Dragon Lens Coverage Information

[Asif Limbada]

There's something odd with the numbers on the Leicas. Phil's records don't seem to match this statement by Jim, and this statement by Jarred.

Phil, could you clarify please.

I second that...its kinda weird. From what I know is that Leica's have 34mm+ Image Circle. Wish I had a DRAGON to check it.

 

[Phil Holland]

Okay. Back home at the ranch now and I can answer several questions.


On FF35 still glass - Like it says in the first post, anything designed for that format will cover Epic Dragon at 6K no problem. Leica Rs, Canon EF, Nikons, Olympus, etc....

On the Tokina/Duclos 11-16mm - Now that's worth testing out. I have a stock Tokina and it does indeed cover FF35 at 16mm. I'll add that to the list of zooms to test. The Ruby 14-24mm should be tossed in too even though it should cover 6K Full. Maybe something happens in the conversion process that might change this. I'll take a look.

On Cooke 5/i Primes - We just weren't able to get a set of those on this particular day of testing. Those, the Kowas, K35s, vintage primes, etc... There's certainly more to test.

On Cooks Mini S4 Primes - Also didn't have these for this particular set. There was a brief moment at Duclos Lenses where Matthew had these months ago and he mentioned that they had very impressive coverage. We'll have to look at those for sure.

On Zoom Tests - The plan is to test zooms next.

On the accuracy of the Leica Summilux-C numbers - Yes. This is the correct information. There was no tequila involved during the test and there were other eyes around who verified these findings. It both reflects my individual measurements made a bit ago and also what visible coverage on Dragon represents. Most of these sets have been tested twice for coverage in reality. I am not error.

 

[Patrick Grossien]

Thank you Phil for the update. And thank you for the great write up as usual

Sorry to be poking further, but since there's the difference in the statements, what is it the Leicas do at 6K from your findings? Do you get hard edges at 6K or is it vignetting?

 

[Will Keir]

Seems the same to me. From our SALT III, LECIA did a better job than MP. Now it's the opposite?

Phil, are you sure you didn't accidentally mix the Leica's with the Master Primes? Jarred has said the Leica's cover fully down to the 16mm which just slightly vignettes at 6K with the rest fully covering.

 

[Phil Holland]

Also, to be clear on the Leica Summilux-C Primes. I have spoken to the engineers at Leica regarding these lenses, as I have an unhealthy addiction towards knowledge on glass. The LSCs were designed with an image circle of 31.5mm in mind. Which is essentially what Mysterium-X at 5K Full is and what Epic Dragon 5.5K is. And they do in fact cover that format nicely.

Here are my "expanded Leica Summilux-C Notes".

18mm - 6K Wide Open usable w/ vignette until T5.6.
21mm - 6K Wide Open usable w/ vignette until T5.6.
25mm - Uneven corner illumination at 6K wide open.
35mm - 6K usable Wide Open but vignettes
40mm - 6K usable Wide Open but vignettes
50mm - Covers 6K, vignettes when iris at T22 and focused past 9 feet.
75mm - Covers all Epic Dragon Formats
100mm - Covers all Epic Dragon Formats

What can be discerned from this? In one way or another they are usable wide open until about T4 on Dragon at 6K and 6K WS. However, the initial graphics involve "clean coverage". This is why Jim is right and the numbers are right. It is a very noticeable vignette roll off however. T1.4-T2.8 and up until T4 you may find success and lovely images.


Some crop factor numbers for y'all.

Okay. So we have to keep in mind relative crop factor here. Meaning if FF35 is your starting point that's 1X, and if S35 is your starting point that's 1X. Here's the info for both with the aspect ratio of Dragon's sensor taken into account:

For people who visualize FF35 as their starting point......

Epic Dragon Crop Factors FF35 Relative to Full Frame 35mm:
6K Full = 1.17x
5.5K = 1.29x
5K = 1.40x
4K = 1.75x
3K = 2.33x
2K = 3.50x
1K = 5.25x

and for S35 as a base.....

Epic Dragon Crop Factors S35 Relative to Super 35mm (Dragon 5K):
6K Full = 0.84x
5.5K = 0.92x
5K = Base = 1x
4K = 1.25x
3K = 1.67x
2K = 2.50x
1K = 3.75x

Now this only reflects the size around the 1.94:1 and 1.9:1 aspect ratios and HD formats are not listed here. This is just a simple starting point to see how your given focal lengths FOV changes across various formats.

Enjoy.

 

[Ketch Rossi]

Phil,

since you are so good at numbers... Can you calculate the coverage of the Medium Format Leica S series?

I have been contemplating something with this lenses for the past few years, and Jarred is aware of my likes for this glass, just don't have the patience to calculate as you do...

 

[Will Keir]

there were other eyes around who verified these findings... I am not error.

Awesome that you took it upon yourself to get this information out, but I feel this is also the issue with these results. Who else was there?

