Dalsa demo tonight

[David Mullen ASC]

I attended the presentation at Dalsa for the ASC. Saw the mock-up for the new smaller-sized Dalsa Evolution (not a working model) -- was a more reasonable size than the Origin, like an Arri-535 with a 400' mag to put it in rough perspective.

 

Saw some tests from the Origin projected on a 2K DLP Christie projector. Looked very impressive, very clean, nice skintones, like 50 ASA 35mm photography.

 

Also saw the first anamorphic lens they have built to achieve a scope aspect ratio onto a 1.96 : 1 sensor. It was built by Eric Peterson and Dan Sasaki at A&S Precision; both used to work for Panavision.

 

It's a beautiful lens. It has a front anamorphic design, 50mm T/1.4, with a 1.22X squeeze (very mild). The size is similar to a normal 50mm Cooke-S4 or Primo spherical. Minimal distortion & breathing when rack-focusing. Built from elements from a Leitz medium-format lens.

 

They are creating a whole spherical set for the Origin from these Leitz elements, maybe some others, even though only 35mm cine lenses wider-angle than 32mm will vignette on the Origin; it seemed better to build a matching set in all the focal lengths that projected an image that definitely cleared the edges of the sensor.

 

Eric said that he wanted to build complete sets with a contrast that would be better-suited for the Origin's digital sensor, i.e. less contrasty than Zeiss.

[Lance Flores]

I attended the presentation at Dalsa for the ASC. Saw the mock-up for the new smaller-sized Dalsa Evolution (not a working model) -- was a more reasonable size than the Origin, like an Arri-535 with a 400' mag to put it in rough perspective.

 

Saw some tests from the Origin projected on a 2K DLP Christie projector. Looked very impressive, very clean, nice skintones, like 50 ASA 35mm photography.

 

Also saw the first anamorphic lens they have built to achieve a scope aspect ratio onto a 1.96 : 1 sensor. It was built by Eric Peterson and Dan Sasaki at A&S Precision; both used to work for Panavision.

 

It's a beautiful lens. It has a front anamorphic design, 50mm T/1.4, with a 1.22X squeeze (very mild). The size is similar to a normal 50mm Cooke-S4 or Primo spherical. Minimal distortion & breathing when rack-focusing. Built from elements from a Leitz medium-format lens.

 

They are creating a whole spherical set for the Origin from these Leitz elements, maybe some others, even though only 35mm cine lenses wider-angle than 32mm will vignette on the Origin; it seemed better to build a matching set in all the focal lengths that projected an image that definitely cleared the edges of the sensor.

 

Eric said that he wanted to build complete sets with a contrast that would be better-suited for the Origin's digital sensor, i.e. less contrasty than Zeiss.

 

I just got off the phone with Ron Vidor when your post came up. He got my message late and missed the presentation. But one of the things we were discussing was the monstrous size of the Dalsa. Ron's doing all our stedicam work and 2nd Unit DoP so he is concerned. Dalsa said they would halve the smaller version ready for our shoot. The test data I have from the camera looks great. Did the announce availability for the smaller format? I'm wondering if they will have it ready for us.

[David Mullen ASC]

Did the announce availability for the smaller format? I'm wondering if they will have it ready for us.

 

Sorry, but it sounded to me like they said "sometime next year". I could have misheard them. I was thinking though that the upcoming RED camera would probably make a good Steadicam camera (smaller size & weight, plus the lack of optical viewfinder not an issue), with the Dalsa as the studio "A" camera.

 

Or use a 35mm film camera for the Steadicam...

 

Or live with 2K for the Steadicam shots and use the Phantom HD. Or use an HD camera like the Viper.

[Lance Flores]

Or use a 35mm film camera for the Steadicam...

 

Or live with 2K for the Steadicam shots and use the Phantom HD. Or use an HD camera like the Viper.

 

I thought it might go that way. Ron just finished a shoot with Arri digital. He liked it. Thought it did a good job. I done any evaluation with it. have no data. Good idea .. could use the Phantom 2K. It would be mostly used for the INT. portion of night raid scenes. It's got great sensitivity at low light and virtually no noise. Thanks David.

Edited April 11, 2007 by Lance Flores

[David Mullen ASC]

I thought it might go that way. Ron just finished a shoot withth Arri digital. He liked it.

 

Yes, RAW data mode from the Arri-D20 may be another option; it's not too big of a camera. You may have to live with being tethered unless there is a flash-mag option (which I think there is.)

