dan kessler

The scratches and random color fogging are a simulated film look. In reality, this kind of fogging is usually only seen at the head or tail of a roll, due to loading and unloading the camera.

All the old issues again. Even as a little kid I never liked TV soaps because they looked "too real." Movies (film) always looked better.

Arri 35-III's have a single registration pin, but the BL models have always had dual-pin registration. Regarding 35mm short-end prices, here's an excerpt from a quote I got back in February from Comtel (818) 450-1122 : The price for the film depends on the type of film and the length. For example, 5219 is the most popular stock. If you want to buy some film at lengths between 300' and 390', the cost would be .10 cents per foot. Shorter lengths would be .05 cents per foot. For a less popular stock, the prices would be lower for longer lengths. You can get lengths of 5201 over 400' for just .08 cents per foot.

You don't say what kind of lights you're using. Basically, with reflective or transparent subjects like glass, water, etc., you want to rely more on soft lighting. Brilliantly lit areas are really just reflections or refractions of the large, soft sources. Opaque subjects with surface texture and color respond better to hard, directional lighting, like fresnels.

I wonder if the technique is being correctly described here. Rotoscoping is essentially a process of hand-tracing elements within a shot for matting purposes or hand-painting elements to match. It is laboriously and painstakingly executed frame-by-frame, and while the results can be superb, it is very unlikely that this is how the sample video was done. Where actors are to be composited over background plates or incorporated into other visual effects, they are usually filmed against a blue or green screen. It's the colored surround that provides the basis for a relatively swift matte extraction. If you had to do it by hand, it would take agonizing months to complete.

Yep, film. Like Mark Twain once told an audience, "Reports of my death have been greatly exaggerated."

The Nagra 4.2 is fine, the workhorse of film sound until digital took over, but that isn't really the issue here. Sure, you could record your music onto the Nagra, or straight onto your mag recorder, for that matter. Your original is the album, in this case. As far as your stereo output into mono is concerned, why don't you use your mixing board? The Nagra was typically used to record voice in sync with the film camera. The sound was then transferred to mag film. The mag film is what you cut and mix from, not the Nagra 1/4" What you call a + b is the right idea for cutting mag tracks. All the gaps between sound takes are filled with leader. Not black leader, but white, and make sure you splice it so the base runs over the heads, not the emulsion. A lot of books and articles have been written on the fine art of sound cutting and mixing. Time to dig into them.

Hm, this is a little confusing. You say you have a print. Okay, does that mean you have cut your soundtrack the old way, on mag film, to match your workprint? If so, then you should have separate mag tracks for dialogue, effects, music, etc. Any syncing would be accomplished by physically cutting or repositioning the mag tape in a gang synchronizer to match the picture. You or someone else will do an interlock mix from those tracks, prior to sending it off to the lab for an optical track and marriage to your picture. Or, if you were cutting the film online, you wouldn't have mag tracks, but you would still have separate tracks that you could edit, then mix out. Is this what you are doing? Trying to patch gear together as you are describing doesn't make sense to me, otherwise.

So you like grainy texture, eh? Check this out:

Of course, all of this begs the question, what will a rig like this give you that you can't get with a single camera and After Effects?

As Chris indicated, your drawing needs to be changed. The camera that is presently on the left will be moved below the prism along the same axis as the lens. The camera on the right will stay as is. One of many other issues to consider is the back focal length, or flange focal distance. With a PL mount, you are using lenses that are designed to focus an image 52mm behind the front surface of the lens port. You must carefully compute the total distance through the prism to the image sensors. That one problem alone could scuttle your design. If you want the image position to be exactly the same for both cameras, then you must provide the means to align the rig with pixel precision, then lock up ROCK SOLID. Nothing can move. The tiniest wiggle will ruin it. Are you good with machine tools?

