I hope you can as well :)
It might be more doable today with a bunch of gopros all rigged to shoot.
Indianapolis, Indiana USA
On Jun 27, 2012, at 12:58 PM, Trausti Hraunfjord wrote:
You mean that I missed the "Headaches ahead" road sign? :)
I know that you have extensive experience in this area, but I still hope that we can get a solution.
On Wed, Jun 27, 2012 at 11:41 AM, Matt Smith <masmith@...>
It is the sync between the rows that is the issue. Even if you shoot.with multiple cameras, you can not mount the item in a way to get all of the rows. Additionally if you were.to shoot a "144" frame rotation, then you would need 72 video cameras to shoot all of the rows in unison. Remember that the object has to set on somethong.most of the time and you can not c.ring the cameras so that theu get every angle. I bring this all up because I have been down this road.in the past.
They do make great single row.rotations though.
-------- Original message --------
Subject: Re: [PanoToolsNG] Re: Large Object VR
From: Trausti Hraunfjord <trausti.hraunfjord@...
The video sequence version is meant for a fully automatic turntable that turns at a constant speed (motorized). That way, the issue with synching multiple rows will be a very minor issue. Manual rotation would pretty much be out of the question though. We also have other options on the table, which should give a much smaller file sizes (better compression) than when jpg's are used, and yet keep similar (or better) image quality. In theory (where it is at right now), 72 images (every 5°) total size should be no bigger than in a jpg project using 36 images (10°) of the same perceived image quality. That in and of itself would be a big improvement,
On Wed, Jun 27, 2012 at 7:51 AM, Matt Smith <masmith@...>
I can not in any way figure out how you are going to shoot the multi-rows. The sync up with the multiple rows will be difficult. Additionally, the amount of post edit in all of the frames will make the idea counter productive.
Good luck on it, I hope you figure out a few of the bumps in the road
Indianapolis, Indiana USA
On Jun 26, 2012, at 5:52 PM, Trausti Hraunfjord wrote:
That is a logical question to ask. My programmer is working on an Object Movie Generator (OMG), and it is based on the traditional image-by-image method as we are used to, but I wanted him to make one based on a video sequence, which would make the whole thing much more fluid and nice experience, and should also be a much lighter than a traditional object movie which would contain just as many frames of images. Of course the quality of the images might be lower, but all in all things should result in much better object movies. So the plan and intent was there (and still is)... but it seems as if only a server with a dedicated Flash Media Server running can provide the smoothness and ease of use (as it would be needed with a multi angle Object movie to change between levels in various video streams)... and a license for that used to be around $4500 ... hosting services that offer Flash Media Server can be had at around $100 per month... and $250 per month for interactive content streaming.... so it is not cheap.
We will be putting some more research into this as time allows, since I am absolutely sold on the concept since I got the idea a few years back.
On Tue, Jun 26, 2012 at 4:17 PM, Erik Krause <erik.krause@...>
BTW.: I always wondered why no video compression format is used for
object movies. It is far from optimal to store hundreds of single
images. There are even techniques to calculate intermediate images, f.e.
for artificial slow motion (twixtor, AviSynth MSU Frame Rate Conversion