ARRI CEO: "We want to have the worlds best digital camera this year".

[Lauri Kettunen]

"Our Goal is to have the best digital camera of the world, within 12-14 months." (Nov.2007) ... "I am certain that digital camera technology will develop much faster than everyone expects".

Yes, all this makes sense. Companies compete with each other, and in reality the level of knowledge is very much the same all over the world. People know about the same things, and thus knowledge does not make a big difference between various groups. Of course within RED they should feel they are the best, and even far better than the others. That's a good thing but also another issue.

As far as I understand, one of RED's main advantages is a small but dedicated R&D team and also a dedicated net of users. This is something which is difficult to establish, and Jim's timing was also very succesful.

So, don't have any doubts that ARRI were not able to develop a better camera. I expect their challenge will be the costs and marketing the product. This REDUserNet is the first genuine realization of the so call new business relying solely on the web I've seen. Establishing a similar forum of keen people is not easy. For example, can't avoid feeling bit sorry for SI. Technically they have a great product, but for their bad luck RED came out about the same time. A comparison of the number of posts written everyday in these two forums tells a lot. It is easy to imagine how things were different for SI if Jim never decided to develop a camera. So, one needs some luck as well, and one never knows about that.

 

[Graeme Nattress]

Maybe projected you don't notice the noise so much, but pixel for pixel on a 4K LCD monitor, even the best DLSR with 4K sensors do not hold up at full res, pixel for pixel, especially low-light shots.

Sure, but "video", pixel for pixel on an LCD doesn't stand up either. RED images are about the best I've seen for that level of scrutiny in the motion world.

Graeme

 

[Craig W. Bickerstaff]

Sure, but "video", pixel for pixel on an LCD doesn't stand up either. RED images are about the best I've seen for that level of scrutiny in the motion world.

Graeme

Are you implying that movies shot on the genesis won't hold up on a 1080p HDTV?
Maybe I'm reading this statement wrong

 

[Radoslav Karapetkov]

True. And sometimes they fuse into something magical
http://www.boacinema.com/misc/lucas2.mov

Bruce Allen
www.boacinema.com

Did that get him the funding for Star Wars? :biggrin:

 

[Andrew M.]

Graeme, I was playing with Canon 1Ds MkIII 21MP images
http://web.canon.jp/imaging/eosd/eos1dsm3/eos1dsm3_sample-e.html
and compared it to 4K images from RED on 2K wide scaled down window in CS3.
The test was not conclusive since I do not have RAW from Canon and RED raw I could push and scale with much better result. (mine Mk3 is still on backorder)

I would speculate though that Misterium II with 24MP and good scaling algorithm to 4K would yield much better result on the big screen? How slow scaling will be is yet another question.

I agree with your statement that 8K on 8K screen looks much better then 8K scaled to 4K
I guess our eyes have much better pixel combining capability then the best scaling algorithm.

You were experimenting with 4K to 2K scaling. I see there is a new way to do it in REDAlert.
What is your experience with it?

 

[M Most]

--If they want to make a HD-SDI centric camera (known, fast, real-time workflow) then they're up against the F23.

--If they want to make a data-based camera, they are up against the Red, whatever Sony will come out with, Dalsa, maybe Panavision, etc. etc.
In terms of image quality, of what I've seen, it looks like Red's images are looking the best. The very low noise is huge... it affects how far you can go in DI.

Glenn, have you worked with any underexposed (intentionally or otherwise) Red footage? Or some footage from the F23 under similar conditions? Because my experience is that the F23 has almost no discernible low level noise regardless of exposure, but the Red has a very significant amount of it - although it's really hard to tell whether it's random sensor noise, fixed pattern noise (there are often obvious vertical stripes), compression artifacts, debayering artifacts (even with full debayer), or a combination of all these things. Now, granted, when color correction is applied, most of this is masked in the blacks. But with any material that you're printing up in DI from an underexposed original, the noise is very, very significant, especially compared to the F23. And I've seen this on both the earlier cameras and one that was delivered only 2 weeks ago.

I think there's a gap right now in that not all people know how to use their color tools that well... I'm seeing footage where the skies are clipping out... what I think people should be doing is applying curves-based adjustments so you have highlight handling like film and to avoid that blown-out look.

While I would basically agree with this, I would also point out that the Red, like any electronic capture device, can and does clip highlights, and clip them hard, if the exposure is not controlled.

Workflow-wise, I think Red has a leg up on the competition among data-based cameras. (Cineform too, but not in a 4k camera currently AFAIK.) The artifacts from good compression are very negligible, yet they can make the workflow a lot more practical.

Well, the only current cameras that I would consider data-based are Red, Dalsa, and Silicon Imaging. Of those, I would say that Silicon Imaging has what is arguably the most complete and finished workflow. It has a built in color management system that allows looks to be created (they even have a built in version of Speedgrade right in the camera software) that are carried through post in metadata (something talked about but thus far not delivered by Red), it has real time debayering and color implementations (thanks to its use of and alliance with Cineform), and it implements in Premiere Pro on the PC everything that Red has promised (but not yet delivered) with Apple - and has done all this for at least a year and a half. I would say that Dalsa has a very complete workflow as well. Although it's not designed for desktop do-it-yourselfers, that doesn't make it any less polished. For the market Dalsa is trying to appeal to, the use of a Codex recording device is not out of line - and it provides raw capture and transcoding to Quicktime and Avid compatible files, as well as viewing copies, right in the box - and almost in real time. That's a pretty sophisticated approach, even if it doesn't happen to run on a MacBook Pro.

