Maybe some of the non-professionals night be interested on amateur setup I am using to create some video's as a hobby. Here is the link:YouTube - ComPH8's Channel (http://www.youtube.com/user/ComPH8#play/all/uploads-all/0/kYQFLm4Q71U) and http://www.youtube.com/watch?v=kYQFLm4Q71U.

I've improved on the amateur 3D rig, which works very well and is based on the Merlin. Here is a YouTube link: YouTube - Stereoscopic 3D setup with Steadicam Merlin and Sony AVCHD Camcorders (http://www.youtube.com/watch?v=THU5XeRIeOA) .Better quality (but anaglyphic - requiring glasses) are at these sites: Stabilized 3D Stereoscopic Rig - Red and Cyan Anaglyph on Vimeo and http://exposureroom.com/members/ComPH8.asp...2b20c9b928e0d7/. The system is quite light and portable.

Pavel, nice job. Can you tell me about the bar on which the camera are mounted? I need one.

Thanks

Thanks Paul. I assume you are asking about the Steadicam version. The main bar is made of aluminum. The dimensions are 6x30x145mm. It has two holes in the center area, for the attachment to a small lower piece with M4 screws. The lower piece is 6x30x22mm. In the center there is a 1/4-20 thread to fasten it to the stabilizer and be almost balanced. Other two M4 threads are for the attachment to the upper/longer piece. If you don't use the Steadicam stabilizer, the Cullmanns have 1/4-20 threads in the bottom plates as well, for attachment to tripods, so the lower piece would not be needed, just the upper connecting/allignment bar.

There are small plastic bridges inside the Cullmann quick adapters, that I removed.

I did it this way so that the assembly is light and simple. It took about an hour to make it using hand tools. There are commercialy available bars from really inexpensive ones to fancy expensive professional ones as well.

Some maybe interested but not know that YouTube now supports 1080p, including 3D. One must have fair amount of MIPS under the hood to play it smoothly though. YouTube - Merlin 3D SBS Audio 1080p (http://www.youtube.com/watch?v=bBvQ634a3ss)

Some maybe interested but not know that YouTube now supports 1080p, including 3D.

Ahoj, Pavel. I noticed YouTube asks you which of the different 3D viewing methods you prefer. Which begs the question how exactly did you encode your video. It would seem you gave it both the left and the right images in full. How?

Ahoj Adame. I encoded the video in side-by-side format. It actually looses half the horizontal resolution that way, but there are many other image quality degradation due to the bandwidth limitation of the YouTube anyway. I use the yt3d:enable=true tag, so the player outputs the video in about dozen formats of choice, depending on the display device or viewing preferences. To my knowledge they cannot deal with full images, but need to fit both L/R onto the same frame. They do have a tag for very flexible selection of the corners of the views, so one cane do, for example, over-under as well. Search of their site gives loads of explanations.

Thanks, Pavle, I got it now. I missed the part about the horizontal squeeze first time I read the instructions on YouTube.

I have a Loreo Lens-In-Cap for Canon EOS. It has two lenses and produces a left and right image but without the squeeze, so the result is a 3D image that is taller than its width. Once I get my Redrock 35 mm adapter, I want to use the Loreo with my EX3. So, I will have to squeeze it horizontally then. I guess I will have to upload a video that has 1080 lines but only 960 vertical lines then. Or maybe crop it to 540 horizontal lines and squeeze the vertical dimension to 960.

I will play with it once I receive my Redrock adapter.

I don't know their limits. I think that you may still be able to do it. They have a tag: yt3d:(left/right)=x1,y1,x2,y2; which allows to set the dimensions of the left and right views. I would give it a try. I don't think that you have to squeeze the sides. Worst case you should find a way to compress and composite the two views, but I don't think that you have to. There is also yt3d:aspect=x:y tag, that should help with that. You could also try to ask the "youtube pete". He made contacts via this thread: youtube 3d feature (yt3d:enable=true) - some issues - YouTube Help (http://www.google.com/support/forum/p/youtube/thread?tid=56b6f6f15dabf994&hl=en) . Your setup is actually cool, you should have the image split on the sensor already, correct? Alignment and various matchings of video characteristics are taken care of in the camera. I'd think that you are set with the Loreo. You have something like this, correct: Loreo Products - LOREO 3D Lens in a Cap (http://www.loreo.com/pages/products/loreo_3dcap.html) ? There is a Japanese fellow who does over/under, where squeeze is complete opposite from mine (x vs.y) and his video looks fine, so I don't see why no scaling should cause problems, as long as you can deliver both views in the same frame.

Thanks, that is very helpful. Yes, that is the Loreo lens I have.

Well, I got my Redrock adapter and, unfortunately, it has a serious design flaw which makes it unsuitable for the Loreo. The problem is that the Redrock is not centered properly. That is, the entire image is shifted to the left. Because of that, the center of the image is to the left of the center of the frame.

While the Loreo can fill up the full 16:9 frame, the Redrock design flaw makes the left image smaller than the right image.

When I asked about it at the Redrock forum, they quickly moved my question to a private area accessible only by those who have already bought their product. It turns out this is a very common problem with them, but they blame it on all the camera manufacturers, including Sony.

This is very disappointing, as one of the main reasons I bought their adapter was to use it with my Loreo.

