d20 grabs

[Keith Mottram]

i'm sure some of you already have this link from the cinematography.net site but here are some d20 images and they look pretty sweet.

 

http://www.arri.de/prod/cam/d_20/d_20_images.htm

 

keith

[Filip Plesha]

I'm suprized nobody had any comment..

 

 

I think it certainly looks better than the star wars stuff (or is it just better lighting and a REAL set?)

 

It's difficult to tell without a film comparison

 

The initial impression is that it looks a lot like film. Maybe the look would "fall appart" in different light, who knows.

Acutally I looks like a DSLR image, which is a good thing I think.

 

It is a pleasant suprize, and I'm definitivley over my old fear that some day I will not enjoy images in cinema the way I do today because everything would look like star wars or a brasilian soap opera. It is definitvley an enjoyable look.

[David Mullen ASC]

I'm suprized nobody had any comment..

<{POST_SNAPBACK}>

 

It's well-lit but what can you say about a small, static image on a computer screen shot under ideal circumstances?

[Tomas Koolhaas]

Hello,

let's not forget it's been graded aswell, which all frame grabs from new cameras seem to be before the cameras are actually released to the rental houses. I would like to see some 'raw' footage from the D-20. Although something tells me that if anyone could afford to rent one of these they would be able to get some beautiful images out of it.

I was wondering whether the D-20's 'film' mode is very similar to the Viper's 'film stream' mode (4:4:4 uncompressed), does anyone know if this makes the output image on the monitor in real time somewhat unreliable in terms of viewing colour and brightness, as I heard was the case with the viper's 'film stream' mode on Collateral (AC).

Is this why Russel carpenter used a light meter and waveform monitor instead of a well calibrated monitor, or did they just want to be super accurate with their exposure.

Cheers.

Tomas.

[David Mullen ASC]

Well, how "reliable" is the video tap image on a 35mm camera? If you're recording RAW or unprocessed from the Viper, Genesis, D20, Dalsa, etc., then the monitor image will either be very flat and washed-out, or go through some LUT box or something to simulate a look that you plan on giving it later in color-correction. Does that make it "inaccurate" or "unreliable"?

 

I believe the D20, like the Viper, can send out an unprocessed signal like "film stream" mode in the Viper, or a processed signal like the Viper in "HD stream".

 

The ability to color-correct the image from an Arri-D20 to look good sort validates it in my mind, not invalidates it. I'm not sure what you'd get from seeing an ungraded image other than notice how flat and washed-out it looks.

[Mike Brennan]

Well, how "reliable" is the video tap image on a 35mm camera?  If you're recording RAW or unprocessed from the Viper, Genesis, D20, Dalsa, etc., then the monitor image will either be very flat and washed-out, or go through some LUT box or something to simulate a look that you plan on giving it later in color-correction. Does that make it "inaccurate" or "unreliable"?

 

I believe the D20, like the Viper, can send out an unprocessed signal like "film stream" mode in the Viper, or a processed signal like the Viper in "HD stream".

 

The ability to color-correct the image from an Arri-D20 to look good sort validates it in my mind, not invalidates it.  I'm not sure what you'd get from seeing an ungraded image other than notice how flat and washed-out it looks.

<{POST_SNAPBACK}>

 

The D20 will not output log according to Arri.

And are we sure filmstream is actually any better than a decent ad mapping and gamma curves?

 

 

Mike Brennan

[David Mullen ASC]

I was just reading an interview with John Kroll in "Film & Video" saying that recording 10-bit log unprocessed from a digital camera like the F-950, compared to using video gamma, was a waste of bandwidth, because the 10-bit Cineon log format is really optimized for color negative, not digital images.

[Keith Mottram]

It's well-lit but what can you say about a small, static image on a computer screen shot under ideal circumstances?

<{POST_SNAPBACK}>

 

David, just in case you hadn't noticed you can actually download a full res tiff file. I also thought it was interesting that they recorded these images onto SR. Obviously they are graded, but having played around with these images in Photoshop I feel that they are quite impressive. To me they seem sharper and more natural than the viper (in filmstream mode)- even considering that it was shot in ideal conditions. I also wonder if the fact that they shot to SR means that ARRI is now considering the SR to be they're choice of recording format.

 

Keith

[Guest Kai.w]

I was just reading an interview with John Kroll in "Film & Video" saying that recording 10-bit log unprocessed from a digital camera like the F-950, compared to using video gamma, was a waste of bandwidth, because the 10-bit Cineon log format is really optimized for color negative, not digital images.

<{POST_SNAPBACK}>

Not having read the article yet, I'm not sure what he is getting at, cause for instance the Viper Filmstream mode is not necessarily a standard cin lut conversion (if there is actually such a thing ;-)) , no ?

