Hi Mike, guys.
Sorry to slander his Lordship with the library accusation - it's not easy to write and there are libraries out there. Knowing this is in the LM code is wonderful as it means we can fine tune the output as mike suggests.
Output.
Commercially, having two images streams is ideal: Currently there is a standards war going on and SMPTE et al have yet to agree which will be the final distribution standard: There are some very big players out there and time is not on their side. Commercially, they need an answer.
For this reason, people seem to be rendering two images/movie streams then combining in whatever distribution format is needed.
TDVision seem to have cracked the turnkey problem and got it into the market place: It is a very strong contender for the standard, as it is part of the mpeg-4/H.264 standard, and (vitally important) is backward compatible with 2d existing systems. And its okay at 1080p60, the Holy Grail of the playback standards.
Last week I met the Hong Kong guys and they have a stereo edit system from japan, which works with two media streams, in real time. It runs on a dedicated 8 HD raid chassis
At this stage, I would not advise having an internal AS stereo codex - let the dust settle from the LA/Europe debate - expect the standard to take around 18 months to agree)
Side by side.
Having a side-by-side is useful as its good for rapid checking, especially when viewed as a 'cross-pair' (the left image is on the right, right image on the left: If you look at it cross eyed, you see the full depth). Sending crossed-pair test renders to clients/coworkers by email is also invaluable. No need for any special viewing systems. If its printed on glossy paper, cross pairs look great.
The answer is:
1: Two separate left and right outputs.
2: Cross-pair/ side by side is good for testing.
New Forum topic:
Over the last few days, I have come to the conclusion (from industry feedback) that HD3D is going to be huge, as big as the move from black and white to colour: The US TV channels are already experimenting with such broadcasts and Singapore is pouring megabucks into a state backed 3D facility. Worldwide the investment is in the $billion category - 90 features completed or in production in the last 12 months.
Considering AS can/will be able to output at 1080, in full colour, at 25/30/50/60 fps, it will be the first HD3D compatible animation program: Smith Micro should sit up and take notice of that: Flash and other rivals simply aren't or cannot be. Stereo-enabled on the box cover is going to turn heads. When is the next industry show at Vegas? This feature alone might dictate the next (earlier) release date.
So new topic forum. I think a new topic category is in order: The problems of how to do this, get things looking right will be a steep learning curve: from some rough early tests, I suspect the Track camera Z setting of 1.7321 is too close to the Z(0,0) point: If the camera rotates, distortion occurs at the edge of field - since the cameras are separated, this distortion is different: This might cause eye-strain/look odd. I backed the camera off to 4 and played with the zoom (Focal length/angle of view)settings and this killed the distortion: But the z layer separation then needs to be increased.
If we have a new topic area, any charts/ ideal settings we stumble upon can be kept in one area.
Development:
Whilst Mike might get the 3D renderer completed relatively quickly, we will be dependent upon Smith Micro and its commercial timetable on when v5.7/v6.0 is released.
I would propose we crack on with the script option until the next version is released. It will allow us to figure out all the requirements, things like where the output image stream is put: I favour a render directory where a new folder is created with a left and right sub-folder containing the output. Keeping track of multiple synced image streams is going to be tough and any automated housekeeping help from the program is going to be a huge benefit.
The question is whether Mike will be releasing an interim LM 3D script, based on the existing anaglyph code, or we continue with the current formula/script work. I am happy ether way, as we will have a tool to get to work on the real problem of the optics. A script now would means I can start generating a set of tutorials to release here.
Okay, back to the day job.
Great to have Mike in on the discussion - his input as always is invaluable. I have a really good feeling about where this is going. We are going to have tools to keep us right at the cutting edge.
Rhoel
BTW, the diagram if the rig is exactly how the script code is being constructed.
and the 60-66 mm offset: Fully agree on how hard that is to guesstimate - that was most of last weekend and I still have only the roughest of ideas of how to find an approximate workable equivalent.
--edit---
Sorry, some silly typos is this post: Corrected.
So I could render out other formats as well if that makes sense. What did you have in mind? Some possibilities would include rendering each eye as a separate movie/image sequence, or rendering both eyes side-by-side in full color in the same movie/image sequence. I'm not sure if there's a standard in say After Effects or Final Cut Pro for working with stereo footage. What would you guys like if you had your choice?
).
