DALSA Evolution

[Ilmari Reitmaa]

Copy-pasting from a DALSA news release:

 

DALSA Digital Cinema reinforced its leadership in 4K motion picture capture today with the announcement of new 4K camera models, an on-board 4K data recorder and new 4K anamorphic lenses. The new products will be on display at the upcoming National Association of Broadcasters (NAB) trade show (booth # C-9423), April 16th to 19th, 2007 in Las Vegas.

 

"Basically, we?re talking about the ability to shoot at 4K resolution, 16-bit, uncompressed, untethered, using the highest quality anamorphic lenses. I think cinematographers will be particularly thrilled that, for the first time, a digital camera will be able to capture the CinemaScope 2.40:1 aspect ratio without compromising image quality."

 

Available "early 2008".

[Andrew Ray]

new 4K anamorphic lenses.

 

Does it mean that there are no good lenses out there that can work with the full 4k format?

 

Are there any lenses that can resolve 400Lp/mm and could be used with the cine type camera?

 

Andrew

Edited April 10, 2007 by AndrewR

[David Mullen ASC]

I don't know if the new Dalsa camera solved this problem, but they had this problem that the sensor was a little larger than Super-35 and thus some 35mm cine lenses would vignette on it.

 

I'm glad to hear that Dan Sasaki is designing anamorphic lenses for the camera, although with a 2x1 sensor, the Dalsa was the one camera that I thought didn't really need anamorphic lenses, since you could either crop a little to get scope (and also deal with any vignetting in the corners) or crop a little the other way to get 1.85. Maybe they changed the shape of the sensor too, to something closer to 16x9.

 

I get to go to a Dalsa presentation tonight so I'll learn more.

[John Holland]

So are they doing a presention in L.A prior to NAB ?

[David Mullen ASC]

So are they doing a presention in L.A prior to NAB ?

 

Yes, tonight for the ASC over at the Dalsa headquarters.

[John Holland]

That sounds very interesting , be nice to hear what you learn .

[Lance Flores]

So are they doing a presention in L.A prior to NAB ?

I received a vip invitation to one so I would say yes. I've settled on the Dalsa since no one else is ready for a 4K feature production with a camera directly compatible with our workflow and provides what I want for the capture.

[Max Jacoby]

David,

 

Obviously the word 'anamorphic' makes me salivate already :P

 

I'm sure you'll ask what the squeeze factor of these lenses is. Since the chip is 2:1 one would only need a 1.2 squeeze, which would yield a look closer to spherical than anamorphic. Also is the anamorphot at the front or at the back? And finally what focal lenghts, size and speed these lenses come in.

 

Thanks

[Lance Flores]

David,

 

Obviously the word 'anamorphic' makes me salivate already :P

 

I'm sure you'll ask what the squeeze factor of these lenses is. Since the chip is 2:1 one would only need a 1.2 squeeze, which would yield a look closer to spherical than anamorphic. Also is the anamorphot at the front or at the back? And finally what focal lenghts, size and speed these lenses come in.

 

Thanks

 

List of lenses available:

 

List of Lenses

[Brian Drysdale]

List of lenses available:

 

List of Lenses

 

Unfortunately that's their current spherical lens rental list.

 

If new lenses have such a small squeeze factor, they'll lose one of the nice side effects of the current anamorphic lenses: a long focal length for a particular horizontal angle of view.

[Lance Flores]

Unfortunately that's their current spherical lens rental list.

 

If new lenses have such a small squeeze factor, they'll lose one of the nice side effects of the current anamorphic lenses: a long focal length for a particular horizontal angle of view.

 

Yes. I'm not too concerned. I'd rather not soften things and can compose the shots to work for the story. As we work through previz we can work out feel we need with the DoP and 2nd Unit DoP/stdycm. We'll have lenses made for us if need be.

Edited April 10, 2007 by Lance Flores

[John Holland]

Soften things ?

[Andrew Ray]

Soften things ?

 

Gentleman, as you see from post number, I am new here.

I do underwater and cave exploring and filming.

We would like to move to 4K format this year and from our limited experience we see that there are no lenses on the popular market that will have resolving power to support 4K sensor, that is about 200 photo sensing sites per mm. Equivalent of 100Lp/mm

The lenses have to have minimum double of that so 90% of resolving power will be preserved. We, as well as most of people here on this forum, are new to 4K.

Can anybody help us with good advice on this subject, please.

 

Andrew

[Brian Drysdale]

Gentleman, as you see from post number, I am new here.

I do underwater and cave exploring and filming.

