Epic and Scarlet release in 2010

Status
Not open for further replies.
What's ironic is that Scarlet 3K's viability is limited in that it's incompatible with the Jim Jannard's oft-repeated "The future is 4K (and above)" mantra. And he doesn't mean just at the upper levels of the mega-budget Hollywood production stratosphere -- he's been emphatic that he means in the home, too. "Rendering obsolescence all but certain" is one way to describe Scarlet 3K. And before anyone gets pissed off, yes, it's true, there's nothing remotely comparable in its price range. But still, resolution-wise, it's the only Red product that doesn't fit in with Red's 4K vision.

This is something that I've often wondered about. The 5K Epic and Scarlet fit perfectly with the vision of a 4K tomorrow. But 3K really leans on the modular aspect, in my opinion, to act as the "Best 2/3" Camera of Today with a Support System for Tomorrow."

Of course, it's a 2/3" sensor which will probably stand on its own right within certain fields (ENG, events, documentary, etc) for years to come. I just don't know what would happen to price or DR if someone decided to make a 5K 2/3" sensor. Boggles the mind a bit.
 
Barry, you have a point. But if you look the history of this business we find more billionaires to fail than to succeed in this industry.

Having money means you have succeded in something in the past. But most of the times it's not 100 successful products in 100 different markets, it's some product in a very small niche. It hardly proves you are competent. Being the carton package king does not mean you have what it takes to be the pizza delivery king.

Having money also has other effects on people. It makes one overconfident. It's easy to exceed your capacity in management and drift beyond logic. You become you own enemy. You can't solve every problem by throwing money at it, and you can't always buy all the talent you need for the job. You can't even set reasonable goals, ego carries you away and you forget that there are forces that are orders of magnitude more powerful. The everyday entepreneur is used to obstacles and limitations, but the billionaire enterpeneur must be very frustrated when problems persist.

Sometimes excessively supply of capital is not the answer to a good product and service. You need to have a background, you need to have knowledge on the subject of your work, you need to know the business, and you need to love what you do. You have to love every single detail about it. Otherwise, the journey of development feels like unnecessary homework.
 
Having money means you have succeded in something in the past. But most of the times it's not 100 successful products in 100 different markets, it's some product in a very small niche. It hardly proves you are competent. Being the carton package king does not mean you have what it takes to be the pizza delivery king.

Having money also has other effects on people. It makes one overconfident. It's easy to exceed your capacity in management and drift beyond logic. You become you own enemy. You can't solve every problem by throwing money at it, and you can't always buy all the talent you need for the job. You can't even set reasonable goals, ego carries you away and you forget that there are forces that are orders of magnitude more powerful. The everyday entepreneur is used to obstacles and limitations, but the billionaire enterpeneur must be very frustrated when problems persist.

Sometimes excessively supply of capital is not the answer to a good product and service. You need to have a background, you need to have knowledge on the subject of your work, you need to know the business, and you need to love what you do. You have to love every single detail about it. Otherwise, the journey of development feels like unnecessary homework.
My point was basically the difficulty of a hard business like this one.

The second aspect, what the history has proven to us. To guess on these waters is not so rare like to be here to try to have the clue for the next lottery extraction.

The remainder you said it all ;-)
 
Seems the main reason some people have written off Scarlet from their list of potential *future* options is sensor size alone.... I just keep thinking about the oscars the last two years.... didn't best cinematography go to films shot on 2/3" sensors the last two years? Obviously everyone wants to continue to move towards larger sensors, but some don't want to do it at the expense of other features and/or performance... and 2/3" remains a very viable and practical format (imo). But, that's kinda the point. Some people cream over large sensors and that alone, and I get to watch a ton of videos of misused/overused shallow DoF shots from people who don't use it as an available creative option to help convey what the shooter intends, but rather b/c they think that's what makes a filmic image. So, to each his own there, I say. From where I stand, sensor size is not at the top of my list. but, that doesn't mean I think people who place it there are somehow wrong or that a larger sensor camera is dead in the water b/c it lacks x,y, or z as a compromise.