[Eric Steelberg ASC]

Wait...what the hell? Dan Sasaki left Panavision?! When the hell did that happen?

[Kevin Zanit]

Yeah, he did, left a few weeks ago.

 

I used to know Eric as well, but he's been gone from Panavision for some time. I was at Panavision on Dan's second to last day, it was weird. I was also there today, seeing his empty work station was even weirder.

 

He told me he was going freelance with Eric, I expect to see some really good stuff come from them.

 

Kevin Zanit

[Max Jacoby]

It's a beautiful lens. It has a front anamorphic design, 50mm T/1.4, with a 1.22X squeeze (very mild). The size is similar to a normal 50mm Cooke-S4 or Primo spherical. Minimal distortion & breathing when rack-focusing. Built from elements from a Leitz medium-format lens.

Are you sure it's build from a medium format lens and not a regular 35mm stills lens? I am not aware of any Leitz medium format lenses, but on the other hand Dalsa have already adapted 35mm stills lenses for their camera because they cover the whole sensor.

 

Do you know if they're going to offer a 1.22 desqueeze in the optical viewfinder or leave it as it is? Since the squeeze favtor is so small I can already see actors demand that their close-ups be used unzqueezed so that they can look a bit slimmer on screen...

[Lance Flores]

Since the squeeze favtor is so small I can already see actors demand that their close-ups be used unzqueezed so that they can look a bit slimmer on screen...

Max you've just been looking in the wrong places. You confuse them with tele-centric and lens-centric. Look for the boxes labeled egocentric.

[David Mullen ASC]

Are you sure it's build from a medium format lens and not a regular 35mm stills lens? I am not aware of any Leitz medium format lenses, but on the other hand Dalsa have already adapted 35mm stills lenses for their camera because they cover the whole sensor.

 

Do you know if they're going to offer a 1.22 desqueeze in the optical viewfinder or leave it as it is? Since the squeeze favtor is so small I can already see actors demand that their close-ups be used unzqueezed so that they can look a bit slimmer on screen...

 

Yes, you're probably right since a 35mm still camera lens would cover 8-perf 35mm, and thus would cover the Dalsa sensor. Anyway, I think they only used certain elements from Leitz -- it seems to me that they are basically making a version of a Primo spherical but somehow managed to keep the anamorphic version from getting large.

 

They are working on the de-squeezer for the viewfinder; Rob Hummel has basically pushed the whole anamorphic approach for scope images but last night when I asked him about the viewfinder, he said "yes, it didn't occur to me until later that we'd have to do that, I was so excited about making the anamorphic lenses!"

 

Someone told me that Panavision in the 1960's secretly made an anamorphic lens with slightly more than a 2X squeeze for shooting Elizabeth Taylor's medium shots, to slim her down. Sounds like it could be a popular idea...

[Max Jacoby]

Someone told me that Panavision in the 1960's secretly made an anamorphic lens with slightly more than a 2X squeeze for shooting Elizabeth Taylor's medium shots, to slim her down. Sounds like it could be a popular idea...

I think there's probably quite a bit of money to be made with the development of anamorphic lenses where one can vary the sqeeze factor. Actors would ask for these lenses every time!

[Phil Rhodes]

If a lens was "less contrasty", wouldn't that just imply that it was diffusing light into shadow areas like a mist or fog filter, and wouldn't that be a negative impact on resolution?

 

If it were possible to build a lens which had reduced contrast with no other side effects, we'd expect to have perfect contrast-reduction filters with zero resolution loss - and we don't.

 

Phil

[David Mullen ASC]

If a lens was "less contrasty", wouldn't that just imply that it was diffusing light into shadow areas like a mist or fog filter, and wouldn't that be a negative impact on resolution?

 

If it were possible to build a lens which had reduced contrast with no other side effects, we'd expect to have perfect contrast-reduction filters with zero resolution loss - and we don't.

 

Phil

 

We're talking about a narrow range of contrast difference, like how Cooke-S4's are less contrasty than Zeiss Master Primes, with Primos (and these Dalsa lenses probably) being somewhere in-between. It's not like putting in a mist filter. I don't think any lens has perfect flare resistence with zero loss of contrast anyway -- we're talking about light travelling through multiple pieces of glass afterall.

 

Yes, the more contrasty a lens is, the sharper it tends to look. But some people would rather have a lens with a more "pleasing" contrast rather than something super-snappy. Lens-making is more than a technical science, there's a bit of an art to it, like making a violin.