We would like to move to 4K format this year and from our limited experience we see that there are no lenses on the popular market that will have resolving power to support 4K sensor, that is about 200 photo sensing sites per mm. Equivalent of 100Lp/mm

The lenses have to have minimum double of that so 90% of resolving power will be preserved. We, as well as most of people here on this forum, are new to 4K.

Can anybody help us with good advice on this subject, please.

 

Andrew

 

Perhaps you should contact a motion picture lens manufacturer like Cooke who will might be able to give you more details.

 

http://www.cookeoptics.com/

 

Also Zeiss cine/motion picture lenses.

 

http://www.zeiss.com/photo

 

You'll also find details of the Zeiss lenses here:

 

http://www.arri.com

 

Under cameras - lenses & accessories

 

Dalsa list these companies lenses as part of their 4k lens rental list.

[Andrew Ray]

Dalsa list these companies lenses as part of their 4k lens rental list.

 

Thanks! Brian.

We tested Zeiss UP and MP and Cooke S4 and anything below 65 do not have sufficient 200Lp/mm resolution away from the center of the frame. Also chromatic aberration is limiting the resolution to less then 100Lp/mm except the very center of the frame.

I was thinking about some lenses that are designed for 70mm format and the adapter to PL mount, along this lines?

[Brian Drysdale]

I was thinking about some lenses that are designed for 70mm format and the adapter to PL mount, along this lines?

 

The 65 mm shooting format lenses wouldn't be as modern as the 35mm lenses.

[Max Jacoby]

Well the Master Primes are the sharpest lenses available anywhere. 65mm lenses are less sharp, because of the increased size of the negative compared to 35mm does not require such high optical performance. The Arri 65mm lenses are all rehoused Zeiss medium format stills lenses released in the early 90s.

[K Borowski]

As someone that shoots 6x7 cm film almost every week, let me tell you, you're worrying too much about edge sharpness falloff. As is, that happens with all lenses. As long as you have a lens that has enough coverage for the full frame, the jump in resolution is going to be dramatic. People's eyes have resolution falloffs aroudn the edges too, you know, even 35mm film is going to have falloff around the edges, not drastic, but definitely there. I know Zeiss is sharp glass, but it's not THAT much better than every type of lens everyone else makes.

 

Even Zeiss is going to have distortions and aberrations. We live in the real world after all. *Nothing* is perfectly formed, that is impossible, especially when it comes to lenses.

 

I think that you should look to the movie equivalent of 6x7, which would probably be 15-perf 65mm to see how they deal with loss of resolution around the edges. My personal bet on what they'll say, is that they ignore it. This isn't going to be noticeable vignetting like in the first silent movies, unless your lens doesn't cover the whole sensor.

 

Frankly, with digital, you should be more worried abou the light from the rear element of the lens not hitting the sensor from a perfect normal angle. Otherwise it causes all kinds of aweful distortion. I'm sure the current line of 16mm-sized HD lenses and 35mm-full frame digital sensor-compatible lenses have been redesigned with a flattened rear element, but be very cautious before you go slapping a large format film lens in front of a digital sensor. You might not like what you see. . .

 

Also, let's just get this straight, because lp/mm is probably the most misunderstood and abused principle in all of imaging: The resolution of a SYSTEM is inversely equal to the inverse of the sums of each of its component elements:

 

So we have a frame of 4-perf Super 35. I'm going to round, and use English Units where I see fit, hope you don't mind:

 

I have a sharp sharp Supersharp Superexpensive lens that I bought, resolving 200 lp/mm, right off the bat, in line pairs/inch that correlates to 5080 lp/in. I have a film that resolves your typical 63 lp/mm in average real-world lighting conditions (note that digital also suffers from a loss of contrast when not shooting test charts, but I am not going to go into that issue here. I have no desire to start a ten-page long debate on the merits and demerits of either medium today). 63 lp/mm is equivalent to 1600 lp/in. Let's just assume that the Super 35mm 4-perf frame is close enough to 1 in by 3/4 of an in. so it will make my calculations much easier. The total area of the frame then is 3/4 of a square inch. That is the area upon which waves of light are focused by the lens.

 

Therefore the inverse of the resolution of the master negative (resolution sub total) (not including optical printing, contact printing, scanning, film-out, shrinkage during processing of acetate stocks, etc.):

 

1/(res.-sub-tot.) = 1/(5080lp/in.) + 1/(1600 lp/in.)

 

res.-sub-tot. = 1216.7665 lp/in.

