Has anyone done a proper F3 resolution test yet?

[Duke M.]

Some production models are out. I did a Google search and didn't find one.

Is anyone doing a proper resolution test?

[TimurCivan]

i got a res chart with my F3. tom and i are going to perform one on sunday.

Also, i'm going to be using Cooke lenses, so sharpness from the lens should be a moot point. The Cookes resolve 4x more than a 5k chip can resolve. It will really show off what the Chip itself is doing.

[Duke M.]

i got a res chart with my F3. tom and i are going to perform one on sunday.

Also, i'm going to be using Cooke lenses, so sharpness from the lens should be a moot point. The Cookes resolve 4x more than a 5k chip can resolve. It will really show off what the Chip itself is doing.

Woot! Be still my beating heart. I'm looking forward to it. At last some facts to go along with the assumptions! :happy:

[PhantomVideo]

The Cookes resolve 4x more than a 5k chip can resolve.

Where's that info Timur? (or is it a knowen fact) be good to read about

[TimurCivan]

Les, the owner of cooke told me.

I tested the new leicas today. They resolve. 7x more than 5K can hold.

they are also $30,000 a piece. They are so sharp, the tech from band pro is going to have to invent a new resolution chart.

[Duke M.]

Those sound absolutely fantastic, but seem to be over kill.

If Nikon's resolve about 3x more than 1080p the results should be similar other than any 'look' imparted to the image.

[G.P.]

I will be conducting a few tests at BandPro with our Sony F3, today I was there with the F3 and F35 and we were doing some simple visual examinations with a 24'' HD broadcast camera. The resolution is quite good, will be updating everyone next week with the full results and inviting people to come view the tests in house.

[nugat]

Les, the owner of cooke told me.

I tested the new leicas today. They resolve. 7x more than 5K can hold.

they are also $30,000 a piece. They are so sharp, the tech from band pro is going to have to invent a new resolution chart.

7x more than 5k is ca 700 lp/mm, that must be in ultraviolet.
I hope they test them on the proper lens projector/mtf bench.
Even if it was half of the figure, it's meaningless without all resolution parametrs:contrast, saggital/meridional, focal length/distance, aperture, distance from the picture center....
In other words the proper mtf--modulation transfer function.
Are the leicas really 30k at that vendor?
Any 1920x1080 camera can resolve nyquist x kell factor or for a solid state imager ca 970 TVL. That's the maximum theoretical figure for diffraction limited glass and the rest of the system mtf =1, which does not exist in real life. However seeing spurious detail happens (eg. by mothers-in-law).

[LeavingTheCandy]

Wow...my brain hurts after reading that one.


The field of nonlinear optics studies the changes to material polarization density under extreme electric fields. The polarization waves generated are related to the generating electric fields through the nonlinear susceptibility tensor. If the polarization P is not linearly proportional to the electric field E, the medium is termed nonlinear. To a good approximation (for sufficiently weak fields, assuming no permanent dipole moments are present), P is given by a Taylor series in E whose coefficients are the nonlinear susceptibilities:
http://upload.wikimedia.org/math/b/4/c/b4c8c7265a671dfe00905cb78cdc9f04.png

[TimurCivan]

I will be conducting a few tests at BandPro with our Sony F3, today I was there with the F3 and F35 and we were doing some simple visual examinations with a 24'' HD broadcast camera. The resolution is quite good, will be updating everyone next week with the full results and inviting people to come view the tests in house.

Hey you were at band pro yesterday?

What time? i was there between 1 and 3:30...

[TimurCivan]

I will be conducting a few tests at BandPro with our Sony F3, today I was there with the F3 and F35 and we were doing some simple visual examinations with a 24'' HD broadcast camera. The resolution is quite good, will be updating everyone next week with the full results and inviting people to come view the tests in house.

Hey you were at band pro yesterday?

What time? i was there between 1 and 3:30...

[G.P.]

I was there around that time too... I must have talked to you and not even known it haha... I had on my NYFS baseball hat. I brought my 35mm T2.2 with me and put it on the F3, that was me.

[TimurCivan]

you have beard? blue jacket?

i was the guy in all black. w/ horn rim glasses. drooling over the F35

[G.P.]

Yep that was me, we talked haha, yeah gotta love that camera. A little much for some work, but god its a beautiful thing, got to use it a few times. Have you played with the 9000PL yet? its a dream to work with and is just like an F35 image wise.

