ARRI HD D-20 camera test screening

[keidrych wasley]

Saw the new arri d-20 the other day. This is the newer version than the prototype revealed last year and will be given for productions soon though i think not yet for rental, though that will obviously follow. It is not confirmed whether it will ever be for sale. This camera is arri's answer to the genesis and from what i saw in the 5 minute test is extremely impressive though the test was graded and not raw footage. To my eye it looked like film (was shot progressive i think, and displayed on HD monitor) but it would have been good to have seen the same test shot on 35 for comparing. I haven't used HD so i know little about it's capabilities (other than what i've seen in the cinema), and i haven't seen the genesis test footage so i can't compare it to that either. Rental costs will obviously be huge but the technology is very exciting. The test was shot by Sue Gibson bsc, it would be interesting to know what she thought of the ungraded rushes. The camera takes 35mm lenses and has an optical viewfinder so it feels just like film. The camera also looks like a film camera. It is single chip 2k resolution and can go to 150frames.

[Wendell_Greene]

It is single chip 2k resolutionand can go to 150frames

 

 

Yes, the D-20 uses a single CMOS sensor instead of CCDs.

 

My question is this: what did they do in the new model to make this possible? How did they solved the bandwidth limitation on the HD-SDI interfaces? Because with the prototype, even if you used all three of the camera's HD-SDI interfaces in parallel, the camera could only go as high as 90 pFPS.

[John Sprung]

Where did you see it? Where can we see it?

 

 

 

-- J.S.

[keidrych wasley]

I'm a cinematography student at the NFTS and a guy from arri came down to show the kit. Didn't get to shoot anything though. How they got to 150 fps i have no idea but if i see the arri guy again i'll ask, though all this flies over the top of my head. Give arri a call and i'm sure they'll give info on when it can be seen etc.

[Michael Nash]

The test was shot by Sue Gibson bsc, it would be interesting to know what she thought of the ungraded rushes.

<{POST_SNAPBACK}>

 

What's the point? These newer generation HD cameras are designed to capture as much info as possible, for the express purpose of grading. "Ungraded" footage from the Viper looks flat and green, but that doesn't mean your footage will come out that way.

 

The ungraded "look" is irrelevant. The technology is not used in the same manner as film (i.e. contact printed), so it doesn't have to follow the same standards of comparison as film. The info that's captured, and how well it can be manipulated, is what's relevant with digital cameras.

[Phil Rhodes]

Hi,

 

> I'm a cinematography student at the NFTS

 

Good for you.

 

> all this flies over the top of my head.

 

Okay, so what are they teaching you?

 

Phil

[Morgan Peline]

> I'm a cinematography student at the NFTS

 

Good for you.

 

> all this flies over the top of my head.

 

Okay, so what are they teaching you?

 

Phil.>

 

 

Well, to be creative mostly. How the d-20 can do 150fps is not essential knowledge really, is it? It's not about being a technocrat here, though you can be if you want, i'm just not. Besides, i've just started and training is predominantly in film at the moment so maybe all this is to come, though i hope not, because i don't think it will make me a better storyteller.

>

 

What's the point? These newer generation HD cameras are designed to capture as much info as possible, for the express purpose of grading. "Ungraded" footage from the Viper looks flat and green, but that doesn't mean your footage will come out that way.

 

The ungraded "look" is irrelevant. The technology is not used in the same manner as film (i.e. contact printed), so it doesn't have to follow the same standards of comparison as film. The info that's captured, and how well it can be manipulated, is what's relevant with digital cameras.>

 

I guess i mean in terms of the colours taken from the initial image that are then enhanced. The footage that is shot in 'video' mode and how it looks before any enhancements next to exactly the same process with film. The colours were fantastic in the video shown, red was used a lot, which is a difficult colour for video. But i haven't used the HD kit yet, so i don't have the knowledge or perspective of the process that you do.

[David Mullen ASC]

Even with film, we don't obviously use the low-con, orangey negative image directly but make a positive, etc.

[Wendell_Greene]

How the d-20 can do 150fps is not essential knowledge really, is it? It's not about being a technocrat here, though you can be if you want, i'm just not.

>

 

 

 

Knowing the top speed the latest model BMW is capable of reaching may not be view as important or essental information to me. I may be more concerned with knowing how many miles per gallon city and highway the car gets and how comfortable the seats are going to feel to someone who will end up sitting in his car for hours on end in Los Angeles traffic. But if the top speed of the cars rolling out of production are higher than that of the prototype the sales and marketing people were pushing and listing in their brochures and sales pitches and I notice it because it relates to other aspects of the car's performance, and my research tells me that the engineers had to make modifications, then I'm asking the question until I get a satisfactory answer.