[Sam Wells]

I'm way over my head here but it seems to me if you take a very aggressive aproach to correcting spherical aberration, using aspheric elements etc you can acheive a very contrasty and fine sharp focus which will be delineated by a fast fall off - Zeiss style - whereas a less agressive aproach could give you a more balanced contrast between higher and the middle spatial frequencies. You could find faces and even overall scenes benefitting.

 

This is nothing new, but I don't know enough about AA filtering to speculate further on these choices.

 

-Sam

[Rupe Whiteman]

Are you sure it's build from a medium format lens and not a regular 35mm stills lens? I am not aware of any Leitz medium format lenses, but on the other hand Dalsa have already adapted 35mm stills lenses for their camera because they cover the whole sensor.

 

... Zeiss have been making Hasselblad lenses for years... Perhaps he meant these?

 

Rupe

[Max Jacoby]

Dalsa have adapted both Leica/Leitz and Nikon 35mm stills lenses for use on their cameras. Unlike the ones that David mentioned these are not anamorphic.

[Jarin Blaschke]

This is leaving the topic, but Hasselblad's lenses for the last few years (H1 era) have actually been made by Fuji.

 

 

... Zeiss have been making Hasselblad lenses for years... Perhaps he meant these?

 

Rupe

[Arni Heimir]

Sorry, but it sounded to me like they said "sometime next year". I could have misheard them. I was thinking though that the upcoming RED camera would probably make a good Steadicam camera (smaller size & weight, plus the lack of optical viewfinder not an issue), with the Dalsa as the studio "A" camera.

 

Or use a 35mm film camera for the Steadicam...

 

Or live with 2K for the Steadicam shots and use the Phantom HD. Or use an HD camera like the Viper.

 

Wouldn't mixing digital cameras be like using two different stocks?

[Lance Flores]

Wouldn't mixing digital cameras be like using two different stocks?

In essence, yes. What we do is run a camera through characterization and normalize the data to the target profile, let's say the profile of the Dalsa Origin. Then we build correlation a database or algorithms. Then we decide on the profile from one camera with all the LUT etc. data as the target then extrapolate the data captured from any of the client cameras. The big problem is on the storage of the data. Once your sure you've got the interpolation correct you can archive the original. So basically what we would do is to take, for example a 2K image, the data and interpolate the chroma, etc information and then extrapolate it into the target 4K image applying various algorithms to adjust edges so they have the right contrast, and all that other silly stuff.

 

So what you end up doing is building meta files so that you don't have to generate large temporary intermediate files .. then you send command lines and metadata to a render farm and it does all the work. What you end up with is a file that virtually matches the 4K images from your target 4K camera model. You then archive the original capture. Sound complicated .. guess it is ... but after the first few runs you no longer need to go through the whole process. Once you build all the translation information you just simply submit the file location of the target data to the render engine with the appropriate command file.

Edited April 11, 2007 by Lance Flores

[David Mullen ASC]

Wouldn't mixing digital cameras be like using two different stocks?

 

Don't movies do that all the time? Within the same scene, yes, it requires more work in post color-correction to match as well as possible, but there are often other mismatching factors as well anyway, like a zoom lens shot mixed with primes, an underexposed shot mixed with normal exposure, changes in sun and shade, etc.

 

So while it is not recommended, sometimes the practical advantage outweighs the issue of perfect matching of quality.

[Illya Friedman]

They are creating a whole spherical set for the Origin from these Leitz elements, maybe some others, even though only 35mm cine lenses wider-angle than 32mm will vignette on the Origin; it seemed better to build a matching set in all the focal lengths that projected an image that definitely cleared the edges of the sensor.

 

Thanks for the wrap up David, it was good to see you again last night.

 

Something that I don't think was mentioned in the meeting (and most people don't immediately catch) is that because the Palomar sensor is slightly wider than a 35mm Academy 1.85 frame the sensor has an LMF of 1.25. In other words, to find the equivalent spherical 35mm lens horizontal field of view on the Origin; simply multiply lens the focal length by 1.25.

 

Any lens you put on the Origin already has a wider horizontal FOV when compared to it working on a 35mm film camera. Here's an example, for the same horizontal field of view as a 20mm lens on a 35mm camera, you would use a lens with a 25mm focal length on the Origin. (20mm x LMF (1.25) = 25mm, 100 x 1.25 = 125mm etc.) Currently the widest full coverage lens is the 19mm Dalsa L Series prime. On the Origin it's about the same FOV as a 15mm lens on a 35mm camera.

 

Illya Friedman

Dalsa

illya.friedman@dalsa.com