We should get together and combine F3 forces if needed for multicam work.

[TimurCivan]

Shoot i wish we would have connected better. WEll no matter, come check out my glass. LAter this week DP Clayton Haskell and i are going to do a test out doors. you should come. Shoot me an email.

[TimurCivan]

7x more than 5k is ca 700 lp/mm, that must be in ultraviolet.
I hope they test them on the proper lens projector/mtf bench.
Even if it was half of the figure, it's meaningless without all resolution parametrs:contrast, saggital/meridional, focal length/distance, aperture, distance from the picture center....
In other words the proper mtf--modulation transfer function.
Are the leicas really 30k at that vendor?
Any 1920x1080 camera can resolve nyquist x kell factor or for a solid state imager ca 970 TVL. That's the maximum theoretical figure for diffraction limited glass and the rest of the system mtf =1, which does not exist in real life. However seeing spurious detail happens (eg. by mothers-in-law).

Thats what they said at the demo...

[nugat]

Thats what they said at the demo...

30k is more than I thought they would cost. More than Zeiss Masters anyway...
As for resolution...
The Leica-C designer Iain Neil admitted the goal was to get 8k resolution and still excellent contrast. Without any MTF's released it's difficult to guess what he meant, but...
Zeiss usually produces best top glass without the hype Leica surrounds it's own products with. The optical progress over the Zeiss Masters can be only incremental, the Leica price premium of 30-50% over Zeiss is ususally connected with the "magic" that surrounds the name, not really any significant differences in performnce.
The highest ever resolution recorded (in the lab) on a production lens was Zeiss Biogon at ca 400lp/mm, close to the diffraction limit (and that must have been at 10% contrast or so, barely discernible, useless in every day use).
My bet is the summilux-C is around the level of Masters, ie. delivers 40-50lp/mm across the frame with ca 70% contrast.

[TimurCivan]

What the guys at Band pro showed us was that when projected the leicas easily displayed 200 LP all the way across the frame. He said the lenses can resolve more than the charts show and a new chart may be neccesary.

In addition, Les at cooke said his lenses, "and his competetors" can all resolve upwards of 200. Which is apparently more than 5k can resolve. The bandpro demo seemed to infer the leicas were far beyond even that. I heard a number thrown around. But perhaps without documented data let's wait till real facts are released.

[nugat]

What the guys at Band pro showed us was that when projected the leicas easily displayed 200 LP all the way across the frame. He said the lenses can resolve more than the charts show and a new chart may be neccesary.

In addition, Les at cooke said his lenses, "and his competetors" can all resolve upwards of 200. Which is apparently more than 5k can resolve. The bandpro demo seemed to infer the leicas were far beyond even that. I heard a number thrown around. But perhaps without documented data let's wait till real facts are released.

+200 lp/mm from projection for top cinematic lenses like Cookes, Zeiss, Leica is quite possible at low contrast . Put on the MTF tester that would be 10%, barely discernible shades of grey. Therefore 200lp/mm is purely a technical feat of no use in picture making. We deem a picture to have a good contrast when it's 50%--for such, the resolution falls to say 60-80 lp/mm for the same top glass. Excellent contrast of eg. 70% is achieved at resolutions of 40-50lp/mm (Zeiss Master Primes and such). The main progress in lenses over the last half century was in contrast and decreasing aberrations.
From 40% to 70% for the same lp/mm resolution is the whole world of difference.
5k is a sensor 5000 pixels wide. It has the physical width of 25mm (24.9 mm is S35 Full Aperture).
5000/25=200 lines/mm or 100lp/mm. To say that a lens "outresolves" 5k or 100lp/mm is meaningless without the contrast figure*. Eg. Zeiss MPs get to +100lp/mm wide open (f1.2) with ca 40% contrast.
*and aperture, focus distance, f-length, distance from the optical axis

BTW, Band Pro had offered Leicas-C for prices similar to MP's:

"Hi Guys,

It's been a while since I've posted here, been busy.

The 8 lens set (18. 21, 25, 35, 40, 50, 75, 100) is available from Band Pro for $178K USD. We are taking 5% deposits and sets ordered today are currently expected to deliver late-feb early-march.

16 & 65 lenses will be following some time in 2011.

Best,

Randy Wedick
Technical Consultant
Band Pro Film & Digital"

[Duke M.]

