DRAGON size comparison - some help needed

[Neil W. Smith]

6k FF is 6144x3160, 6k 2:1 is 6144x3072

Duh! ... thanks John.

Neil

 

[Neil W. Smith]

the economics of camera choice ...

the economics of camera choice ...

Why do producer's concern themselves with numbers and technical specs? And why is it necessary for a cameraman to be able to explain the math that compares the dragon sensor to other sensor sizes? I guess I'm from the school of thought that all that really matters to me about a camera is the image it creates and the way it operates. Obviously it works for many camermen to be very technically proficient but I'm wondering why a producer would care.

Fact of life, David ... it happens on many level of project from Indie feature to Burbank TV Show ... we're one of the sponsors of the 'Emerging Cinematographer Awards' (ECA) which is organized by IATSE/ICG (also known as Local 600 - the camera guild) ... at the sponsor's lunch yesterday which took place in the lovely ASC (American Society of Cinematography) clubhouse in Hollywood, Steven Poster, ASC, the President of IATSE/ICG made that very point - Producers are making camera choices based on all kinds of criteria and not leaving the decision in the capable hands of the one person who knows what's best for the story - the DP.

So being able to explain both the technical and financial benefits of shooting with the DRAGON is something that many DPs and camera guys are going to have to learn .... just showing a Producer an image from one camera versus another is not going to work.

Your question gets to the heart of why I started this thread ... even explaining what 6K is and why its important can be a challenge with certain Producers (these guys are typically not very tech savvy) and that's before you start on the financial discussion around why Camera A is better than Camera B even though it costs more to buy or rent.

Neil

 

[Phil Holland]

From the concept of clear communication to producers, executives, and studios in general there's a few ways I handle it. And much of it is similar to the way I go about discussing 4K, UHD, and the past/future of cinema.

Most of these discussions end up in me showing the visual benefits and using sort of real world examples and analogies. Many here have heard me mention the "window effect" or discussing motion resolution and relating it to print media at various viewing distances and display sizes. These concepts aren't very difficult to grasp and more importantly they aren't that difficult to demonstrate. And the general concept of 9x the resolution of 1080p are nice lead in selling points if you are looking to impress.

Now from a budgetary perspective there's lots and lots of different types of productions. A single cam 3 day commercial shoot is a very different beast compared to say a 60 day 3 cam feature. Also workflow and pipelines must be considered. I tend to work on VFX heavy projects and there are certainly things to consider and gain from having that extra resolution to work with. Color workflow is another thing to look at as there's several ways to go about it. Also, how a project is being edited and delivered comes into play. I see folks still using rather archaic methods on the editing front and that's something that should be closely examined. If spending $??,???? on a new system could potentially save you $???,??? in labor/post hours that's something that must be examined closely and not simply ignored because somebody is using 5 year old software.

As David mentioned in regards to Epic Dragon's 6K, 5K, 4K and formats found within are all options for 4K productions. From my experience over the last year, mainly after discovering the actual 6K resolution details back in August 2012 and knowing exactly how the Red system works in regards to the WS and HD standard I predict we'll see a good amount of productions exploring 6K WS and 6K HD workflows for cinema and broadcast delivery as well. This has been much of the "how do we get ready for this" talk that's been going on from my localized perspective (which isn't that local really). Lots of things have changed in that year with things like the Red Rocket-X being announced and suddenly having GPU acceleration. This effects workflows literally right now.

There's some other interesting concepts like debayer mathematics and the gains netted from super sampling that are bit harder to explain, but again, one merely just needs to reference the old school print resolution concepts that we've been using forever now. Often that can be summed up by "details really shine".

One distinct advantage that Red has over say other systems is indeed REDCODE RAW and how much media space that takes up versus some other similar raw formats. And it's also something you can control. We know the standard production practices with Mysterium-X. 5:1 for features, 8:1 for nearly anything else, 3:1 for high frequency detail, 4:1 for anamorphic, and higher compression ratios have been used for 2K delivery as well as over cranked material. Again, that's something we can control. Dragon's going to be a bit different, but likely in the same general range, although we might be able to get away with higher compression ratios yielding higher quality results due to the lower noise floor. Testing is needed and will be happening really, really soon. My early guess from my initial observations is we'll have more freedom in general on this front to produce great looking 4K deliveries.