 

And how do you think I'd respond to a saleperson or anyone else for that matter whose response was "Ah, you don't need to know that, it's not essential information, since you'll never drive that fast!" when I posed the question?

 

My point is this. How do you know that what knowledge isn't essential to me? Why you would assume that the camera's ability to perform as advertised isn't crucial to my needs as a cinematographer to order for me to support the story I'm being asked to photograph? You don't. It's a poor assumption.

 

I've noticed an increasing tendency on the part of some people here, and especially students to post when they don't fully know the answer to a question and then to react defensively when they're called on it, by saying "well I'm just a storyteller, not a technician" using the word technician perjoratively.

 

Since when has a cinematographer's need to ask questions in order to better understand the technical aspects of the tools companies are asking us to use, come to be viewed in a negative light? And in all places a board related to Cinematography? We continually ask to justify why we want to use this or that piece of equipment to producers, and conversely, they'll come to use wanting to use equipment that they've heard about. We have to know, not guess, as cinematographers we can't afford to be luddites.

[Chris Keth]

How is the depth of field situation? Does most HD have the same trouble as other video, namely too much depth of field?

[Keith Mottram]

How is the depth of field situation? Does most HD have the same trouble as other video, namely too much depth of field?

<{POST_SNAPBACK}>

 

35mm lenses, 35mm dof.

[David Mullen ASC]

Depth of field characteristics of lenses are affected by the size of the image target area; regular HD uses 2/3" CCD's so are in the ballpark of Super-16's depth of field. The Arri D20, Panavision Genesis, and Dalsa Origin use 35mm-sized imagers so the depth of field characteristics are similar to 35mm.

[Mike Brennan]

Depth of field characteristics of lenses are affected by the size of the image target area; regular HD uses 2/3" CCD's so are in the ballpark of Super-16's depth of field. The Arri D20, Panavision Genesis, and Dalsa Origin use 35mm-sized imagers so the depth of field characteristics are similar to 35mm.

<{POST_SNAPBACK}>

 

I am interested in the notion that depth of field becomes more noticable to the viewer as higher resolution capture and displays are used.

 

When HD lenses were introduced I was crowing loudly that depth of field had to be the same as standard def lenses.

Although technically correct the higher quality capture and display of HD made the fall off more apparent. The overall effect is to display a depth of field that appears less than lower res capture and display. (I termed the phrase "apparent depth of field" to refer to displayed depth of field)

 

Bearing this in mind perhaps we should wait and see if the apparent depth of field of 35mm lenses on say Genesis when digitally projected is more less or the same as 35mm film capture and projection.

 

Someone could some emperical testing and shoot this apparent theory to bits....:)

 

Mike Brennan

[Guest Jim Murdoch]

The Arri D20, Panavision Genesis, and Dalsa Origin use 35mm-sized imagers so the depth of field characteristics are similar to 35mm.

<{POST_SNAPBACK}>

 

The problem is that with video cameras, much of the perceived sharpness comes from so-called "detail correction", which is really only the processing computer's "best guess" as what edges were meant to be sharp and which were not.

 

The problem is always that when you deliberately try to de-focus something, the detail correction circuitry helpfully tries to "crispen" it back up for you! Until somebody can figure out a way to fit a USB socket into the back of the cinematographer's head to inform the processing software what is intended, video cameras are always going to produce un-natural looking images!

 

As it is, I do rather wonder why it took so long to post-produce Star Wars II, which after all was a fairly ordinary-looking Sci-Fi flick by today's standards. Unless it was because they had to touch up all the detail correction by hand....

 

Compare this with over-sampled film. When you over-sample 35mm film (eg 6K for 4K output) in that case the digital processing circuitry does know which bits are meant to be sharp edges and which are not and it can apply the detail correction "makeup" much more intelligently. It can also means that the optical low-pass filtering can be applied in software rather than in a physical spatial filter, and so the "severity" of this can be varied at will. That means that the filtering is only be applied to the parts of the image that really need it, and the result is a considerably sharper image than theory would predict possible.

 

This is one reason downconverted HDTV always looks better on SDTV sets than images originated on SD equipment.