So for data-centric digital cinema cameras, I think it will be very hard to compete with Red.

Only if you're talking about the low budget market. I'm not sure I agree with that statement if you're talking about the entire movie-making universe.

I don't see Arri competing on price or workflow, so they have to really excel in image quality.

And they very well might.

 

[Graeme Nattress]

Are you implying that movies shot on the genesis won't hold up on a 1080p HDTV?
Maybe I'm reading this statement wrong

The only images from the Genesis I've seen on my monitors here were terrible, and probably not representative of the camera's true quality as I saw Panavision project at 2k in their display trailer.

And of course, beauty is the in the eye of the beholder....

Graeme

 

[Graeme Nattress]

I agree with your statement that 8K on 8K screen looks much better then 8K scaled to 4K
I guess our eyes have much better pixel combining capability then the best scaling algorithm.

You were experimenting with 4K to 2K scaling. I see there is a new way to do it in REDAlert.
What is your experience with it?

The new scaling direct from 4k works rather well, and also works well with a high OLPF compensation too, to make a very sharp image. The code is still in development, but should be very good / fast indeed when done. I think it's very nice now though...

My issue with scaling down is that you've got to balance three requirements:

lack of aliasing
ringing / halos
resolution

in your filter design. We all know nearest neighbour downscaling makes for very sharp images without halos or ringing, but highly aliased. Gaussian makes for no aliasing, no halos or ringing, but low resolution. All other filters trade that lack of halos or ringing for increased resolution while trying to keep aliasing down.

Graeme

 

[M Most]

The only images from the Genesis I've seen on my monitors here were terrible, and probably not representative of the camera's true quality as I saw Panavision project at 2k in their display trailer.

And of course, beauty is the in the eye of the beholder....

If you've watched any number of television series (Night Stalker, What About Brian, and currently Samantha Who? and Cashmere Mafia) you've seen Genesis images that "hold up" just fine on any HD monitor. Not to mention the numerous features shot with it (Click, Apocalypto, Flyboys, and Walk Hard are some of the better examples), all of which are just fine in their video releases.

Honestly, Graeme, Red has a great product. But the Genesis has been proven many times over, and its images have been scrutinized by some of the most discerning and demanding eyes in the world. To imply that the images it has thus far created are sub par flies in the face of the evidence. One doesn't have to discredit others to validate one's own product. We already get enough of that with presidential candidates.

 

[Graeme Nattress]

Nope, never seen any of them. The only Genesis footage I've seen was in Panavisions's trailer at Cinegear, and that looked great. The only DPX (and hence ability to scrutinize here) I've seen was from CML and that was not nice, with awful clipped blacks, but I have no idea how that image was made, and hence my comment that it's unrepresentative.

I have not implied anything at all, other than saying the exact words I have written. I am not discrediting the Genesis, as based upon Panavision's own presentation, the image was great. I am saying that the only DPX I have access to from that camera was not nice, and given my previous, but limited exposure to footage, was totally un-representative.

My comment about beauty is saying that of course I like the way that RED's pixels look given my deep involvement with the RAW development of said pixels - in other words, I'm biassed.

Graeme

 

[M Most]

I am not discrediting the Genesis, as based upon Panavision's own presentation, the image was great. I am saying that the only DPX I have access to from that camera was not nice, and given my previous, but limited exposure to footage, was totally un-representative.

My comment about beauty is saying that of course I like the way that RED's pixels look given my deep involvement with the RAW development of said pixels - in other words, I'm biassed.

Fair enough. ;-)

 

[Graeme Nattress]

Thanks for making me explain what I was saying so that it was clear.

Graeme

 

[Chris Parker]

Well, let's hope that cameras that are over 100 grand, and in some cases approaching 200 grand have better images than a 20-thousand dollar camera. If they don't...it's over.

ARRI, Panavision, etc. are not in this market to sell 4,000 cameras. They want to service the highest end of film production.

I have a sneaking suspicion though that RED has changed the game forever. Sure, ARRI will still be there. So will Panavision. I am just not so sure that many of their previous stongholds will continue to shoot exclusively with their products.

Where I am from, almost ALL commercials were shot with ARRIs 5 years ago. Now, MOST of them are shot with ARRIs. Next year at this time, I bet about HALF of them will be shot with ARRIs. The year after (when ARRI's new 'best-ever camera according to them' comes out), many people may simply be accustomed to making commercials on RED, and they will not pay the extra money needed to go back to ARRI for their digital needs.

RED has turned this business upside down, and I think it's absolutely incredible. As long as the REDs don't have a hidden achilles heel that hasn't surfaced yet, I would say they are well on the road to being the leader (or at least one of the leaders) in the digital cinema camera world.