The 8 lens set (18. 21, 25, 35, 40, 50, 75, 100) is available from Band Pro for $178K USD. 16 & 65 lenses will be following some time in 2011.

A bit off topic, but is that a common range for a prime lens set? Some of those seem too close together. From 18 to 21 mm is only 3mm difference.

For my Nikons I have:

17, 24, 35, 50, 85, 105.

[TimurCivan]

Its normal for higher end glass.

[Duke M.]

That's going to leave a lot of people hurting with the 3 lens PL set Sony's offering then.

[nugat]

That's going to leave a lot of people hurting with the 3 lens PL set Sony's offering then.
what the sony set is missing is a cinematic wide 18-20mm. other than that, it's a sufficient set to shoot for oscars.

[Everts]

Its a coming, 11-16 and 17-50 f2.8 or T2.8

[Duke M.]

what the sony set is missing is a cinematic wide 18-20mm. other than that, it's a sufficient set to shoot for oscars.

That's kind of what I was wondering. My Nikon 6 prime lens set is 17, 24, 35, 50, 85, 105 and pretty much covers everything without breaking out the zooms. (I know some people want ultra wide.)

The three prime lens set seems too limited and that 8 lens set seems overkill with the close to overlapping lenses.

[TimurCivan]

You can easily shoot a whole film on the three lenses in thse set. It will require some planning but its doable.

[Cosimo Bullo]

You can shoot with 1:

http://en.wikipedia.org/wiki/Yasujirō_Ozu

[Duke M.]

Sure, you could shoot a movie on an I phone or and HV20, but it would be a PITA and take longer (therefore more expensive on a pro shoot) so that's not the real question.

To paraphrase the question: Balancing time and cost what lenses would most pros pick as the set reasonably necessary to shot the average movie quickly and easily.

[TimurCivan]

The standard 5 lens set is there for a reason. 18 25 35 50 85. Maybe a 100.

[Lee Saxon]

Which is apparently more than 5k can resolve

I really dislike when companies make these resolution-based claims, whether it's "this lens outresolves 5K" or "this lens covers full 5K"

That tells me nothing. You can put 14 megapixels onto a 1/1.2 compact camera sensor or a 6x4.5cm digital back.

Just a little pet peeve of mine, I guess.

[Duke M.]

Back to the original post, can anyone just set up a chart and give us a proper still?

[nugat]

Back to the original post, can anyone just set up a chart and give us a proper still?

Alan Roberts (BBC/EBU tests) just wrote at dvdoctor.net (NEX thread) that he'd be posting F3 results next week.

[hunter richards]

Found this (scroll down) looks as sharp as it gets for 1080p: http://www.kamrat.tv/

[Duke M.]

Alister Chapman posted some Imatest results. One thing I like about Imatest is that, even tough they can't be compared to non-Imatest results, because it is a computer doing the evaluations comparisons between different cameras with Imatest are consistent.

31052 31053

MTF50 is the most common test. These results indicate 798 LW/PH V and 958 LW/PH H, which is right up there. Very close to EX1/3 and about at the Nyquist limits.

[Razz16mm]

MTF50 is the most common test. These results indicate 798 LW/PH V and 958 LW/PH H, which is right up there. Very close to EX1/3 and about at the Nyquist limits.

That is about as good as 1080p gets under real world conditions.

[Duke M.]

+1

[Fohdeesha]

Can we get some bad news about this camera to balance all this out? :2vrolijk_08:

[BobbyMurcerFan]

Alister Chapman posted some Imatest results. One thing I like about Imatest is that, even tough they can't be compared to non-Imatest results, because it is a computer doing the evaluations comparisons between different cameras with Imatest are consistent.

31052 31053

MTF50 is the most common test. These results indicate 798 LW/PH V and 958 LW/PH H, which is right up there. Very close to EX1/3 and about at the Nyquist limits.

Are these done off of XDCAM-EX or out the HD-SDI? Thanks.

[Duke M.]

XDCAM files. I'd think that HD-SDI would be (very) slightly better and pretty much the same as EX1 resolution.

[Lee Saxon]

Can we get some bad news about this camera to balance all this out? :2vrolijk_08:

I can't afford one :(

[Cinesin]

The viewfinder sucks, that's all I can come up with so far.

Wait till I hook up a KiPro to one of these this week.... although you may not want to see that stuff! ;)