I also think about 120 fps for under 5 grand... Framerates are expensive. For people looking for high framerates that is still extremely attractive. For people looking for something that doesn't have the compromises necessary of current DSLR cams, it's still a potentially very attractive option, b/c other than it, there are small chip video cameras as the only current option, and an announced new panny product that uses a DSLR sensor, which may or may not overcome the shortcomings of the DSLR's image (aliasing, etc)

There is plenty of room in the market for a wide array of products and a wide range of price points... We really have no clue what Scarlet will be if/when it hits the market, but one thing is still pretty certain (to me) and that is that there isn't any other product on the market like it... It still stands as a pretty unique product in comparison to what is available and what is known to be forthcoming. So, it may find its own market.

In the under 10k market, we've got our traditional small chip video cams, we've got consumer priced DSLR's, an announced AF-100 (which is also different than any product that has come before considering how hybridized it is between DSLR and Video Camera), and an announced Scarlet.... All very different types of cameras, and I'm sure they will all appeal to different people for different reasons. I just don't get this whole "x camera kills y camera" every time a new camera is announced and what not.... everyone has different needs, and I'm sure that Scarlet (even in a 2/3" format) will remain very attractive to some. The consumer price range folks hoping for a 3k for 3k have DSLR's now anyways and that market has plenty of released and forthcoming options already.

I'm not sitting here holding my breath waiting for some holy grail or getting all emotional about an unreleased camera... But, i still recognize the unique offering that the 2/3 scarlet is likely to be considering in 2+ years, there still isn't really anything spec'd like it... and (if you leave out sensor size) there is a lot to like about what is on paper, and also a lot to like about the performance of it's sister sensor in the S35 MX. So... I'm not sitting on my thumbs waiting for anything, but still seems like a pretty compelling product to me (and, I'm just me).

my 2 pence anyways.... err... maybe 6.

later,
Jason
 
LOL Your 6 pences are always welcome and fair ;-)

Do not forget the most important part of Slumdog Billionaire was still shot on Fuji Eterna 35mm film stock; The Hurt Locker still in film; and the overuse of DOF is in fact true, but with it you also have the 35mm bokeh what doesn't necessarily mean shallow DOF. The quality of bokeh shines out through deep DOF as well (aka 'film texture' or at least part of it).

And Sony has revised the hybrid concept for $2000!
 
Last edited:
Yes, it is true. I've also read something similar. That's why I edited for «the most important part of» and we can't be wrong anyways. LOL :-)
 
if you shoot it right shallow DoF will take care of itself, some people just get too cought up on it
there are much more important things to worry about than just shallow DoF imo
 
True, up to a point. All I'm saying is that zoom lenses could possibly be made smaller. The zoom on the NEX-VG10 is small but slow at the long end. Perhaps there could be some compromise between bulk and greater speed, and with fully electronic control, a constant T-stop could be possible.

The lens on the Sony is a perfect example of the inherent compromises. Small size = slow speed and limited range on the long end. It is also slow on the wide end too. My Angenieux 12 -240 is constant f4.2 designed for a 16mm format camera. It is significantly larger than the lens on the Sony with a front element diameter of 104mm. The equivalent long end magnification for the Sony would be in the 600mm focal length range. For the same f4.2 it would be more than 140mm in diameter.
 
This is something that I've often wondered about. The 5K Epic and Scarlet fit perfectly with the vision of a 4K tomorrow. But 3K really leans on the modular aspect, in my opinion, to act as the "Best 2/3" Camera of Today with a Support System for Tomorrow."

Of course, it's a 2/3" sensor which will probably stand on its own right within certain fields (ENG, events, documentary, etc) for years to come. I just don't know what would happen to price or DR if someone decided to make a 5K 2/3" sensor. Boggles the mind a bit.

3k is probably the practical upper limit for a 2/3" sensor if one also expects to balance sensitivity and dynamic range into the equation. There are damn few lenses that will resolve 3k on a 2/3" sensor and none that could possibly resolve 5k.
One of Scarlet's outstanding features is the ability to shoot for the 2k digital cinema format in both standard 1.85:1 and 2.4:1 cinemascope without the need to upscale from a lower resolution. It will downscale to both formats for superior noise and resolution compared to any native 1080p HD camera. 4k may be the cinema format of the future, but 80% of first gen digital theaters are 2k equipped and they will be around for at least 10 years. The projectors cost too much to not milk them for every hour of useful life.
2/3" is a perfectly valid creative format choice for any kind of production, including feature films. I expect Scarlet will have a long and profitable service life for RED. Plus since it is part of a modular system one can add an S35 or even an FF35 brain to the mix when those are available.
 
