Misterium and wide angle lenses

[Andrew M.]

While trying to pick good wide lenses for RED camera, I came a cross the film versus CMOS compatibility problem.
We all know about compatibility problem of lenses manufactured for 3 x CCD systems where focus of R G B color is not landing on the same plane and what to do about it, see related thread where this subject was extensively discussed.

The CMOS sensor versus film in the view of 10mm or 8mm focal length lenses is bit more complicated.

Some 10mm and 8mm lenses are constructed with the short (about 30 to 40mm) back focal distance.
http://en.wikipedia.org/wiki/Wide-angle_lens

Misterium may or may not have microlenses used on top of the 5X5 micron light sensor (pixel). Also the construction of the photoelectric sensor is sometimes as deep as size of the pixel itself.
http://www.micron.com/innovations/imaging/pixel

Taking the short back focal length under account we may end up with poor performance of wide angle lenses, unless specially designed for the particular CMOS sensor.
Using retrofocus wide angle lenses may improve the performance but 70 – 90 degree light beam angle could be a must depending on the Misterium pixel geometry.
http://en.wikipedia.org/wiki/Retrofocus

I know that Misterium details will be released in just a few weeks but judging on other manufactures of the cameras we do not have much to work on.
On top of other common CMOS characteristics will we get following?:

Fill in factor (pixel area versus photosensing area)
Vertical (cross section) geometric characteristic of the pixel sensor:
Microlenses characteristics if used.
http://www.lumerical.com/fdtd_microlens/cmos_image_sensor_pixel_microlens

I will be much more confident buying RED 10mm wide angle lenses though.

[Mike the beginner]

The good things about all this is that Jarrad has indicated sometime ago that he would do lens tests with as many lenses as possible.

I hope ypur right about red possibly coming out with a wide 10mm prime lens.

I think it is an odds on cert they will announce their intention at NAB to provide a set of primes possibly sometime next year, just guessing, but would they make the widest one 10mm or would they play safe with volumes and stick to 12mm?

Any guesses as to the likely "red price" for a wide 10mm? Would they sell enough of them?

Mike the beginner

[Andrew M.]

Since the 4K camera lends itself for cine format 1: 3.5 or 1: 2.4 so well and high res screen shots look so impressive, the 100 deg FOV is a must.
It is where the power of the 4K image is coming out. You feel like you are almost there. I know, Jim got himself 300mm for the racing shots first, hey he is the founder of RED so his dreams come true first. But if Jim will get few aerial panoramic shots with top notch 8 or 10mm of Alps or underwater caves, he will fall in love again. After I did some visual tests, IMHO 2.4 ratio screen at 4K on large screen will be just small step away from 70mm IMAX.

[Brook Willard]

Even if wide-angle lens tests are not performed, I'm hoping that there will be some good inter-company friendliness at NAB. Maybe somebody from a lens company booth can bring one of their specialty lenses over to the RED booth for 10 minutes for some shits and giggles.

[Dan Blanchett]

A RED 10mm would be amazing. But I would be pretty happy with the announcement of a 12mm, or any wide angle RED prime for that matter.

[Andrew M.]

Even if wide-angle lens tests are not performed, I'm hoping that there will be some good inter-company friendliness at NAB. Maybe somebody from a lens company booth can bring one of their specialty lenses over to the RED booth for 10 minutes for some shits and giggles.

If not, as soon as we can get our hands on single digit serial number RED we can test quite a few lenses, 5 hours flight anywhere from here, will not hold me from trying a good lenses. Alternatively any single digit reservation holder wants a limo VIP treatment here for a weekend?

I expect some serious ripple vignetting or interference patterns out of 8-10 mm wide angle and les than 40 mm back end focus lenses. Could be some sharp edges fringing as well. All could be not significant I hope.

Anyway, lenses designed for flat film surface are what they are.
BTW the film has its own problems with in film light dispersion at sharp angles and film curvature focus or rather lack of it.
Add to it loss of resolution when you copy the material from master and you end up with 2MP maximum. Digital coping is 100% perfect, isn’t it?

What is killing the film is the copy process, masters are quite good these days and this is the most common mistake when people compare film to digital. They forget about the fact that we are all watching the copy, aaa... except DPs:-)

[Stephen Williams]

What is killing the film is the copy process, masters are quite good these days and this is the most common mistake when people compare film to digital. They forget about the fact that we are all watching the copy, aaa... except DPs:-)

Hi,

There is no reason when doing a DI not to make multiple original negatives, so no further copying is required.

If you want higher quality than normal 35mm then Anamorphic photography or 65mm is probably the way to go.

Stephen

[Andrew M.]

Hi,

There is no reason when doing a DI not to make multiple original negatives, so no further copying is required.

If you want higher quality than normal 35mm then Anamorphic photography or 65mm is probably the way to go.

Stephen

Stephen, I agree with you when 65mm is concern, but 65mm or DI, the light is passing 3 times through lenses before you end up with the copy.
In digital process though, especially when you display on the non projecting screens, you go through lenses only once, when shooting.

Andrew

[Brook Willard]

Extreme wide-angle lenses with aberration correction based upon the three [or four] color layers in film will be of greatest concern to me. Because very high-end lenses will split R, G and B beams a calibrated amount around the edges of the frame to compensate for the deeper red and green layers of film, they will likely have the opposite effect on a Bayer pattern sensor. Because all pixels on a Bayer sensor are at the same depth, the "correction" for film will probably read as aberration on RED.

Just a guess.

[Chris Forbes]

Maybe somebody from a lens company booth can bring one of their specialty lenses over to the RED booth for 10 minutes for some shits and giggles.

WARNING: Spoilers ahead.

http://www.thevideodepartment.net/4500.jpg

Sorry guys. Really I am.

[Andrew M.]

Extreme wide-angle lenses with aberration correction based upon the three [or four] color layers in film will be of greatest concern to me. Because very high-end lenses will split R, G and B beams a calibrated amount around the edges of the frame to compensate for the deeper red and green layers of film, they will likely have the opposite effect on a Bayer pattern sensor. Because all pixels on a Bayer sensor are at the same depth, the "correction" for film will probably read as aberration on RED.

Just a guess.

Let see...quick check, correct me if I am wrong.
http://www.fujifilmusa.com/JSP/fuji/epartners/bin/AF3-865E.pdf

looks like layers are 4 micron each
Curvature of film 20 microns
Looks like most problem with film when focal plain distance is concern, is film curvature.
Resolution of the good quality film, 100 lines pairs per millimeter (5 microns per each line, black line and white line)
Max resolution of 8R 200 lines pairs per millimeter.

Misterium pixel size 5.5 microns so 181.81 pixels/millimiter

[Graeme Nattress]

5.4 microns, actually...

Graeme

[Andrew M.]

Thanks! Graeme.

We have more news from Arri.
The distance from the closest glass to the film plain on 8R is 29.57mm
However we want to know the pivotal point distance of the beam to figure out the beam angle of the light falling on the Misterium in the far corner.

30mm is awfully close if Misterium has microlenses on top of CMOS cells.

[Brook Willard]

The only thing over the sensor is an OLPF, as far as I know.

[Graeme Nattress]

That's correct Brook.

Graeme

[Andrew M.]

Any influence on the beam going through the OLPF thin glass plate at say 45 deg angle?
I think it is right above the CMOS surface, correct?
Putting it straight as a layer on the CMOS is not recommended?