I Need A Headslap!!!

[number6]

This could be posted in the workflow thread but since Scarlet is the first step in achieving it, I will post it here.

I'm needing guidance in figuring out how 3-D (from two Scarlets) will affect a workflow? Does it double? Do you import the data from each (synced) camera or will the dual Scarlet configuration have a common storage that will record the data as one combined set? What software is available to seamlessly work with 3-D?

I confess to being a RED One 4K snob... but I am starting to believe that this Scarlet... that's some idea for a camera!

 

[Jeff Kilgroe]

Shooting 3D with Scarlet (or RED) first assumes that you have the proper hardware rig for mounting the cameras and all that. Once you have your footage, the most direct way of handling 3D stereo footage is with SCRATCH or possibly SpeedGrade as they both have stereo 3D modules. Some of the much more expensive high-end systems do as well.

If you bring in your R3Ds in SCRATCH as a stereo pair, you can trim, color, and finish your footage to your needs. You can conform to edits done in Final Cut done with just the right eye... Managing a stereo pair and editing it isn't the tough part. It's the extra bits you'll have to deal with for additional FX and compositing that will add a whole new dimension (hehe) to your work requirements. Output and delivery is another consideration... For example, will you be going to a system with LCD shutter glasses? Polarized projection? Out to broadcast or DVD for home viewing with anaglyphic (red/blue) glasses?

The workflow options are there. But they're not all that cheap... You can go cheap and do it the brute-force, manual, "I've got more time than money" way too...

 

[number6]

Shooting 3D with Scarlet (or RED) first assumes that you have the proper hardware rig for mounting the cameras and all that. Once you have your footage, the most direct way of handling 3D stereo footage is with SCRATCH or possibly SpeedGrade as they both have stereo 3D modules. Some of the much more expensive high-end systems do as well.

If you bring in your R3Ds in SCRATCH as a stereo pair, you can trim, color, and finish your footage to your needs. You can conform to edits done in Final Cut done with just the right eye... Managing a stereo pair and editing it isn't the tough part. It's the extra bits you'll have to deal with for additional FX and compositing that will add a whole new dimension (hehe) to your work requirements. Output and delivery is another consideration... For example, will you be going to a system with LCD shutter glasses? Polarized projection? Out to broadcast or DVD for home viewing with anaglyphic (red/blue) glasses?

The workflow options are there. But they're not all that cheap... You can go cheap and do it the brute-force, manual, "I've got more time than money" way too...

SLAP!!! Thanks Jeff... I needed that:biggrin: Hope 3-D will receive more attention on REDuser though... I think Scarlet is going to make its mark in this way of shooting.

(I think Scarlets should be sold as a brace, as well as individually)

 

[Dan Hudgins]

Crossed eyes stereo 3D as 2D frames to edit/CC

Crossed eyes stereo 3D as 2D frames to edit/CC

This could be posted in the workflow thread but since Scarlet is the first step in achieving it, I will post it here.

I'm needing guidance in figuring out how 3-D (from two Scarlets) will affect a workflow? Does it double? Do you import the data from each (synced) camera or will the dual Scarlet configuration have a common storage that will record the data as one combined set? What software is available to seamlessly work with 3-D?

I confess to being a RED One 4K snob... but I am starting to believe that this Scarlet... that's some idea for a camera!

If you take the two image sets and format them in some graphics program to single frames such that they are in "crossed eyes" format on one image then you can view the proxies running or stopped without any hardware in color 3D, and edit in any 2D system, even view from a video tape on a projector in 3D without glasses or special hardware.

Two 2K images will fit side by side in a 4K image half height, i.e. 4096x1152 and such. If needed that 4096x1152 can be letter boxed into 4096x2048. The right image goes on the left, and the left image goes on the right.

If you need two images later you can just extract the 2K image pair from the finished CC 4K frames. In a 4K projector the "crossed eyes" pair can be projected for glasses with mirrors and filters or shutters.

You can see some examples of stereo 3D formatted images at my Web site, there are examples for most formats including "crossed eyes". Over-Under may give higher resoulution in the proxies viewing, but you would need a viewing hood with mirrors and lenses on your monitor.

Here is a link to the "crossed eyes" without letter box,

http://www.dancad3d.com/1024/DMXSM030.JPG

And in a letter box,

http://www.dancad3d.com/1024/DMXSM031.JPG

and Anamorphic for special projection lens,

http://www.dancad3d.com/1024/DMXSM040.JPG

The text for these examples is in Section: 4.20.51.0 under display modes 30, 31&32, and 40.

If you have not done "crossed eyes" before you may need to hold your finger mid-way between the monitor and your nose and fix your eyes on that, then move it in and out until you get the yellow border to show THREE images side by side, then slowly try to focus your eyes so that the center image becomes sharp then lower your finger out of the way. It may take some hours of trying before you can do this with ease, but once you learn how you can view stereo without aids in this way any time any place.

Some people have a weak eye, or cannot see stereo 3D, and may not know that they lack that skill. You may need glasses with pos or neg lenses if you are older in order to focus on the monitor screen while having your eyes fixed on a closer point, but if the monitor looks sharp before you cross eyes, you should be able to learn how to focus and point your eyes independently.

