Interesting information...

[Hugues Wisniewski]

Those who'll dare to ask for a 617 uncompressed option will need an aircraft carrier of raid drives to store that ;-) and another one loaded with cash for the deposit
Jim talked about 4.8GB/s for an epic uncompressed. I'm still wondering who would be interested in transforming its Red into a Viper with giant raid arrays in the background.

Shooting 3D with one camera could be handled if all the shots were done on a dolly at constant speed, with the camera's angle positioned at 90 degrees from the tracks.
Then duplicating the video track and moving one track by a few frames would do it. That's how they do 3D aerial photography.
Uneasy for dramatic close ups ;-) I'll give you that, but using a 617 would look amazing on those giant wild west landscapes and canyons

 

[Craig Ryan]

So 617 goes beyond the scope of both human eyes...incredible. And up to 30 fps too! I guess with an aspect ratio like that, it would completely wrap around your peripheral vision entirely...its the new Cinerama!

 

[Eric Lange]

Those who'll dare to ask for a 617 uncompressed option will need an aircraft carrier of raid drives to store that ;-) and another one loaded with cash for the deposit
Jim talked about 4.8GB/s for an epic uncompressed. I'm still wondering who would be interested in transforming its Red into a Viper with giant raid arrays in the background.

Shooting 3D with one camera could be handled if all the shots were done on a dolly at constant speed, with the camera's angle positioned at 90 degrees from the tracks.
Then duplicating the video track and moving one track by a few frames would do it. That's how they do 3D aerial photography.
Uneasy for dramatic close ups ;-) I'll give you that, but using a 617 would look amazing on those giant wild west landscapes and canyons

Funny you should say that, as I was going to post that I would love to take a 617 to Ghost Ranch in New Mexico (ware I am based) to carry out 3d recording. The location has been used for a lot of movies and would be good to have in full 3d and modeled. As a photogrammetrist that builds 3d cameras and software I can say the 617 would be truly awesome if it only shot stills (raw or otherwise). The camera shifting and aerial case is good for static subjects I have used this technique on 3d recording of historic architecture in the U.K. in the past, however dynamic subjects such as human subjects require (for the most part) two lenses and simultaneous exposure. The 617 could certainly take two lenses for narrow base stereo work, but two 617’s with wide angle (low distortion) Schneider or Rodenstock lenses at very wide stereo base separations for capturing fine 3d detail from large subjects would be awesome and also very needed. We used to use surveying cameras with large glass plates 130 x 180 mm on special super precise Zeiss UMK’s with 100mm Biogon lenses. These plates used to be scanned to 15 micron (huge files), as yet there is no single shot digital system than can come close in resolution to the large glass plates we used to use; that is until the 617. I feel that Jim has a special interest in the 617 as a Jim special project close to his heart; I’m sure he would say all his "children" are special but the 617 can only come from some one who is truly passionate about the image and photography.

 

[Tony de Vries]

As amazing as the eye is, that I believe that sensitivity number includes the rather slow chemical-adaptaion the eye undergoes when the overall luminance of the scene cahnges drastically (i.e. the "eyes adjusting to the dark" phenomenon). That adaptaion slides the "window" of detectable contrast range up and down.

This would be akin to being able to add and remove ND filters in front of the lens.

The actual contrast ratio your eye can resolve at any one time within a scene is far less that... maybe 16-18 stops...

-Steve

It's even better, our eyes can resolve 24 stops in a single scene, on top of which we can adjust our eyes instantly to see even more detail in the bright or dark portions in that scene.
Now all RED has to do is create some kind of interface and record straight from the retina. I propose the name REDtina. :shiftyph34r:

 

[Eric Lange]

It's even better, our eyes can resolve 24 stops in a single scene, on top of which we can adjust our eyes instantly to see even more detail in the bright or dark portions in that scene.
Now all RED has to do is create some kind of interface and record straight from the retina. I propose the name REDtina. :shiftyph34r:

I was “taught” that the eye can only distinguish (as seperate) about max 24 grey levels simultaneously under a static lighting condition (there are special response curves also for different levels of illumination, (roughly logarithmic). It is the IRIS in the eye that allows us to distinguish faint star light to blazing sun shine. Is my thinking out of date on this? Do you have a reference for 24 stop simultaneous perception?