I know its totally impracticaly, but I would buy that for the simple fact that it would look quite interesting perched on a shelf. Or perhaps cradled in a brass desk display? Hmm...

I have an antique 18-180mm Angenieux f1.8 zoom made for a 1" format TV camera. The front element is 6 inches in diameter and it weighs 21lbs. The lens is much bigger than my GL2 camcorder.
 
2/3" is a perfectly valid creative format choice for any kind of production, including feature films.

I totally agree. If I had the resources I'd buy both a 2/3" and S35 because I know there would be times I'd like to use each. Of course, another benefit of the modular concept, is that I can just buy a 2/3" brain when I have the money.


Plus since it is part of a modular system one can add an S35 or even an FF35 brain to the mix when those are available.

Recent posts suggest that RED has already started R&D for their next-next sensor. Of course...I'd like to have a camera in my hands.
 
I have an antique 18-180mm Angenieux f1.8 zoom made for a 1" format TV camera. The front element is 6 inches in diameter and it weighs 21lbs. The lens is much bigger than my GL2 camcorder.

Not looking to sell it, are you? I could do with some office swag :2vrolijk_08:
 
3k is probably the practical upper limit for a 2/3" sensor if one also expects to balance sensitivity and dynamic range into the equation. There are damn few lenses that will resolve 3k on a 2/3" sensor and none that could possibly resolve 5k.
One of Scarlet's outstanding features is the ability to shoot for the 2k digital cinema format in both standard 1.85:1 and 2.4:1 cinemascope without the need to upscale from a lower resolution. It will downscale to both formats for superior noise and resolution compared to any native 1080p HD camera. 4k may be the cinema format of the future, but 80% of first gen digital theaters are 2k equipped and they will be around for at least 10 years. The projectors cost too much to not milk them for every hour of useful life.
2/3" is a perfectly valid creative format choice for any kind of production, including feature films. I expect Scarlet will have a long and profitable service life for RED. Plus since it is part of a modular system one can add an S35 or even an FF35 brain to the mix when those are available.

Nobody really upscales 1080p to 2k. We know a 2.4:1 crop from 1080p works well. That's the format everyone distributes 2.4:1 HD in. 2K is 2048x1080 and 1080p is 1920x1080. Only 6.7% less for 2,4:1. Identical resolution for 16:9.

A 5um pixel is a practical limit for a lens that expects to be any good wide open. It's 100 line pairs/mm. 5um is also important for diffraction and better use of the full aperture range. 5um is what a 2/3" 1080p camera uses. Pixels smaller than 5um in a 2/3" sensor are typically limited by the lens resolution.

The 3K camera has pixels half as large in area vs the 1080p camera. There is no noise advantage, quite the contrary. One stop of dynamic range and sensitivity is lost and there is an extra resampling filter that the signal needs to go through for 1080p output. On the native 1080p camera you get the original sensor MTF. On the 3k camera you get higher MTF for the sensor but MTF drop in the downsampling filter, downsampling alias because no filter is perfect, and the original lens will be the limit in almost every cine lens (designed for about 50 line pairs/mm, 3x less than what a 3k sensor requires). The only advantage of 3K is the higher chroma resolution, but distribution is 4:2:0 so that chroma cannot be distributed, 1080p bayer already captures all chroma the distribution can use. The resampling filter will dowsample chroma by a strange ratio which can introduce artifacts combined with the compression color sampling. The 3K chroma will probably be inferior to the 1080p chroma for 1080p blurays etc.

3K on 2/3" for 1080p distribution is simply consumer oriented marketing on the simplistic notion that 3K > 1920 which ignores all limitations imposed by physics and required processing. 4K for s35 makes sense, the pixels can be larger than 5um and a 2:1 downsampling works well. But 3K on 2/3"? Sony and Panasonic would never do that, which is why all professional Sony cameras including the s35 F35 are native 1080p resolution and transfer 1:1 pixel to 1080p distribution. Light gathering and dynamic range are too valuable to waste on marketing.
 