 

-k

[Phil Rhodes]

Hey, look, the candle flames are clipped to white.

 

It's another video camera. Ooh, Matron, did I say that?

 

Phil

[Filip Plesha]

Hey, look, the candle flames are clipped to white.

 

It's another video camera. Ooh, Matron, did I say that?

 

Phil

<{POST_SNAPBACK}>

 

They would not be clipped only in "raw" negative scan, but the low contrast negative image is not viewable, as soon as you convert it to viewing contrast, flames would blow to white. Every medium blows up highlights here or there

[Michael Most]

I also wonder if the fact that they shot to SR means that ARRI is now considering the SR to be they're choice of recording format.

 

Well, considering that SR is the ONLY currently available tape based recording format that can record RGB 4:4:4, everyone would consider it their choice. Everyone who wants any degree of portability, that is.

 

Arri has also mentioned that they are supporting the Thomson Venom RAM-based recording module as well.

[Michael Most]

I was just reading an interview with John Kroll in "Film & Video" saying that recording 10-bit log unprocessed from a digital camera like the F-950, compared to using video gamma, was a waste of bandwidth, because the 10-bit Cineon log format is really optimized for color negative, not digital images.

 

I respectfully and completely disagree.

 

First of all, John is making an assumption that is not true - he is assuming that the Cineon profile is being used for the log curve that is implemented in devices such as the Viper and the Genesis. This is not the case. The Cineon profile is indeed optimized for film, it is based on printing density and as such is valuable in cases where you want to see a simulation of a film print, such as a DI suite. However, the video cameras are not using this profile, they are using their own profiles to maximize the use of the 1024 available levels in the 10 bit encoding. This is quite valuable in that it allots more levels to areas of the image that require more detail (like the lower midranges) and fewer levels to those that don't (like the super-whites). This not only allows better representation of the original image, it also allows more flexibility in color correction and other manipulations because you're not "inventing" information that wasn't there in the first place. How much an image benefits from this is content dependent, but the theory is sound and in many cases provides a distinct advantage.

 

As for "wasting bandwidth," the bandwidth required is determined by the size of the frame, the bit depth, and the frame rate. These are exactly the same regardless of how you encode the individual levels. Log or linear makes absolutely no difference. What he is likely referring to is the use of bandwidth-saving YUV instead of RGB. If that's the case, I disagree with him here as well.

[Michael Most]

Hey, look, the candle flames are clipped to white.

 

It's another video camera. Ooh, Matron, did I say that?

 

Another reason why posting stills on an Internet site as a demo of a digital cinematography camera is a really, really bad idea - people like Phil will make such assumptions.

 

Motion picture photography must be judged in motion, and must be judged in the context of what was represented in the scene. Timing is based on subjective, not objective, opinions, and thus somewhere along the line someone decides just how bright or dark a scene should be, regardless of what the original capture contained in its more extreme areas. A bright candle will always be blown out, on film or on a digital image, if it is that much brighter than what is deemed to be the key element in the image. Timed stills are perhaps the single worst way to represent the abilities of a digital motion picture camera.

[Filip Plesha]

I think motion picture equipment is like photography equipment in motion, so it must first be judged by looking at one frame (to see how good it is as a simple photographic capturing device)

[David Mullen ASC]

I think motion picture equipment is like photography equipment in motion, so it must first be judged by looking at one frame (to see how good it is as a simple photographic capturing device)

<{POST_SNAPBACK}>

 

That only works in a COMPARATIVE situation. A single still frame is not a particularly accurate way of judging a moving image because certain information is revealed or hidden (like graininess) when displaying multiple frames in rapid succession.

 

Where a single frame can be useful is in comparison to another single frame of another camera, lens, format, etc.

 

In other words, I can't really say much about a single frame from the Arri D20 other than to say "looks nice". I've seen moving images on HD monitors from the Arri D20 though and it looks nice that way too.

 

But if I also had a single frame from a 35mm camera shot at the same time as the Arri D20, or another HD camera, then I could make side-by-side comparisons between certain qualities revealed only in that way, with the understanding that everything would look different as moving images.

 

I mean, think of a single frame from something shot in 5218. Would that tell you how much improved it was over 5279 or Fuji 500T without a side-by-side comparison? All you can say is "looks nice" if it looked nice.

 

We need some frame of reference in order to judge something more fully. If the Arri D20 is really meant to compete with 35mm color negative, then a frame of that as a reference would be nice. However, if Arri thinks of it as competing against other HD cameras, then a reference frame from one of those would be useful.

[Filip Plesha]

Yes, bu that's just it, the moving image hides a lot of stuff.

 

for example, many judge the DVD video image by playing a movie on a TV set. Of course it looks great that way.