I really liked the SALT format, people to double check findings and lend a critical eye. It's a more valid test when you have 3-4 DPs, directors and lens experts working collaboratively to stress test a lens. You also get 5-6 opinions to sort through to find your own truth. The "closed door" thing just turns people away.

 

[Phil Holland]

Awesome that you took it upon yourself to get this information out, but I feel this is also the issue with these results. Who else was there?

I really liked the SALT format, people to double check findings and lend a critical eye. It's a more valid test when you have 3-4 DPs, directors and lens experts working collaboratively to stress test a lens. You also get 5-6 opinions to sort through to find your own truth. The "closed door" thing just turns people away.

To be clear and 100% transparent this was an undertaking done with myself and Red at Red Studios Hollywood. And several Red folk were there making notes, helping out, and making observations as well. We had people double checking with critical eyes. There is no "my own truth" there is only the image in the frame and numbers. This was a test done to share information regarding Dragon coverage for everybody in the motion picture community and was conducted as such.

Also, I have been doing lens tests professionally for over ten years on feature productions, but I guess that's not the point really.

And if you recall as Matthew said during our last test there's every reason in the world to have a concise mind and eye applied to the testing of glass by an individual. Which is why discerning tests such as resolving power, field performance, performance across the entire focus range, at various iris settings, etc... etc... etc...

 

[Matt Ryan]

To be clear and 100% transparent this was an undertaking done with myself and Red at Red Studios Hollywood. And several Red folk were there making notes, helping out, and making observations as well. We had people double checking with critical eyes. There is no "my own truth" there is only the image in the frame and numbers. This was a test done to share information regarding Dragon coverage for everybody in the motion picture community and was conducted as such.

Also, I have been doing lens tests professionally for over ten years on feature productions, but I guess that's not the point really.

And if you recall as Matthew said during our last test there's every reason in the world to have a concise mind and eye applied to the testing of glass by an individual. Which is why discerning tests such as resolving power, field performance, performance across the entire focus range, at various iris settings, etc... etc... etc...

I understand your findings, but what I cannot wrap my head around is on a projector at Duclos for the SALT test, the Leica's had greater coverage than the MP's especially on the wide end. So that's where my confusion stems from... It's like the opposite now.

 

[Phil Holland]

I understand your findings, but what I cannot wrap my head around is on a projector at Duclos for the SALT test, the Leica's had greater coverage than the MP's especially on the wide end. So that's where my confusion stems from... It's like the opposite now.

It was made clear that by Matthew himself that that wasn't the best way to test for coverage. And remember we tested wide open and occasionally stopped down to T2.8 when those lenses were on the projector. Wide open the Leica's all projected the target "cleanly" with a bit of fall off. And if you remember correctly the Summilux-C 75mm was the one we were all impressed by it's image circle projecting that target. Which not only covers, but covers through the iris range. Great lens IMO.

There's a variety of ways to test coverage with several devices. This particular test is all about how these lenses perform on Dragon. It's important to toss the lenses on this camera as there's various things that could come into play that a simple measurement may not provide. Also very important to see the image in the frame itself.

 

[felipe meneghel]

Thanks guys!
And keep "discussing"!

 

[Patrick Grossien]

Thank you Phil for taking the time and effort! The expanded notes really help !

Really appreciate all your extensive work and openness ! Thank you!

 

[charles lim yi yong]

Hey Phil, the larger the sensor. The better the resolving power for all lenses across the board?
Does it mean that for the anamorphic, it's possible wider apertures?

 

[Jarred Land]

Just remember guys, " Slight Vignette " means different things to different people. All lenses have transmission falloff from center to edge. Some have more , some have less. So when Phil says the Leica's vignette at 6K WS, his definition may be different from yours.

For example, here is a Dragon with a dirty sensor and the Leica 25mm at 6K WS at about T32 ( or whatever it is right before the Leica's closing iris snaps shut ) pointing at a lightbox. Sure there is fall off to the corners, but you likely would never shoot at this stop ( the lenses dont even go past to T22 ) and the fall off is barely noticable. At 6K full frame, you will see some harder edges on some lenses in the corners if you close down as Phil said, But I just want to make sure when he says " Usable at 6K with slight vignette " the translation is such. Phil isn't technically wrong.. he is being super clinical ( as he should be).

 

[Phil Holland]

Phil isn't technically wrong.. he is being super clinical ( as he should be).

Yep, this is more or less a by the numbers test with no real creative interpretation. For me under most the most common shooting conditions the Leica Summilux-Cs are very usable lenses at Dragong's 6K Full that cover at the T Stops you'd likely be shooting at anyways.

And these are actually currently my favorite super speeds out there for a variety of reasons. Mainly the character, resolving power, and somewhat miraculous coatings. Size too.

Hey Phil, the larger the sensor. The better the resolving power for all lenses across the board?
Does it mean that for the anamorphic, it's possible wider apertures?

Hi Charles. A few things to take note on in regards to true resolving power.

Quick thing to keep in mine, for the immediate future when it comes to motion pictures we'll be creating a 4K master for a 4K delivery. That's going on industry wide at this point.