 

[Mark L. Pederson]

ARRI, Panavision, etc. are not in this market to sell 4,000 cameras. They want to service the highest end of film production.

I have a sneaking suspicion though that RED has changed the game forever. Sure, ARRI will still be there. So will Panavision. I am just not so sure that many of their previous stongholds will continue to shoot exclusively with their products.

100% agree.

 

[Steve Sanacore]

This is very funny! If I were the Arri CEO I'm sure I'd *want* to have the best camera too...

Presumably they will unveil something stupidly expensive that does 6K? If you read their newsletters you can see they're already talking about 6K a lot...

Seeing as Clairmont lists the D-20 as $3000 per day (vs the F23 as $3200 per day), I don't think I'll be making much use of an Arri D-30 or whatever they'll call it...

Bruce Allen
www.boaicnema.com

Maybe 6K could replace IMAX, but I thought 4K is already enough resolution for that too. But one thing is for sure, unless ARRI comes up with a super-duper raw compression scheme, what can possibly stream, record and store all that data?

If the industry does go 100% digital, then ARRI will have to come up with a 4K camera at a reasonable price if they want to compete with Sony and Panasonic down the road.

But RED is years ahead of them all right now.

 

[Craig W. Bickerstaff]

Maybe 6K could replace IMAX, but I thought 4K is already enough resolution for that too. But one thing is for sure, unless ARRI comes up with a super-duper raw compression scheme, what can possibly stream, record and store all that data?

If the industry does go 100% digital, then ARRI will have to come up with a 4K camera at a reasonable price if they want to compete with Sony and Panasonic down the road.

But RED is years ahead of them all right now.

Don't get me wrong I think Red is a great camera but I certainly wouldn't agree that it's years ahead of Panavision, Sony or even ARRI.

 

[GlennChan]

Glenn, have you worked with any underexposed (intentionally or otherwise) Red footage? Or some footage from the F23 under similar conditions?

Hmm this is based off what I've seen of the stuff posted on cinematography.net. As Graeme points it, it's not necessarily representative. / indicative of what happens in real situations.

Only if you're talking about the low budget market. I'm not sure I agree with that statement if you're talking about the entire movie-making universe.

Hmm perhaps I am a little cynical about how cheap producers are.

In Canada, there are (low-budget) TV series shot on HDV. Sometimes:
A- Producers will try to "save" money.
B- The image quality of HDV is good enough.

Certainly there are many different tiers in the industry... but I think price will be a factor for many segments of the industry. Even on well-funded American TV series, it seems some productions shoot super 16 instead of 35... even there price matters.

 

[M Most]

Hmm this is based off what I've seen of the stuff posted on cinematography.net. As Graeme points it, it's not necessarily representative. / indicative of what happens in real situations.

I would go further and say it's almost never representative.

Certainly there are many different tiers in the industry... but I think price will be a factor for many segments of the industry. Even on well-funded American TV series, it seems some productions shoot super 16 instead of 35... even there price matters.

True, although there aren't many. And of those, some actually choose 16mm intentionally for either practical (more compact camera size, potentially longer running time per mag) or creative (grittier look) reasons. "Friday Night Lights" is a good example of this.

 

[Dj Joofa]

If you really think about it... I think even there Nyquist has been violated.

At 60 cycles/degree, it doesn't quite seem possible to have zero ringing, little aliasing (*depends on what you're looking at), AND resolve 60 cycles/degree. Theoretically, you can't have all three at once.

In any case it's a moot point... human vision doesn't follow Nyquist.

The actual resolving power of human eye is quite less than 60 cpd. I don't remember the actual numbers but luminance is about close to 9-10 cpd under normal illumination, red-green less than that, and blue-yellow even further less, perhaps as less as 4 cpd according to some studies (if I remember correctly)

Human vision does suffer from aliasing. You can search on the internet for reference images. The sampling (spatial position) of L,M,S photoreceptors (responsible for red, green, and blue dominant signals) is not uniform -- hence, it is kind of stochastic sampling, that is why eye has such excellent noise immunity. Stochastic sampling results in lowering of several objectionable patterns and noise, which to a uniform sampler such as a CMOS/CCD grid would be visible.

A prime example of aliasing / frequency limitation type mechanism going on the eye is that the photoreceptors responsible for blue-dominant signal are much less in number than the red and green ones, whose numbers are pretty close to each other. It is a great design, because of reasons such as diffraction at blue wavelength, the "blue image" is already very blurry, and there is no reason for over-sampling which would have happened if the number of blue cones were the same as red/green.

In addition, one reason we do not see so much of these low-level image processing issues is that a lot of it is filtered by mid and higher-level vision processes. Cognitive vision and perception is a much-explored area but our understanding of this area of vision is not as good as the lower-level processes going on in immediate neighborhood of the retina, which are responsible for image processing issues such as sampling, noise, aliasing, etc.

 

[David Birdy]

I do no know of any producer or production company that is NOT constantly trying to reduce costs, (after 20 years of this I wish I did know a few!)

If we can define and refine a good work flow to go along with this camera Producers will embrace this new product...by the thousands..