 

Which is really another way of saying that no commercially available SDTV cameras were ever able to fully exploit the capabilities of both the PAL and NTSC systems, and that current generation "HDTV" cameras are really only represent the technical perfection of SD cameras.

 

If and when true 1920 x 1080 screens become commonplace in living rooms, only then will people begin to see the difference between film scanned at 4K and down converted to 2k, and film simply directly scanned at 2K.

[Guest Kai.w]

The problem is always that when you deliberately try to de-focus something, the detail correction circuitry helpfully tries to "crispen" it back up for you! Until somebody can figure out a way to fit  a USB socket into the back of the cinematographer's head to inform the processing software what is intended, video cameras are always going to produce un-natural looking images! 

<{POST_SNAPBACK}>

Oh yeah. As if film looks natural...hmmm...

 

As it is, I do rather wonder why it took so long to post-produce Star Wars II, which after all was a fairly ordinary-looking Sci-Fi flick by today's standards. Unless it was because they had to touch up all the detail correction by hand....

<{POST_SNAPBACK}>

No offense, but this is getting absurd.... What did you say about getting some fresh air cause the y2k bug did not prove that bad... you know there are just way too many totally weird conspiracy theories out there on the internet. Lets not add some more absurd ones.

 

-k

[Alvin Pingol]

>>Until somebody can figure out a way to fit a USB socket into the back

>>of the cinematographer's head to inform the processing software what is

>>intended, video cameras are always going to produce un-natural looking

>>images!

 

Always?

 

Setting detail/sharpness to 0, or Off, effectively solves this (assuming, of course, the camera is capable of delivering sharp images without the need of electronic edge enhancement! :D )

[Michael Nash]

>>Until somebody can figure out a way to fit  a USB socket into the back

>>of the cinematographer's head to inform the processing software what is

>>intended, video cameras are always going to produce un-natural looking

>>images! 

 

Always?

 

Setting detail/sharpness to 0, or Off, effectively solves this (assuming, of course, the camera is capable of delivering sharp images without the need of electronic edge enhancement!  :D )

<{POST_SNAPBACK}>

 

I don't see edge enhancement as a major problem for video/HD cameras, as it is controllable as Alvin points out. If you feel the raw image with no in-camera detail added is too soft, you can add some sharpening in post. This is common with 16mm-video transfers, as the camera can't add enhancement, yet you can squeeze out a slightly sharper looking image in transfer by adding a small amount of enhancement.

 

BTW, setting the detail to "0" in Sony HD cameras is not the same as turning the detail off. Sony uses a scale of -99 to +99, where zero represents the middle of the scale and the "nominal" amount of sharpening. Not that that's what you should use...

 

There are other ways to take the "curse" off edge enhancement, like using lens diffusion or naturally softer lenses. Nets on the lens for example can accentuate the difference between sharp-looking details and those that are out of focus, even though technically the net is softening everything.

 

It's all a matter of degree, taste, and control -- which is the basis for the art and craft of cinematography.

[Alvin Pingol]

BTW, setting the detail to "0" in Sony HD cameras is not the same as turning the detail off. Sony uses a scale of -99 to +99, where zero represents the middle of the scale and the "nominal" amount of sharpening.

<{POST_SNAPBACK}>

 

Thanks for that important correction. I tend to forget about negative values for things like sharpness because, frankly, I don't think it makes sense - "Negative" sharpness should technically be adding blur... but oh well. :P

Edited March 20, 2005 by Alvin Pingol

[Michael Nash]

I am interested in the notion that depth of field becomes more noticable to the viewer as higher resolution capture and displays are used.

...

 

Bearing this in mind perhaps we should wait and see if the apparent depth of field of 35mm lenses on say Genesis when digitally projected  is more less or the same  as 35mm film capture  and projection.

 

Someone could some emperical testing and shoot this apparent theory to bits....:)

 

<{POST_SNAPBACK}>

 

It's not just a notion, it's a known factor called "circle of confusion."

 

You don't hear the term used much with HD, since technically the term refers to an actual dimension on a piece of film. But in any case the principle is the same; the finest resolution available with the optics, imaging, and display being used is one factor used to determine depth of field. With film you often have a choice of several values for the circle of confusion for any given film format; for example projection will require a smaller tolerance than an SD TV transfer would.