(please don’t feel I’m being snotty, I am genuinely interested in what you have said and suspect my knowledge may be very out of date)

 

[Carsten Kurz]

Jim, after incredible sunglasses, after incredible cameras,........ will you build a stargate?

Hehe - what do you think where Jim pulls out all that spacy stuff RED sells?
Develop it? Here? ON EARTH? Spend some more time on Terminator II and try to understand the concept!



- Carsten

 

[Jason Diamond]

...We can see all sorts of combinations with existing equipment and tremendous opportunities for "new thinking". We are just happy to be here...

Jim

is that a subtle AVID reference ....

 

[Steven Caesare]

It's even better, our eyes can resolve 24 stops in a single scene, on top of which we can adjust our eyes instantly to see even more detail in the bright or dark portions in that scene.
Now all RED has to do is create some kind of interface and record straight from the retina. I propose the name REDtina. :shiftyph34r:

That's a contrast ratio of over 16,000,000:1.

Even with dark adaptation most numbers I've seen are closer to 10,000,000:1, which is closer to 23 bits. But the dark adaptation time is a matter of minutes, so this isn't a "single scene" scenario.

Do you have any reference for your claim?

-Steve

 

[Florin Andrei]

That's insane! But I'd rather have you guys beat the dynamic range of the human eye, which is almost 24 f stops. :shifty:

The "sensor" per se in the eye is even better than that: it can resolve single photons, that is a single quantum of light is enough to trigger the receptors!
But the neural net behind it has a noise filter with a threshold of about 5 to 9 photons within 100 ms - without that noise filter, our perception in darkness would resemble the image from a cheap ultracompact digital camera.

http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.html

So the 24 f stops is the overall range after the noise limiter is applied. The range is purposefully limited towards the dark end.

Fascinating stuff.

 

[George Wilcox]

Will there be a projector to go with this?

Will there be a projector to go with this?

I'm dyin' to know Jim - are you developing projectors to show all those fabulous pixels to their very best advqantage? :usd:

 

[Florin Andrei]

My brain doesn't have to debayer Well, I guess it might, but it does it in real time.

The human eye has really crappy, awful optics. The image on the retina is all mangled and distorted, with all sorts of aberrations.
The brain actively compensates for that by "wagging" the eyes in all directions constantly, in brief but rapid and constant little motions, so the image projected on the retina moves constantly (even when you try to keep your eyes focused in a single point - they still move involuntarily). Due to that process, the brain then is able to apply some sort of dynamic filter which cleans up the distortion and renders the image clean.

http://www.journalofvision.org/4/4/4/

Mama Nature is clever.

 

[George Wilcox]

617 RAW can be done...

617 RAW can be done...

The latest OCZ Vertex 250 GB 2.5" SSD drives could handle shooting 617 RAW with a RAID 0 of 30+ drives with 160MB/Sec write speed per unit with a pair of 16 to 24 port RAID controllers. More drives means better reliability of transfer speed. 48 drives would provide of 12 TB that would allow you 41 minutes of shooting time! It would cost around $42K for the hardware.

http://www.ocztechnology.com/products/flash_drives/ocz_vertex_series_sata_ii_2_5-ssd

 

[Tony de Vries]

I was “taught” that the eye can only distinguish (as seperate) about max 24 grey levels simultaneously under a static lighting condition (there are special response curves also for different levels of illumination, (roughly logarithmic). It is the IRIS in the eye that allows us to distinguish faint star light to blazing sun shine. Is my thinking out of date on this? Do you have a reference for 24 stop simultaneous perception?