OK, I have spent this entire week writing a Sony Vegas plug-in to allow me to cut out parts of the image so I can use the Loreo with the Redrock adapter. Today I shot some quick footage, processed it with the plug-in and uploaded it to YouTube (http://www.youtube.com/watch?v=_yJyABy3BV8).

That has shown me that I need to add a little more functionality to the plug-in before I can have some half-decent 3D videos made with the Redrock/Loreo combination. I will keep you posted of my progress.

By the way, I named the plug-in Bororo 3D in honor of one of my favorite movies, Akce Bororo.

Actually I like it. The alignment is good. Impressive result, I am interested in your progress.

The alignment is good.

Thank you. I meant that I need to add more functionality to the plug-in even though it worked for this video.

I have added a little more functionality to my plug-in and here is another video (http://www.youtube.com/watch?v=xF4NZJjE94k) processed through it. Now, if I could somehow convince Visual Studio to compile a 64-bit version, I would be happy. Right now I can only use it with the 32-bit version of Sony Vegas.

This is the first Sony Vegas plug-in I ever wrote. I want to write two more which will allow additional 3D features. When I have all three finished, I will make them available for download. That is, if anyone is interested.

Also, is it just my computer being too slow, or is the video a bit jerky?

I am interested in your plug-ins. This 3D technology path is lot more difficult to transition to than other technologies were. There are just too many things the producer can do badly. I think that the player makes the video bit jerky, perhaps stream rate at 1080p. Try to publish the same thing in anaglyph and see the smoothness. I think it will improve. Thanks for sharing your work.

Very well, I'll keep working on those plug-ins, though they will only be for the 32-bit version of Vegas, unfortunately.

I was able to incorporate my second plug-in into the first one (here is an example of what it does (http://www.youtube.com/watch?v=2HJTeruEFAA)). That way I do not have to write three plug-ins, only two. It is 3 am now, so I think I'll call it day and go to bed. :)

I have just discovered a fluke in that YouTube 3D player. I made yet another test video (http://www.youtube.com/watch?v=IfwphTAxwyg) (this time with a static 3D photograph I took last summer). It worked just fine at first. Then I thought it would be nice to add some sound to it and used YouTube's "AudioSwap".

Now it has some wild music in the soundtrack but no longer shows in 3D. Very strange.

By the way, this morning I watched all my test videos with the Stereoscopic Player software, and they all looked good, no retinal rivalry which I have been getting on YouTube. So all the problems I have been having with watching my tests are due to the limitations of the YouTube 3D player. This is good to know because I have been blaming myself.

As for the plug-in, the first three parts (combined in two plug-ins) are pretty much finished. The first part takes videos shot with the Loreo adapter and makes sure the left and right sides are positioned properly within the video frame.

The second one takes a 2D track and moves it closer to the viewer or away from him in 3D. This is useful when adding text to a 3D video.

The third one works on two tracks (left and right) and combines them into a single stereoscopic track either by shrinking them to half size and positioning them next to each other or above each other, or by creating an anaglyph. The video I mention above, the one that no longer shows in 3D on YouTube is an example of that.

And this test video (http://www.youtube.com/watch?v=U3lXhnq_o2Q) was made using all three parts of the plug-in. I used the first one to separate the two views into two separate tracks, the second to recombine them into one track but shrunk to half-size so it can get some simple background on YouTube, and the third one to add some text.

Today I am going to be working on the fourth part, which I was not considering when I first started working on this plug-in. Namely, I am going to try to create some 3D transitions. I am not promising anything because I have to see if it really works first. And if it does not, I will remove the fourth part. The plug-in will still be quite useful without it.

I have added a transition to the plug-in. It was rather complicated to figure out how to do it so it really is a 3D transition, especially since it is a very basic transition, just a small square appearing on the screen and growing until it fills the frame.

I am not even sure it is obvious that it really is a 3D transition, so I need your opinion (I have been staring at it for so long that I am no longer objective about it and, as usual, tend to be very critical of it) whether working on those transitions is worth it.

I have made a very simple test, which shows a stereoscopic photograph (I wanted it to fill the screen and my videos only fill half of it) for two seconds, then takes ten seconds to transit to another photograph, which stays on for another two seconds. I made the transition last so long, so you can pause it easily and examine it, though, admittedly, examining it too much is what is making me so critical of it.

YouTube | Bororo 3D Transition Test (http://www.youtube.com/watch?v=IRnDk5iAYMQ)

I'd really like some feedback here. I am putting a lot of time and energy into this plug-in, so I would like to know if anyone might find it useful...

Anyway, I have tried a different method of implementing the 3D transition. This one seems somewhat more clearly stereoscopic but it is causing me serious retinal rivalry. You can see it here (http://www.youtube.com/watch?v=9CyhXJU-knA). Does it seem more stereoscopic than the one mentioned in my last message before this? Is it too harsh on the eyes? I wish I had a way of seeing it with a method other than anaglyphic glasses.

I have been trying to decide which of the two methods to put in when I release the plug-in for download. Right now I am leaning towards leaving both methods in and letting the user decide which one to use. But if the second method causes more eye strain, will I be contributing to people having headaches just because some less experienced user will choose the wrong method for his videos?