The above might be too much for some to absorb. However, visual quality is something that all should understand clearly. Also camera choice by a DP shouldn't be ignored and I feel for those who are being dictated to use other systems if they prefer the look and feature set of another system. I don't typically have that problem as an owner I can literally just shoot a test and show them exactly what's possible. I understand that there's different budgets for many outlets, but disk space honestly isn't the worst thing in the world. Man hours are the real thing that cost productions cash. If money needs to be saved they need to really investigate streamlining or optimizing workflows rather than limiting the quality of the material for any reason.

I just consulted on a $10k short (25 minutes) and they seem to feel a 2 Epic, full support, lenses, lighting, 12 day shoot is a reality with cast and crew. That's gonna be a tough one and that's thinking big and pulling favors for sure as they didn't calculate any costs that come with post production at all. They aren't offering any differed payment for anybody either. But I guess it really depends on how that money is spent and what people are willing to work for and what the people on the top of the project are willing to expect from the final results.

Much of this is the reason I own my own gear these days and do all that I do. Productions I produce tend to get a good deal on my equipment and abilities, which allows me personally to do more with less. And how this particularly relates to productions I get hired to work on is they have the advantage of having somebody who can see a production through from pre-post. That's more of my personal model though. Some solutions benefit from throwing more bodies and money into the fire, some benefit by having a multi-tasker executing at a high level. Just depends. Heh. This last paragraph is a bit off subject I guess, but it's something to observe as it technically relates.

 

[Alain M]

From the concept of clear communication to producers, executives, and studios in general there's a few ways I handle it. And much of it is similar to the way I go about discussing 4K, UHD, and the past/future of cinema.

Most of these discussions end up in me showing the visual benefits and using sort of real world examples and analogies. Many here have heard me mention the "window effect" or discussing motion resolution and relating it to print media at various viewing distances and display sizes. These concepts aren't very difficult to grasp and more importantly they aren't that difficult to demonstrate. And the general concept of 9x the resolution of 1080p are nice lead in selling points if you are looking to impress.

Now from a budgetary perspective there's lots and lots of different types of productions. A single cam 3 day commercial shoot is a very different beast compared to say a 60 day 3 cam feature. Also workflow and pipelines must be considered. I tend to work on VFX heavy projects and there are certainly things to consider and gain from having that extra resolution to work with. Color workflow is another thing to look at as there's several ways to go about it. Also, how a project is being edited and delivered comes into play. I see folks still using rather archaic methods on the editing front and that's something that should be closely examined. If spending $??,???? on a new system could potentially save you $???,??? in labor/post hours that's something that must be examined closely and not simply ignored because somebody is using 5 year old software.

As David mentioned in regards to Epic Dragon's 6K, 5K, 4K and formats found within are all options for 4K productions. From my experience over the last year, mainly after discovering the actual 6K resolution details back in August 2012 and knowing exactly how the Red system works in regards to the WS and HD standard I predict we'll see a good amount of productions exploring 6K WS and 6K HD workflows for cinema and broadcast delivery as well. This has been much of the "how do we get ready for this" talk that's been going on from my localized perspective (which isn't that local really). Lots of things have changed in that year with things like the Red Rocket-X being announced and suddenly having GPU acceleration. This effects workflows literally right now.

There's some other interesting concepts like debayer mathematics and the gains netted from super sampling that are bit harder to explain, but again, one merely just needs to reference the old school print resolution concepts that we've been using forever now. Often that can be summed up by "details really shine".

One distinct advantage that Red has over say other systems is indeed REDCODE RAW and how much media space that takes up versus some other similar raw formats. And it's also something you can control. We know the standard production practices with Mysterium-X. 5:1 for features, 8:1 for nearly anything else, 3:1 for high frequency detail, 4:1 for anamorphic, and higher compression ratios have been used for 2K delivery as well as over cranked material. Again, that's something we can control. Dragon's going to be a bit different, but likely in the same general range, although we might be able to get away with higher compression ratios yielding higher quality results due to the lower noise floor. Testing is needed and will be happening really, really soon. My early guess from my initial observations is we'll have more freedom in general on this front to produce great looking 4K deliveries.