(please don’t feel I’m being snotty, I am genuinely interested in what you have said and suspect my knowledge may be very out of date)

You're right. :help:

You can start here:
http://www.luminous-landscape.com/columns/eye-camera.shtml

More here:
http://www.clarkvision.com/imagedetail/eye-resolution.html

It's difficult to compare the eye to the camera. Take resolution for instance, we can find out how many cones and rods the eyes have, but you can also ask yourself, how many megapixels do I need to resolve this scene with as much detail as the eye can see. That means something completely different, since most of the resolution in the eye is concentrated tot the center, so we can scan a scene at a much larger resolution than the amount of cones and rods dictate. Same goes for dynamic range, we can quickly adjust the iris when we scan over a scene, remember the details and let the human brain do the rest.

 

[Pawel Achtel]

Jim - could you do 3d with one sensor? seems neat and tidy to me.

No, because you can't adjust for convergence.

 

[Pawel Achtel]

There are more pixels in the RED 617 Mysterium Monstro sensor (261,352,000) than there are rods and cones in both your eyes (126,000,000 x 2 eyes = 252,000,000).

Just in case you were wondering...

Jim

All we need is a lens that would resolve it :whistling:

 

[Mark Pugh]

No, because you can't adjust for convergence.

Hi Pawel
I realised that shortly after my post. Was waiting for someone to mention it...
took a while...
So I've been wondering if it can be corrected in post by sliding the two images slightly left or right. Why wouldn't that work?
I'm certainly no 3D expert... seems to me though, that could be an easier solution than worrying about convergence on set.

 

[Mark Pugh]

Sorry to see you're not having much of a Saturday night either, Pawel.
I've been sitting here with my two-year old sitting on my lap, me looking at Reduser on the left of the screen, and we're singing along with something I shot for Disney Channel (on Red) that's playing on the right side of the screen!

 

[Sanjin Jukic]

All we need is a lens that would resolve it :whistling:

Master Primes obviously not .

Maybe Schneider-Kreuznach Digitar lenses.

"The focal lengths range from 28 to 150 mm with image circle diameters of 60 mm and above,
and are therefore most suitable for high resoltion digital photography with current image sensors.
As a result of their MTF values and the apochromatic correction, they are also largely suitable for
image sensors of the next generation."

More about Digitar lenses.

"What is line resolution and how does it compare

Line Resolving Power

Resolution is measured by how many lines a lens can project side by side into 1 millimeter of
space with the lines still clearly distinguishable from each other. One black and one white line
make one line pair.

Capture Area and Resolving Power

The lens used with the small digital chip has to be able to project x times the lines onto the
same space then the large format lens! The higher the resolving power of the chip the more
crucial is the performance of the lens.

Chip Resolution

Sensors on chips are aligned in straight lines. The new chips with 6.8 micron sensor size have
approximately 148 ccd sensors per millimeter which equals apprx. 70 line pairs per millimeter.

Resolving Power in Comparison

Medium Format apprx. 40 to 60 line pairs per millimeter
Large Format apprx. 30 to 50 line pairs per millimeter
Digitar Lenses apprx. 90 to 200 line pairs per millimeter

Existing large format lenses are not a good choice for digital work in terms
of their resolving power because they have not been designed with the
small area of the chip in mind."

Also must read:

Optics for Digital Photography

 

[Pawel Achtel]

Hi Pawel
I realised that shortly after my post. Was waiting for someone to mention it...
took a while...
So I've been wondering if it can be corrected in post by sliding the two images slightly left or right. Why wouldn't that work?
I'm certainly no 3D expert... seems to me though, that could be an easier solution than worrying about convergence on set.

It is not just a matter of shifting the images, but the focus plane of left image must be at an angle to focus plane of the left. Maybe if you used shift lenses.

 

[Mark Pugh]

Unfortunately not because image planes must be at an angle to each other. It is not just a matter of shifting the images, but the focus plane of left image must be at an angle to focus plane of the left. Maybe if you used shift lenses.

yep - I'd thought about shift- tilts... not very sensible, I think! hehe.
Would having image planes not on an angle matter for anything except really close-focus shots, really?
anyway... I don't think I'll be shooting 3D on this camera in a hurry.