Okay. Now onto the Dragon sensor.

First up is the pixel size. Pixels are smaller on Dragon compared to Mysterium-X, so in terms of sensor "real estate" you get more pixels in the same physical area.

The next thing to consider is that capturing at 6K, 5.5K, and 5K finished for a 4K deliverable is going to take advantage of a concept most call "super sampling". Meaning capturing at a higher resolution then your desired final output and scaling down to your desired output size. Various re-size and scaling methods yield different results, however all benefit by having more information. And we're not even getting into sharpening talk yet, which can extract even more detail.

This is going to benefit all lenses really. Most modern optics from the last few decades resolve 75-200 line pairs per millimeter. Commonly 140-200LP/mm center frame, lower then 140LP/mm is considered fairly soft these days. Some lenses hold that 200LP/mm across the field, some soften up a bit. Most soften drastically towards the outer field of S35 coverage. This is pretty much how they have been designed and also some physical limitations behind the elements used.

The advantage here is if you capture at 6K with say a lens that's in the 140LP/mm range and scale down to 4K for your finish you'll have a fairly impressive result right away.

Great resolving lenses tend to be in the 200LP/mm and up. Depends on how your measure that really. When you see an "easy 200" you're looking at some extremely nice glass.

 

[Elsie N]

I'm so pleased I went with FX Nikon glass originally.

For the time being, my only concern is paying for Dragon updates knowing my lenses used on my Epic MX will transfer smoothly to the Dragon sensored camera.

I made the statement in the past that the RED camera is so good it makes pretty pictures even with lenses that have previously been discounted as not good enough. Of course Leicas and the other great cine lenses will look great for those who work in the "best image available" world of imaging, but from an Indie perspective, this camera in all its iterations make my stills Nikons look great.

I consider the Dragon will not only upgrade my camera, but my lenses as well.

 

[Mike 'Fireman' Ross]

Jeff, I don't think the vII has any coverage differences vs the vI.

Are you sure about the Tokina not covering 16mm at Full Frame? Maybe it vignettes and we haven't noticed? If you look at this piece here, the opening "wide" shot is the 11-16 on a 5D, whereas the rest of the piece is Red One MX with Ultraprimes.

Either way. Point still stands. This "little powerhouse" as you call it (I like that!), should be TERRIFIC on Dragon.

If you don't need to go quite so wide I'll keep banging my drum for the Nikon 14-24mm f/2.8. Astonishing quality for a mass-produced lens at that price.

And of course absolutely NO coverage issues on Dragon at 6K at any focal length...

Mike

 

[Jeff Kilgroe]

Jeff, I don't think the vII has any coverage differences vs the vI.

Are you sure about the Tokina not covering 16mm at Full Frame? Maybe it vignettes and we haven't noticed? If you look at this piece here, the opening "wide" shot is the 11-16 on a 5D, whereas the rest of the piece is Red One MX with Ultraprimes.

My Nikon mount versions -- I own two of them -- of the 11-16 both fall just shy of total coverage at 16mm on my D3X and D800. So I guess that's where I'm coming from with it. I'm also nitpicking -- it's mostly a portholing effect with noticeable falloff. And ever so slight start to vignette in the very tips of the corners. I also own the Duclos PL modded version of the lens, never have had it on a camera with a full-frame sensor. It's one of my all-time favorite lenses to shoot with!

Either way. Point still stands. This "little powerhouse" as you call it (I like that!), should be TERRIFIC on Dragon.

Totally agree!

 

[Nick Morrison]

If you don't need to go quite so wide I'll keep banging my drum for the Nikon 14-24mm f/2.8. Astonishing quality for a mass-produced lens at that price.

And of course absolutely NO coverage issues on Dragon at 6K at any focal length...

Mike

No doubt. Many people consider the Nikon 12-24 2.8 one of the great Full Frame wide-angles available. It's an absolute BEAST. The Tokina 11-16 is not as good, it's just a whole lot cheaper.

Pound for pound, the Zeiss 15 2.8 and Nikon 12-24 2.8 are hard to beat if your looking for FF wides. The Zeiss 21 2.8 is a beast too, but not an ultrawide like the other two, so I personally categorize it in a separate "bucket".

 

[Will Keir]

How many stops do we gain in the darks with Epic Dragon from Epic MX?

I was excited to hear Mark comment that the Dragon shoots the image he sees with his eyes. That's why I signed on for this ride in 2005 and put money down in 2007. Both the R1 and Epic were not quite sensitive enough for me, which is why I turned to high speed lenses. Now I'm super badass at pulling the shallow depth of field work, unintended but useful.

Is a Epic MX + Master Prime at T1.3 = Epic Dragon + Ultra Prime at T1.8 in terms of light sensitivity?

And remember we tested wide open and occasionally stopped down to T2.8 when those lenses were on the projector. Wide open the Leica's all projected the target "cleanly" with a bit of fall off.