The above might be too much for some to absorb. However, visual quality is something that all should understand clearly. Also camera choice by a DP shouldn't be ignored and I feel for those who are being dictated to use other systems if they prefer the look and feature set of another system. I don't typically have that problem as an owner I can literally just shoot a test and show them exactly what's possible. I understand that there's different budgets for many outlets, but disk space honestly isn't the worst thing in the world. Man hours are the real thing that cost productions cash. If money needs to be saved they need to really investigate streamlining or optimizing workflows rather than limiting the quality of the material for any reason.

I just consulted on a $10k short (25 minutes) and they seem to feel a 2 Epic, full support, lenses, lighting, 12 day shoot is a reality with cast and crew. That's gonna be a tough one and that's thinking big and pulling favors for sure as they didn't calculate any costs that come with post production at all. They aren't offering any differed payment for anybody either. But I guess it really depends on how that money is spent and what people are willing to work for and what the people on the top of the project are willing to expect from the final results.

Much of this is the reason I own my own gear these days and do all that I do. Productions I produce tend to get a good deal on my equipment and abilities, which allows me personally to do more with less. And how this particularly relates to productions I get hired to work on is they have the advantage of having somebody who can see a production through from pre-post. That's more of my personal model though. Some solutions benefit from throwing more bodies and money into the fire, some benefit by having a multi-tasker executing at a high level. Just depends. Heh. This last paragraph is a bit off subject I guess, but it's something to observe as it technically relates.

I truly believe that this kind of info about sharing different workflows truly helps us to learn and open our minds to the approaches taken by great filmmakers from this community. I particularly have an approach and I feel we can all learn from eachother if we go info detailed workflows and post them here. Real life learning from rather than going nuts on numbers only.

My 2 cents.

 

[Neil W. Smith]

Storage requirements ...

Storage requirements ...

Many thanks for everyone's feedback and suggestions ... very useful thread ... got some great ideas out of it.

Re: disk storage requirements - Phil and David made a suggestion about having a chart for the amount of storage that different .r3d compression levels produce .... anyone have one handy or care to create one that shows something along the lines of disk storage used for 1 x hour of 4K, 5K and 6K .r3ds at each of the various compression levels covering:

6K RAW (2:1, 2.4:1)
5K RAW (Full Frame, 2:1, 2.4:1 and Anamorphic 2:1)
4.5K RAW (2.4:1)
4K RAW (16:9, HD, 2:1 and Anamorphic 2:1)

with frame rates at 23.98, 29.97, 59.94 fps

One of the real benefits of the RED ecosystem has always been their 'secret sauce' REDCODE RAW compression algorithm .... the visually lossless compression that RED uses allows filmmakers to shoot RAW on camera straight onto the REDMAGS without having to be tethered to an exterior recording system - kinda amazing that you can do that at 6K 19 Mpixels per frame resolution.

The combination of 6K resolution plus the wide DR of the DRAGON sensor plus the new color fidelity is a stunning technical achievement - we just need to make sure that Producers understand how affordable it all is.

Cheers,
Neil

 

[John Marchant]

Not exactly what you asked for but related- note that its for 25fps and 16:9 as this was made for my own ref and that's my most common target.

 

[Phil Holland]

Cool chart John. Question, I notice you're using color coding. Is this more or less to indicate things like "about 1 hour or 1.5 hours" worth of recording time? Or is it something more REDCODE related.

I have yet to do any hard core REDCODE comparisons yet and curious how you feel we'll be using the compression schemes with Dragon too. I.E. w/ Mysterium-X 5:1 seemed to be the sweet spot for 4K productions under most scenarios. Where do we find ourselves now with the new lower noise floor on Dragon?

 

[John Marchant]

Yeah, it's just to signify the 'roughly 60/90 min' point at the different settings.

i've yet to get to the bottom of Dragon redcode, so I'm being conservative right now... Plenty of tests to come

low light still hurts the codec IMHO, but well lit shots... 10:1 at 6k looks good on initial inspection

 

[Phil Holland]

Cool. Good to know.

 

[Neil W. Smith]

Great chart ...

Great chart ...

Not exactly what you asked for but related- note that its for 25fps and 16:9 as this was made for my own ref and that's my most common target.

Most excellent work, Sir ... as an ex Brit I can certainly appreciate the power of PAL ... your chart get's us in the ballpark for these damn Yankees!