 

That said, the Genesis is still effectively an HD camera with no better than 1920x1080 pixels in its current configuration. I think it's pretty commonly known what that type of resolution and sharpness that can give, and wouldn't be hard to put a number to as an "effective" circle of confusion. I don't think it's any great mystery, we all pretty much know that HD resolution is softer than the highest 35mm film resolution, so the circle of confusion and depth of field will follow that.

 

In practice, I would expect the depth of field characteristics of Genesis to be fundamentally the same as 35mm film, and almost exactly the same as 35mm film gone through a 2K scan.

[Mike Brennan]

It's not just a notion, it's a known factor called "circle of confusion."

 

You don't hear the term used much with HD, since technically the term refers to an actual dimension on a piece of film...........

 

In practice, I would expect the depth of field characteristics of Genesis to be fundamentally the same as 35mm film, and almost exactly the same as 35mm film gone through a 2K scan.

<{POST_SNAPBACK}>

 

 

Circle of confusion has nothing much to do with what I was referring too, which is the practical application of depth of field on digital cameras.

 

Circle of confusion could be described as the depth of field of the image *at the image plane* of film or digital cameras. Measured in fractions of an inch or in case of 2/3 inch HD fractions of a millimeter.

 

 

I was referring to the display of very sharp pictures. So sharp that the area in front of and behind would normally be regarded as in focus appears, in relation to the point of focus, relatively soft.

 

 

"We can achieve critical focus for only one plane in front of the camera, and all objects in this plane will be sharp. In addition, there will be an area just in front of and behind this plane that will appear reasonably sharp (according to the standards of sharpness required for the particular photograph and the degree of enlargement of the negative). This total region of adequate focus represents the depth of field."

 

The Camera, Ansel Adams, 1980

 

 

"Zones of focus", "critical focus", "acceptable sharpness" are all fairly subject descriptions of depth of field.

 

 

The higher the res of the display (or in Ansel Adams case the larger the neg and print) the less the "apparent" depth of field.

 

 

As for cinematography prospect of 4k high res digital displays and high res film or digital capture may require a rethink on the tried and trusted depth of field tables.

Maybe something to learn from those who shot for 70mm projection?

 

 

Mike Brennan

[Mike Brennan]

BTW, setting the detail to "0" in Sony HD cameras is not the same as turning the detail off. Sony uses a scale of -99 to +99, where zero represents the middle of the scale and the "nominal" amount of sharpening. Not that that's what you should use...

 

<{POST_SNAPBACK}>

 

Aghhh .... quoting numbers from Sony menues is dangerous as the numbers differ between camera models and even versions of same camera. (software is designed by different teams)

 

It appears that on the f950 neg softening starts at -40.

In previous cameras there has been always residual amount of detail. This may not be the case with newer cameras.

 

Mike Brennan

[Riku Naskali]

I was under the impression that for example the Panavised Sony has no edge enhancement?

 

However I quite dislike the effects edge enhancement produces, at least on SD world. I've been shooting some stuff on dvcpro and tried to turn it off. Too soft. When I increase it even slightly, it produces the tasty white/black lines around subjects. So I guess the thing is, it just doesn't have enough resolution.

 

I would love to get my hands on HD, though :)

[Guest Jim Murdoch]

What did you say about getting some fresh air cause the y2k bug did not prove that bad...  -k

<{POST_SNAPBACK}>

 

I didn't say that at all; have another look.

you know there are just way too many totally weird conspiracy theories out there on the internet. Lets not add some more absurd ones.

 

<{POST_SNAPBACK}>

 

One of the most endearing things about conspiracy theorists is they way they often refuse to accept the possibility of the existence of conspiracies that are at least theoretically possible!

 

This whole Digital Cinematograpy thing was started by ex-Panavision CEO John Farrand in a desperate attempt to prove that the company isn't currently worth something like MINUS $400 million! The upper management is going to keep stringing this garbage out for as long as they can until Ronald Perelman finally pulls the plug!

[Guest Jim Murdoch]

I was under the impression that for example the Panavised Sony has no edge enhancement?

 

 

<{POST_SNAPBACK}>

That's what they'd like you to believe.

 

But it doesn't matter whether it's done in the camera or in post; detail correction is not high resolution; it's just the illusion of high resolution. It's a bit like an impressionist painting; its looks fine if you don't look too closely.

[Guest Jim Murdoch]

(assuming, of course, the camera is capable of delivering sharp images without the need of electronic edge enhancement!  :D )

<{POST_SNAPBACK}>

Ha bloody ha! Without detail correction it might make a passable Standard Definition camera.