mike ogden]

I didn't mean you Wesley or Gavin. You guys seem more level headed when referring to 'fanboys' on here :)

But can I point out here that no one has seen 28k, let alone worked with it. Lord knows what sort of super duper computer you're going to need to deal with a 28K image size. Sometimes I think Red runs before they can walk in terms of development.

I sort of wish they'd gone with a 3 chip design for the Red. Would have made a better color workspace from the off and been closer to Cineon specs. But still, people seem to be making good images with it, I guess it can only get better. Wonder why they did go with Beyer technology?

[Graeme Nattress]

Mike, the color space of the RED is just fine. I don't know what Cineon specs you're reading, but the ones I've read don't define any form of colorimetric color space. The specs deal with mapping negative film densities to code values - an encoding curve, and that is it. There's no implied colorimetry, and no reference back to actual linear light data either.

3 Chip designs are practically useless for larger chip sizes (notice lack of 3 chip cameras beyond 2/3" sensors), and vastly complicate the lens designs. I don't think it's quite "you can't make one bigger than 2/3"", but that for all practical reasons you wouldn't want to.

Why Bayer - it's been answered before - it offers, in a single chip design, the best optimization between photosite count and perceived resolution / image quality. With how we work it with a wavelet codec, there are other advantages to in terms of image scalability, which is what will make 28K resolution still useful - you'll be able to easily deal with it at lower resolution extractions, only going up to the resolution you need when you actually need to.

Graeme

[Stefan Christou]

Mike, the color space of the RED is just fine. I don't know what Cineon specs you're reading, but the ones I've read don't define any form of colorimetric color space. The specs deal with mapping negative film densities to code values - an encoding curve, and that is it. There's no implied colorimetry, and no reference back to actual linear light data either.

3 Chip designs are practically useless for larger chip sizes (notice lack of 3 chip cameras beyond 2/3" sensors), and vastly complicate the lens designs. I don't think it's quite "you can't make one bigger than 2/3"", but that for all practical reasons you wouldn't want to.

Why Bayer - it's been answered before - it offers, in a single chip design, the best optimization between photosite count and perceived resolution / image quality. With how we work it with a wavelet codec, there are other advantages to in terms of image scalability, which is what will make 28K resolution still useful - you'll be able to easily deal with it at lower resolution extractions, only going up to the resolution you need when you actually need to.

Graeme

Wise words from Graeme.

I like what Ridley Scott says:

they are two very different mediums, and digital really must be treated as a new, not replacement artform...

both are suitable for storytelling... the latter is all that matters.

And that was where the discussion should have been put to bed

[mike ogden]

Are you an admin, Stefan? I didn't realise...

[David Battistella]

WHen this discussion comes up I am always very surprised at how easily biated everyone is into pushing an opinion on something so subjective.

Why not call the thread:

RED vs. SUPER 8
RED vs. 16mm
RED vs. SUPER 16mm
RED vs. 35mm
RED vs. SUPER 35mm
RED vs. 70mm
RED vs. IMAX

Each one of the film formats laid out are their own specific type of film with their own characteristics.

RED is a digital medium with it's own personality and style.

It's not a competition.

David

[Steve Sherrick]

This is very subjective and has certainly been covered many times, in many forums usually going around in circles because it's a question without a definitive answer.

But sometimes what comes out of it are more questions, ones that are perhaps more important. I find the topic of archiving digital media a far more fascinating topic because it's something that is tangible and could certainly benefit from many intelligent perspectives so as to find solutions that are effective and offer filmmakers confidence that their work will be easily retrievable years down the line. It's these topics that truly have an impact on us.

The film vs digital is a project by project question, one that will evolve over time. If the topic is to really hold merit, it seems like a far more interesting discussion would revolve around tests that people may want to share between say a particular film stock and the RED. Perhaps they might want to discuss how they conducted their tests, why they decided to go with one over the other, and then receive feedback. I like these kind of discussions over general blanket statements and opinions, especially when film and digital each encompass many different flavors within their own realm.

That's just me.

[David Battistella]

Great points Steve,

Storing Data in the long term is going to be very interesting for many of us.

David

[Wesley Scoggins]

I didn't mean you Wesley or Gavin. You guys seem more level headed when referring to 'fanboys' on here :)

But can I point out here that no one has seen 28k, let alone worked with it. Lord knows what sort of super duper computer you're going to need to deal with a 28K image size. Sometimes I think Red runs before they can walk in terms of development.

I sort of wish they'd gone with a 3 chip design for the Red. Would have made a better color workspace from the off and been closer to Cineon specs. But still, people seem to be making good images with it, I guess it can only get better. Wonder why they did go with Beyer technology?

Thanks. But from what I understand in regards to 3-Chip is there are a few issues with it that makes it unviable for 4K image capture.

First very high-quality prisms are VERY expensive and hard to make, and there is a limit to how good the prism can economoically be beyond 2/3" (As Graeme pointed out), also you can't use regular cine lenses on 3 chip designs, you have to have either ones with a special element designed to compensate for the prism or you have to have a special adapter between them.

Then there is the issues of blooming and stuff in regards to prisms, especially ones that big.

I recall reading that if Red has gone with a 4K 3CCD design it would have bloomed the camera price by more than 10 times (since I don't believe there are any 4K 3CCD sensors available, so Red would have to have spent tons of money developing on), for little to no real increase in image quality (Since the limitations of the prism would be hampering the capabilities of the sensors.)

CMOS Bayer chip means it's cheaper, uses less electricity (thus longer shooting times), can use standard cinema lenses, and doesn't suffer from as many defects as huge 3CCD's (all the blooming and streaking and light loss you get from prisms that big).

And true, no one HAS worked with 28K yet, but I assume that Red will have developed a viable option for working with 28K footage (and being able to downscaled it to a 4-8K master for editing) before the brain is released.

The dynamic range and clarity of 28K footage sampled down to 4-8K will be HUGE, I forgot the exact conversion scale (I believe it was David Mullen or Graeme who talked about this in a thread recently) but when downscaling you can see an increase in the dynamic range and a decrease in noise.

Someone talked about if you're downscaling to SD or YouTube size 4K footage can jump from 10-11 stops of DR to like 12-13. So if you're downscaling 28K to 4-8K for a 35mm print, and the Dynamic range of the new sensors is already higher than the Mysterium...

I don't know how much that is, but it sounds like it could definitely beat the dynamic range of Vision 3, or most any other 35mm stocks.