 

This is a 2-D measurement, so we have to take both the length and height

of the frame, respectively, and multiply each of these dimensions by the

corresponding number of line pairs per inch of 2D dimension, then multiply

the two resultant products together:

 

(1216.7665 lp/in. x 1 in.) x (1216.7665 lp/in. x 0.75 in.) =

(1216.7665 lp width) x (912.575 lp height) = 1110390.507 lp^2 frame area

 

Since a line pair is two lines, and two lines times two lines equals four lines,

multiplying this area by four gives us 4441562 and change "waves"

of resolution. This is only about 4.4 megapixels equivalent worth

of information that can be resolved on the S35 4 perf frame,

which would, assuming greater sweetspot and lens coverage would

correspond to 8.9 megapixels on a frame of Vistavision. Again, if you

increase sweet spot, stop down two from wide open, and have enough

effective coverage, the f-stop acts to limit lens falloff around the edges.

Stop down anymore than two stops and you run into aberrations from light

glancing off the f-stop blades, which would cause another type of distortion.

 

Note that the 200lp/mm lens still takes 25% of the resolution away from the film, so your statement that you'd retain 90% of the information on the sensor is incorrect. It would be even lower with 1- or 200 lp/mm sensor sites. Honestly, I don't think that those sites can possibly resolve all 200 sites unless you're shooting black lines on white boards illuminated by studio lights. At 1000:1 contrast, you'll get resolutions around 1-300lp/mm with modern slow color negative films. It's not film that is the limiting factor here, or digital, it is and always will be the lens, regardless of the size. Note that 5 perf 65mm is about 2 1/6 in. x 9/10 in. (39/20 in.^2) This is 2.6 times larger than the area of the 4 perf Super 35 frame, so, as long as your new lens covers the whole area of the negative, it can actually have far WORSE resolving power, because the neg area has increased the resolution othewise to 260% of the 35mm negative. It is often the case that still photographers shoot with 100-year old lenses on their 8x10 inch or larger view cameras. The area of the negative is so huge that you could shoot with the lens Daguerre was shooting and still outresolve a DSLR or 35mm.

 

This is assuming that every beam of light that hits the lens is somehow in sharp focus all at once. We needn't worry about F-stops and chromatic aberrations right now though, they'll come into play later, though it's too complicated to sample calculations without specific data from a lens; even among the same model aberrations and distortions will vary, significantly from lens to lens. The only way around this problem is to basically act like you're Stanley Kubrick when you go to the rental house and individually check and test EACH of their lenses with resolution charts (again keeping in mind that resolution at chart contrast can be more than an entire order of magnitude greater than the resolution of the lens with real-world contrasts you'll encounter when actually shooting)

 

I hope this makes sense. I tried to make it coherent in the time I had to write it.

 

Take care,

 

~KB

[Lance Flores]

Thanks! Brian.

We tested Zeiss UP and MP and Cooke S4 and anything below 65 do not have sufficient 200Lp/mm resolution away from the center of the frame. Also chromatic aberration is limiting the resolution to less then 100Lp/mm except the very center of the frame.

I was thinking about some lenses that are designed for 70mm format and the adapter to PL mount, along this lines?

 

Andrew - you are exactly on point. I've working through similar issues. I'm having three complete sets of 70mm lenses made.

[Max Jacoby]

Andrew - you are exactly on point. I've working through similar issues. I'm having three complete sets of 70mm lenses made.

By whom? You are taking medium format lenses I believe?

[Brian Drysdale]

Andrew - you are exactly on point. I've working through similar issues. I'm having three complete sets of 70mm lenses made.

 

Seems like we're talking about film having three layers where as the single sensor camera has only one. Therefore any edge drop off problems become more apparent compared to film.

[K Borowski]

So few people are shooting 70mm these days, I haven't ever heard of anyone call medium format camera lenses "70mm camera lenses". I'd say 70mm is even more a dinosaur than 65mm is in the cinema. That being said, I love 70mm.

 

I'm sorry if my last post(s, since my first goof up one hasn't been deleted) too technical for some of you to digest. In layman's terms, I'm saying that the size of the sensor makes a thousand times more difference than the lens you're using. And I'm sorry, still photographers shooting medium format aren't so flipping clueless that they go around using lenses with horrible vignetting, distortion, and light falloff around the edges. Be reasonable folks, medium format lens manufactures, and not just Contax/Zeiss, make some damned fine lenses, even lowly Pentax or Mamiya. And honestly, there's been article after article in professional photography magazines comparing Zeiss with the "cheap" Mamiya and Bronica, and Pentax glass (I like to bitch about still photographers doing nothign but counting megapixels these days, this is the more worthy activity we indulged in before that aweful transition ;-) ) would you really be surprised if time after time the most distinguishing pros couldn't tell the difference?