Nobody really upscales 1080p to 2k. We know a 2.4:1 crop from 1080p works well. That's the format everyone distributes 2.4:1 HD in. 2K is 2048x1080 and 1080p is 1920x1080. Only 6.7% less for 2,4:1. Identical resolution for 16:9.

A 5um pixel is a practical limit for a lens that expects to be any good wide open. It's 100 line pairs/mm. 5um is also important for diffraction and better use of the full aperture range. 5um is what a 2/3" 1080p camera uses. Pixels smaller than 5um in a 2/3" sensor are typically limited by the lens resolution.

The 3K camera has pixels half as large in area vs the 1080p camera. There is no noise advantage, quite the contrary. One stop of dynamic range and sensitivity is lost and there is an extra resampling filter that the signal needs to go through for 1080p output. On the native 1080p camera you get the original sensor MTF. On the 3k camera you get higher MTF for the sensor but MTF drop in the downsampling filter, downsampling alias because no filter is perfect, and the original lens will be the limit in almost every cine lens (designed for about 50 line pairs/mm, 3x less than what a 3k sensor requires). The only advantage of 3K is the higher chroma resolution, but distribution is 4:2:0 so that chroma cannot be distributed, 1080p bayer already captures all chroma the distribution can use. The resampling filter will dowsample chroma by a strange ratio which can introduce artifacts combined with the compression color sampling. The 3K chroma will probably be inferior to the 1080p chroma for 1080p blurays etc.

3K on 2/3" for 1080p distribution is simply consumer oriented marketing on the simplistic notion that 3K > 1920 which ignores all limitations imposed by physics and required processing. 4K for s35 makes sense, the pixels can be larger than 5um and a 2:1 downsampling works well. But 3K on 2/3"? Sony and Panasonic would never do that, which is why all professional Sony cameras including the s35 F35 are native 1080p resolution and transfer 1:1 pixel to 1080p distribution. Light gathering and dynamic range are too valuable to waste on marketing.


Tell that to Arriflex. The Alexa downscales 2880 x 1620, a 1.5x horizontal and vertical ratio, in camera to get 4:4:4 Prores 1080p HD output. This is exactly the pixel count and ratio Scarlet will use for 16:9.

Bayer raw does not equate to any encoded video sampling mode. The ratio of red and blue to green pixels is similar to the color sensitivity distribution of the human eye. It has become the most widely adopted sensor pattern for still cameras precisely because it is the closest mimic to the human eye and Bayer pattern sensors are most efficient at producing images with minimum visible artifacts with proper debayer algorithms. Compared to a 3 chip camera, oversampling and down scaling are an effective way to reduce noise and artifacts.

I don't know if you have played with Red's debayer software and files at all, but it is pretty amazing. You can debayer and output in any file format or color space you care to, from full resolution 16 bit 4:4:4 Tiff or Jpeg to scaled HD or SD video formats. The native aspect ratio of the sensors, including 3k Scarlet, is 1.89:1 digital cinema wide screen format.

I agree with you for the most part on lenses, but premium grade modern 16mm lenses are significantly sharper than 100 lpmm, many will hit double that at 50% mtf one or two stops off of full open. This is true of many current generation top line cine lenses for S35 too, one of the reasons they are so expensive. But yes optics will be the limiting factor on resolution, as it is for most HD cameras anyway unless you are buying the top of the line zooms.

One of the most impressive things about the new Canon XF series is the lens performance. That is one very well designed lens/camera combination.
 
Last edited:
Nobody really upscales 1080p to 2k.

This is a bit off-topic, but it's a good learning opportunity for me...and the answer shouldn't spiral out of control.

I've always heard that films that are shot on 1080p cameras (F23, F900, Genesis, Viper, sometimes D20) generally (as in most of the time) do get upscaled for a 2K DI.

As best as I can remember, Episode 3, Sky Captain, Sin City, Reign Over Me, Public Enemies, Alice in Wonderland, 2012, Avatar, and Tron Legacy are all examples of the more popular/larger films that went from the 1080p acquisition through 2K DI to a 35mm (or even 70mm) print, for everything from 1.44:1 to 2.35:1.

So when they do a 2K DI with 1080p source files, do they just leave it at 1080p, eat that extra 128 lines of resolution, and not upscale? Or am I missing something?

Light gathering and dynamic range are too valuable to waste on marketing.