 

but when you play your DVD on a high resolution computer monitor and then freeze the frame you see all kinds of problems. For one thing it is not sharp as a jpg image would be. Then it has noise in shadows, it has compression artefacts. etc.

all that is hidden when you play it on TV, and that is why I think it is best to stop the image and disect it on a computer and let it show its true nature.

[Mike Brennan]

Ref 40mm grab, the jaggies on her right hand, set against the window, are interesting if you blow it up. Multicoloured with a nice shade of blue every now and again. Look under her hand and you'll see random(?) primary coloured pixels poping up in the shadows at much higher luminence.

 

Dare I call these "smarties" an artifact?

Its not something I've seen on SR1W before so perhaps its from the camera?

 

Overall the pictures are sharper than I expected and a little noisier, will be interesting to see how lively the noise is on a large screen once the picture moves...

 

 

Arri in the UK at least are sending the D20 out with an SRW1 and also a Venom.

Idea is that you shoot untethered if you want to with the venom(Arri will call it something else like XXXX pak)

Then playout to the SR.

At end of the day the client walks away with SR tape equipment goes back to Arri.

Venom plays back in real time, (no data transfer mode as far as I'm aware) so I guess you'll need 2 venoms if you can't cope with a cable to the SR.

Arri are calling the xxxx pak the primary means of capture.

 

Looks promising.

 

 

Mike Brennan

[David Mullen ASC]

that is why I think it is best to stop the image and disect it on a computer and let it show its true nature.

<{POST_SNAPBACK}>

 

I don't know if I'd call that it's "true nature" -- moving images are meant to be seen moving, after all. A still frame from an interlaced-scan video camera, for example, looks more film-like than when it is in motion, so the still image is the one that is misleading.

 

But certainly freezing a movie image is revealing, just like looking at it under a microscope is -- as long as you consider that this is not how a motion picture image is meant to be judged.

[Filip Plesha]

well sure, testing the image in motion is the way to see how well does it work in real world and how does it subjectivley look to viewers

 

But all the technical stuff can be analised only from a still frame:

MTF, RMS granularity, characteristic curves etc.

Those are pure photographic properties of the medium, and those specs are devoted to a freezed frame rather than a print in motion.

 

I don't know maybe Im just thinking to much like a still photographer.

When I think about how grainy some MP film is, I think about the grain when you freeze the frame. Or when I think about sharpness, I don't think about subjective sharpness that comes from combining several images in the brain to one, I think about shaprness of one frame.

[Max Jacoby]

Kudos to Arri for showing the candle lit image. Dalsa did not put up that shot on their website and Panavision certainly won't since it'll be a gigantic embarassment to them.

[Keith Mottram]

Kudos to Arri for showing the candle lit image. Dalsa did not put up that shot on their website and Panavision certainly won't since it'll be a gigantic embarassment to them.

<{POST_SNAPBACK}>

 

Not surprised there's no candle shot from Dalsa with all the highlight clipping going on in their other grabs. Inface the Dalsa to me seems worse in every instance to the D20, although its hard to make an accurate comparison.

 

You've got to wonder how bad the Genesis grabs really are?

 

Keith

[Logan Schneider]

I attended the Cinegear Masterclass where these shots were taken. They same shots were also done using the a 435 with 5218 and 5229, a Panavision Genesis and a Dalsa Origin. Afterwards, the film was scanned in at 4K and downrezzed to 2K on a Spirit at Laser Pacific, where they were compared side by side using a 2K projector.

The Dalsa had rendered the image at an incorrect ISO (one of several problems with the Dalsa), so it was not really availible for comparison.

The D20 and the Genesis were much closer. Though the Genesis had a slight CCD defect (a thin verticle flaring from the candles), it seemed by far to be the most production ready of the 3 digital cameras.

The D20 had a different look than the Genesis, but it may have been because it already had settings applied to it in camera (I can't remember what these are called).

The Genesis had primo lenses and the others had the Master Primes, which exhibited zero breathing and no ghosting. That's right, zero breathing, no ghosting. Germans. I do think that the primos were sharper though.

Overall, the film had a richer and crisper feel to it even though it was downrezzed again to 1K because the processors couldn't handle the 2K for some reason. What was most impressive to me about the film was how much could be done with it in comparison to the other cameras. They had a relatively narrow color correction range before there was too much noise. This film kept it's smooth feel, even when abused.

 

This is most of what I remember. Most of the people around me spent the day talking about the girl.

 

Logan Schneider

DP

 

PS: I made a lighting diagram of the shoot, which I will post if I have a chance.

[Filip Plesha]

yea, It's a nice girl..

 

but what was with here eye? It seems like she has some problem, the eye looks bloody

[Filip Plesha]

oh, I haven't even seen the Dalsa pictures.

Very interesting.

 

I think Arri looks more like negative film.