Well, I guess you didn't like the answer I gave you before, but I'm not sure you understand what a beam splitter is, either. Beam splitters are nothing more than partial reflectors. They "split" a light beam by transmitting part of it while reflecting the rest of it. They do not include lenses. They can be positioned behind or between lenses, as they are in film and video cameras, for various purposes. Whatever it is you are trying to build, the beamsplitter will be either a prism, a plate or a pellicle. There may be more than one beamsplitter in the optical path, but any other components, lenses or otherwise, are not really part of the beamsplitter. You need to describe your project in more depth to get any better answers, I think.

There are prism-type (like the Bolex), plate-type and pellicles. All can be made with varying degrees of reflection and transmission. The plate-type might introduce ghost images, depending on the thickness of the glass substrate. Pellicles are extremely thin membranes (to avoid ghosting),but they are very delicate. Some sources for new or surplus: Rolyn Optics Edmund Scientific Anchor Optics Surplus Shed Google will get you to all of them and also link you to all the info you can handle.

It sure wouldn't be the first time I've seen stuff broken up and sold piecemeal on ebay. Nothing more than a way to milk buyers for more money. I just keep walking.

Where were you back in '98? Dean Devlin and Roland Emmerich already did a big-budget CG Godzilla

Models inhabit the real world, whereas everything cg is a mathematical simulation. That's always harder to pull off. Not that it can't be done, because there's plenty of cg work that no one ever realized was cg. IMO, 90% of the cg look is due to over-lighting. After investing all that work in modeling and shading, it's as if they can't bear the thought of any of it falling into the shadows. They want to make sure you see it! Plus, too often cg fx goes so far over the top that it gives itself away purely on the fact that it couldn't be done any other way. It's really a question of discipline. Don't get me started; I'm a guy who will take the 1933 King Kong over the 2005 King Kong every time.

You're in NZ, didn't realize that. Guess that means that Weta is your main source of CG work. My comments were primarily in reference to LA, but apply to the U.S. in general. If you're in Canada or India or China, there are lots more opportunities.

Don't want to be a spoil-sport, but starting at zero with Maya is not likely to pay off as fast as you hope. No need to tell experienced people here that the real heyday of CG in this town is behind us. Many studios are gone and many jobs have fled these shores, like so many other industries. The few biggies that remain all have satellite studios in other parts of the world and are investing heavily in the (less expensive) talent in those locales. If you want to work for them, you need to be capable of working at a level that sets you apart.

locked off camera - check clean plate - check actors and table against green screen - check plenty of coverage - check so far, so good Now go over everything with the editor and compositor BEFORE you shoot

Now that I think of it, I think cameras with beamsplitter viewfinders require the use of a circular polarizer. Don't know if that would work in this particular case. Mirror-type viewfinders are no problem.

I haven't used Blender, but I think you want to stick with that for all your modeling and animating. You also need to load in your background plates as reference for both your animation and lighting. Make everything match as closely as you can in Blender. Then, render your CG out to a standard file format, like tiff, for example. Then, pull your background and CG sequences into AE. Comping shouldn't be much more complicated than A over B with some color and contrast tweaking. Anything beyond that depends on the specifics of your shot. This is the typical workflow for a shot like this.

You could definitely try a set-up like the one I described above. You put sheet polarizers in front of your lights will the same filter orientation and a polarizing filter in front of the camera lens. Rotate the lens filter as needed to knock down the reflections on your subject. In this situation, you can easily compensate for the light loss. Don't know what kind of camera you are using, but beamsplitter-type viewfinders might act weird with polarized light. Not a show-stopper, but be aware and devise a work-around, if necessary.

Light is light. In the old days of downshooter animation stands, it was standard practice to put polarizing filters on the lights and the camera lens to kill reflections in the glass platen and acetate cels.

First of all, yes, it is easier to build your model in a 3d modeler than it is to build an actual model, light it against a greenscreen, shoot it, extract a matte, and then match the elements to your bg plate. Secondly, CG comps of this type usually follow a rather conventional workflow How about more details as what you specifically tried to do and what didn't work?

Maybe take the smoke density down a bit, a little more noise in the transparency to break up the uniformity; it's a little too dark, keep it on the light side. Less is more. (just like real vfx dailies!)