 

I'd say the only thing you really have to worry about with adapting MF still lenses to a 4K rig would be overriding the Iris shutter and rear element flatness. You'll get far more ugly effects from light hitting a digital sensor at a non-normal to the plane of the chip, than you'll get from "lack of resolution" in a lens.

 

Frankly, even the best best lenses, the ones that only show up say in 1 out of 100 Zeiss lenses manufactured, will probably cause at least 10% loss of image information they pass through them.

 

I was trying to figure out how high of a resolving power I'd need to get a slow slow slide film to render a 35mm 8-perf still frame that truely contained 16 million resolved points lines of visual information, and I think it took an insanely-high number, nearing *400* lp/mm to hit that mark.

 

I think a lot of the misconceptions about resolution come from the fact that while someone is shooting with a 12 MP digital camera, I doubt it is resolving more than 8MP of information, as it suffers from the same rules of physics as film cameras do.

[K Borowski]

Seems like we're talking about film having three layers where as the single sensor camera has only one. Therefore any edge drop off problems become more apparent compared to film.

 

For the *third* time, this has nothing to do with "drop off problems" it has to do with the rear lens element not being optimized for digital. Some digital sensors are defective when it comes to registering light, so lens designers had to pick up the slack with redesigned rear lens elements that straighten out the path of the light before it hits the sensor. BTW, this actually introduces more optical distortion, because you're bending the rays of light more.

 

I've seen first-hand how this manifests itself. This was Nikon and Canon equipment, all high-end film glass, and it was as if there were purple ghosts dancing around the edges of the picture. You'll get weird pink or purple ghosting near the edges of the frame, often splotchiness. Forget it if you're shooting a shot where there're drops of water on the front of the filter or falling between you and the subject. Images that suffer from this phenomenon look attrocious, but a scene from a movie would almost certainly be totally unacceptable.

 

If you're smart, you'll go with Hasselblad's H2 system lenses, as they're all compatible with digital backs. Pentax & Mamiya's 645 systems also have been optimized for digital. I have no idea how big the Dalsa sensor is. I looked at their announcement and didn't see the specs. 645 is about 58 mm by 42mm, so you'd get a radius of coverage extending 58mm to before you get falloff. Mamiya's 6x7 line, the Mamiya RZ series, would probably be better for a larger sensor. I know there are several backs that are compatible with thsi camera, so it is another likely candidate for having lenses that don't accentuate the angled light flaw too much.

[Brian Drysdale]

For the *third* time, this has nothing to do with "drop off problems" it has to do with the rear lens element not being optimized for digital. Some digital sensors are defective when it comes to registering light, so lens designers had to pick up the slack with redesigned rear lens elements that straighten out the path of the light before it hits the sensor. BTW, this actually introduces more optical distortion, because you're bending the rays of light more.

 

Just acting dumb here: this effect come down to the difference between how film operates with light sensitive particles suspended in a layer, which can pick up light energy coming in off the normal angle and the "flat surface" sensor, which has a drop off in sensitivity the further the light hits it off the normal.

[K Borowski]

Just acting dumb here: this effect come down to the difference between how film operates with light sensitive particles suspended in a layer, which can pick up light energy coming in off the normal angle and the "flat surface" sensor, which has a drop off in sensitivity the further the light hits it off the normal.

 

Correct! Now we don't have any problems :-) I thought you were intentionally ignoring me when you kept listing this symptom. Well, it's actually more complicated, because all lenses are brighter in the center and get dimmer as you move out to the edge, with digital, film, or Daguerrotype plate.

 

This is why you don't want to shoot lenses wide open when you have the stop down one or two stops from wide open. I forget the Physics behind it at the moment, but it works. I try to print all of my pictures with the enlarge lens stopped down two or three, at the very least one stop. Some digital sensors exhibit further problems though, with weird color shifts and artifacts at the edges, which go beyond just vignetting toward the edges.

 

In any case, I know I got really technical in my first reply to this thread, but if you read through that whole post, as long as it may be, it should give you a good understanding of the other misconceptions on resolution and optics that plague photographers and moviemakers. When you actually realize how the formula works, basically taking the inverses, adding them all together, and then inverting again for your answer, it's kind of like a lightbulb going on. So the concept of system resolution is something you either do get or don't get.

 

To put it as simply as possible though: the advantages of large format acquisition, regardless of the lens you use, will almost always produce results that are far superior to the best results you could get from 35mm. Just don't shoot with your lens wide open in a snowstorm and crank the compression algorhytms up really high.