I also have examples of Anaglyph Stereoscopic Display Modes in Section: 4.20.52.0, but Anaglyph glasses you can purchase off the internet have weak filters and give ghost images. Some Anaglyph glasses have a lens with the filter to help see both colors in focus at the same time. You can improve the filters by taping LEE filter gels over the filters that are there, I give some filter numbers in my docs...

Other information about prism glasses and stuff is in my Docs as well. Shutter glasses format interleaved may work in an uncompressed AVI, but with compression some of the stereo information may get lost, so "crossed eyes" stereo format may be more fool proof.

 

[Bruce Allen]

I used the "crossed eyes" method a bit while doing the Hannah Montana title sequence - once my eyes were sick of every other method.

DanCad you'll be pleased to know 4K is not needed - if you're going squint you really don't notice much going on beyond 1K...

Managing a stereo pair and editing it isn't the tough part. It's the extra bits you'll have to deal with for additional FX and compositing that will add a whole new dimension (hehe) to your work requirements.

THIS IS VERY TRUE.

Bruce Allen
www.boacinema.com

 

[number6]

Thanks DANCAD3D, I think your website will be a valuable resource for 3D understanding.

But if you, Bruce, Jeff, or anyone else can answer something for me, I would appreciate it. What I'm wondering is, if you shoot a scene with a couple of Scarlets, synced, and mounted on some sort of yoke that RED will probably come up with that allows the cameras to maintain proper station even during zooms, would you be able to separate that footage and output it as though it were single camera footage, or alternatively output it as 3D footage?

I would assume this to be true, but having no experience with 3D, I don't really know. If true, boy would that open up the possibilities, at least in my mind.

edit: DANCAD3D... did the crossed eyes thing and while I get the three images, trying to concentrate on the center one just causes my eyes to uncross. Is this to be expected?

 

[Bruce Allen]

would you be able to separate that footage and output it as though it were single camera footage, or alternatively output it as 3D footage?

Yes. You can always make 2D from 3D by just outputting the left or right eye.

Bruce Allen
www.boacinema.com

 

[number6]

Thanks for confirmation

 

[Bruce Allen]

...forgot to mention that the only problem is your titles aren't perfectly centered!

For this and other reasons we actually output wider than we had to. Delivery was 1920x1080 but with the finishers (Fotokem) we settled on 2048x1080 - so they could nudge things left or right in their Quantel if they wanted to - without having to blow up the shot...

Bruce Allen
www.boacinema.com

 

[number6]

...forgot to mention that the only problem is your titles aren't perfectly centered!

For this and other reasons we actually output wider than we had to. Delivery was 1920x1080 but with the finishers (Fotokem) we settled on 2048x1080 - so they could nudge things left or right in their Quantel if they wanted to - without having to blow up the shot...

Bruce Allen
www.boacinema.com

Good point. But I should have been more clear by stating that I was talking about splitting footage before outputting the final product. I was thinking more along the lines of shooting in (3K?) 3D, then outputting to 2D while having the ability to go back and output to 3D later on. I think this gives the Indie movie-makers the ability to put their work out there as 2D while waiting for the 3D display world to catch up. Then, they (we) can re-release the movie as 3D and sell it again!:w00t:

 

[Dan Hudgins]

Stand back to fuse 3D "crossed eyes" images

Stand back to fuse 3D "crossed eyes" images

Thanks DANCAD3D, I think your website will be a valuable resource for 3D understanding.

But if you, Bruce, Jeff, or anyone else can answer something for me, I would appreciate it. What I'm wondering is, if you shoot a scene with a couple of Scarlets, synced, and mounted on some sort of yoke that RED will probably come up with that allows the cameras to maintain proper station even during zooms, would you be able to separate that footage and output it as though it were single camera footage, or alternatively output it as 3D footage?

I would assume this to be true, but having no experience with 3D, I don't really know. If true, boy would that open up the possibilities, at least in my mind.

edit: DANCAD3D... did the crossed eyes thing and while I get the three images, trying to concentrate on the center one just causes my eyes to uncross. Is this to be expected?

Yes, if you are too close it can be hard to fuse the images, if you stand back you can try to get the center image to stand still.

At first the center image will be out of focus since you will have you finger tip just under the center of the center image, then you need to learn to focus your eyes without moving them and the center image will slowly pull into focus.

The first few times you will lose the fussion when you try to move your eyes around the edges of the shape, but in time it should become easy and your eyes will lock on the center image very fast and you will not look at the side images.

You can cut a rectangle the same size as one of the images, and look through that opening at the monitor, that gives your eyes something to block off the side images, although you will have to focus past the black paper.

If you have three lenses (two small one big), you can make an arial image 3D viewer anyone can see 3D on, you put the monitor back, then two projection lenses focus the two images onto a large field lens, then you look at the field lens to see the 3D image floating in space through the field lens, no glasses required. You just need to get the focal lengths and image spacing right so that the two "exit pupils hit the viewers eyes, if they are too far appart or close together one eye's view will be black.

You can also use a large concave mirror (or fresnel lens) and two video projectors to make a large arial image 3D display that anyone can view 3D on without glasses.

---

BTW: To shoot with two cameras you can use a half silvered mirror, and two cameras at about 90 degrees, that lets you adjust the angle and pupil distance without having the cameras bump into each other, you then flip the image shot into the mirror...