This is one of the big question marks for me. S35 M-X is pretty solidly within the 13 stop bracket, with the new Alexa offering a bit more. So I wonder where the 2/3" is supposed to fall. According to the RED team, it's supposed to be "very close"...which makes me really want to see some charts and some numbers!
 
Tell that to Arriflex. The Alexa downscales 2880 x 1620, a 1.5x horizontal and vertical ratio, in camera to get 4:4:4 Prores 1080p HD output. This is exactly the pixel count and ratio Scarlet will use for 16:9.

Bayer raw does not equate to any encoded video sampling mode. The ratio of red and blue to green pixels is similar to the color sensitivity distribution of the human eye. It has become the most widely adopted sensor pattern for still cameras precisely because it is the closest mimic to the human eye and Bayer pattern sensors are most efficient at producing images with minimum visible artifacts with proper debayer algorithms. Compared to a 3 chip camera, oversampling and down scaling are an effective way to reduce noise and artifacts.

I don't know if you have played with Red's debayer software and files at all, but it is pretty amazing. You can debayer and output in any file format or color space you care to, from full resolution 16 bit 4:4:4 Tiff or Jpeg to scaled HD or SD video formats. The native aspect ratio of the sensors, including 3k Scarlet, is 1.89:1 digital cinema wide screen format.

I agree with you for the most part on lenses, but premium grade modern 16mm lenses are significantly sharper than 100 lpmm, many will hit double that at 50% mtf one or two stops off of full open. This is true of many current generation top line cine lenses for S35 too, one of the reasons they are so expensive. But yes optics will be the limiting factor on resolution, as it is for most HD cameras anyway unless you are buying the top of the line zooms.

One of the most impressive things about the new Canon XF series is the lens performance. That is one very well designed lens/camera combination.

Was the Alexa sensor designed specifically for Alexa? I don't know. The Kodak CCD, the Sony (2/3" and s35) and the best Panasonic 2/3" CCDs were all designed for 1080p output and they are exactly 1920x1080 resolution.

Bayer produces a 1920x1080 luma and it has exactly 960x540 R,G,B pixels which exactly describes a 4:2:0 1920x1080 delivery.

Downsampling does very little for noise. It can't make up for the loss of using small pixels in the first place. Maximum possible pixel size is key to low noise and high dynamic range.

The 16mm lenses I have seen are not that sharp wide open. Nothing those interested in a 10,000 euro camera can afford is any good at 150lp/mm wide open. Remember that 16mm film is not even considered HD by most HD channels and lenses are part of the problem. Try SLR lenses. Barely 50lp/mm on the best primes wide open.
 
This is a bit off-topic, but it's a good learning opportunity for me...and the answer shouldn't spiral out of control.

I've always heard that films that are shot on 1080p cameras (F23, F900, Genesis, Viper, sometimes D20) generally (as in most of the time) do get upscaled for a 2K DI.

As best as I can remember, Episode 3, Sky Captain, Sin City, Reign Over Me, Public Enemies, Alice in Wonderland, 2012, Avatar, and Tron Legacy are all examples of the more popular/larger films that went from the 1080p acquisition through 2K DI to a 35mm (or even 70mm) print, for everything from 1.44:1 to 2.35:1.

So when they do a 2K DI with 1080p source files, do they just leave it at 1080p, eat that extra 128 lines of resolution, and not upscale? Or am I missing something?



This is one of the big question marks for me. S35 M-X is pretty solidly within the 13 stop bracket, with the new Alexa offering a bit more. So I wonder where the 2/3" is supposed to fall. According to the RED team, it's supposed to be "very close"...which makes me really want to see some charts and some numbers!

As an engineer, I believe upscaling 1080p to 2K makes no sense. Filling in the width with black will produce better results.

No, Scarlet will not be very close because that's not possible. It should be >1 stops inferior because the pixel area is 2.68x smaller. The Scarlet pixel area is somewhere between that of an 1/3" 1080p camera and that of a 1/2" 1080p camera. MX and Alexa are not 13stops in the tests I have seen. On the Arri test Alexa is 12.1 and Red One is 11.3 but I don't know if that was MX. The original Red One sensor was supposed to be 11.3stops but it's 8-9 in most tests.
 
Status
Not open for further replies.
Back
Top