Maybe between Phil and I we can come up with a NTSC version.

Cheers,
Neil

 

[Ketch Rossi]

In our shoot, I did some Portrait shooting at 6k @ 17:1 86fps... Blacks Incredibly NOISELESS!!

 

[John Marchant]

For what its worth, my methodology was to cycle through the relevant settings with the camera limited to record 100 frames. Made it rather quick to jump through the combinations, and I then just slotted the Redmag into my mac and did a little maths based on the file sizes.

 

[Hrvoje Simic]

So how do we go about running a 'cost/benefit' analysis for DRAGON that makes it easy for Producers to choose the best camera for the job?

Neil

You don't.

Choosing the best camera for the job is up to the cinematographer.
His Craft. His tool. His aesthetic choice.

Not the post guys'or the producer's.

 

[Ketch Rossi]

For what its worth, my methodology was to cycle through the relevant settings with the camera limited to record 100 frames. Made it rather quick to jump through the combinations, and I then just slotted the Redmag into my mac and did a little maths based on the file sizes.

But that is exactly why I love what people like Phil and you take the time to do... All tests and charts very useful to all, were I just can't do that, I just shoot... While taking advantage of your great and patient work...

 

[Neil W. Smith]

You don't.

Choosing the best camera for the job is up to the cinematographer.
His Craft. His tool. His aesthetic choice.

Not the post guys'or the producer's.

You've obviously spent a lot of time in Hollywood, Hrvoje ... thanks for the helpful suggestion ... I'll be sure to pass it on to the Producer's Guild.

Neil

 

[Phil Holland]

Maybe between Phil and I we can come up with a NTSC version.

I have a version of this chart that will go with a bigger data sheet once "all is nailed into the floor". Lots of stuff to add to that. If possible compared to the last estimated recording times published now that we have an understanding of the different flavors/sizes of REDMAGs I may compensate to allow for very accurate recording times. The old way was just to calculate the data rate, however, some of the media is obviously not used. I'm exploring that concept right now.

 

[John Marchant]

I have a nice table on current Dragon performance too- but as you say, that's probably not worthwhile to share currently as things are still changing. More just a handy red for those shooting with ours right now.

 

[egon touste]

Better to have the tools to calculate the difference than posting the % difference here (if you are a cameraman, you should have some basic math, geometry and algebra in your arsenal):

the math for pixel count is the same as calculating the area of a rectangle. multiply the height by the width and that is the total area...or # of pixels. (forgive me if that is obvious)

Here's the formula to calculate a difference in percentage:

First number is A second number is B

percentage difference is (B/A-1)*100

So what is the percentage difference between 60 and 80

percentage difference is (80/60-1)*100= 33.33

This is the same as saying 80 is 33.33% more than 60.

..er, I think you need to check your algebra sir, whats the -1 about? your results are wrong also, I don't think the difference between 60 and 80 is 33% is it? ..actually it might be.. I would go with % difference between two numbers A and B, which is (A/100)*B .. % difference between resolutions would go more like ((W*H)- (W2*H2)) / ((W*H)/100) ..which could be simplified..

 

[Neil W. Smith]

the challenge o 6Kf maths ...

the challenge o 6Kf maths ...

..er, I think you need to check your algebra sir, whats the -1 about? your results are wrong also, I don't think the difference between 60 and 80 is 33% is it? ..actually it might be.. I would go with % difference between two numbers A and B, which is (A/100)*B .. % difference between resolutions would go more like ((W*H)- (W2*H2)) / ((W*H)/100) ..which could be simplified..

This is exactly the reason I started this thread, guys :-( .... we have some of the finest minds on the planet here on RedUser and even then we have trouble with the maths ... think I'm going to stick with Phil's 6K Chart

But seriously, you try explaining all the different permutations to your average Producer and see how far you get.

Anyway, I'm going to combine Phil's Chart with Harry's Table with John's Table to produce a three dimensional matrix which I'll paint on the sides of a Rubik Cube and use that to explain things to non technical producers .. should win lots of friends out there.

Thanks for all the feedback, folks ... now going to get back to finishing Ketch's 'The Red Dress' ... math aside, the DRAGON sure adds up to one hell of a camera which ever way you slice and dice it.

Cheers,
Neil

 

[Lee Saxon]

Wow. In 2013.