Discussion in 'General Discussion' started by 1959nikos, Aug 20, 2012.
The projector itself projects a mirrored image, its a projectors option.
It can also be panorama, depending what tools he used to make that panorama photo he took it can be that there are discontinuations. I use Hugin myself and even it is really good making panoramas it too fails at times, probably depends from photos and how they overlap.
I base this on my experience with panoramas like http://gigapan.com/gigapans/47226
JTbo is right, it is a problem of photo shooting the panorama. The lines are in fact vertical. As nikos points out the projectors take care of reversing the image so they appear correct when back projected.
@Niko; the flat area on my screen is acheived by drapping the projection material across a window cut out in the plexiglass...with one solid piece of plexiglass you would likely wind up with on constant arc rather than the 'flatten bowl' profile I wanted
View attachment 3749
The 1:1.2 is the max zoom factor for that projector, not the throw ratio. The throw ratio for that projector is 1.5-1.8. So, according to this calculator if you want a screen width of 2m, you need to place that projector 3-3.6m away from the screen. (I'd aim for about half-way in the middle [3.3m] that way you have some room for error). Also, I'd recommend mocking it up in a 3d program if you have access to one (even SketchUp would be plenty fine) to make sure that you can position the projectors so the projector isn't being blocked by the back side of the screen itself (i.e. the left side of your semi-cylinder blocking the projector for the right side of the image, and so on...).
As far as the focus goes--no, the software can not affect focus. Focus is purely an optical property--which is where those expensive lenses come into play. You'll be forced to compromise on that, unfortunately.
Honestly, one of the solutions I'm liking best was (one of) the first one(s) that was recommended--the 3 flat panels set at 45° to each other to project onto. You'll get the least amount of distortion and lost resolution. When you're sitting in the driver's seat, you shouldn't really notice a difference between that and curved screens--provided rFactor truly assumes 45° for its 'multiview'. I'll try to put together an example to show exactly why.
I think that with the bowl shape, you have almost the same problem as triple screens have i.e. an abrupt transition from front to sides.
What Im thinking now is, if my approach, half cylinder, is a bit extreme in the end and better opt for an oval shape, something in between of your screen and half cylinder. Perhaps there are less focusing problems that way.
Got you. Thanks for info.
Im already thinking some compromises.
First an oval shape than cylindrical, second a smaller radius, so the difference from edges of screen to middle is smaller, hopefully making focusing problems lesser.
Even with an oval, the flat bowl shape provides less focal change from the front to the back edge than a constant arc would...the transition is not really a concern the way I have it set up with nthusim when the cars move from the flat to the rounded surface the distortion is not really perceptible.
View attachment 3750
Yes, thats true.
I am not following the logic here. When you are talking about a projected image onto a flat surface, the distance to a corner is a lot longer than the distance to the center. This in itself would present focal issues in my opinion. Likely this is taken care of in the lens design though. What I am thinking is that if the projected length remains constant to the screen from the lens, then the focus should remain constant as well on a curved display. Again, the lens design probably plays a huge role in this.
Will have to do a little digging to see what sort of lens is best suited for curved surfaces, their cost and if there is a large problem with rear projection onto a curved surface as the distance from center to outer would be far greater in that respect.
Now that you mention it...it does make sense.
Ill do some calculations about it and some reading if I can find.
As it seems, projectors project infinatelly the image they are seeing.
If you make vertical cuts along the central path of projection, you will get analogous images of the original image.
So, any image will have the same focus on every point of that vertical cut.
Otherwise, all screens would be curved lol.
(nice point though, I thought I solved my problem...)
Analysing this, its like having a picture in front of a tiny hole.
If you had a light exactly behind the picture, the beams of light would have to go through this tiny hole i.e. the lens of the projector, forming the exact mirror image of the picture, as they went to the other side of the lens.
Pure geometry at its best...(wish I could sketch it, but I think you got it)
as a guy said around here, my memory is all over the place, cause this is standard theory for cameras and I did shoot a lot sometime ago...
That sounds like the pinhole camera I made as a kid. Took quite a few pictures with that thing. I was amazed every time that it actually took pictures!
I do remember it as a minor school project.
You should warn us Mike LOL
Pic of curved screen running fr2
View attachment 3757
View attachment 3758
For me that looks really good, incredible good colours too. Side sections probably look to be same brightness to eye?
That bit of unfocus at edges might be actually good thing, but of course if one would use track-ir in conbination, then sides would need to be in focus too. That is probably where 2 or 3 projectors will help, but I really can't see any resolution issues in that pic. Of course monitor will have bit sharpe borders of things, but it really is noticeable only if stationary, when moving it is impossible to really notice much of such things.
But it is same for super high detailed cars and skins, you really don't notice 512 or 4096 textures at racing pace, especially in F1 cars, eye just is not that fast, so I would not worry about resolution too much.
I don't really think that is a curved screen, though I could be wrong on this. The image looks way too much like triple projector screens or even triple rear projection screens with the definite lines where they are going vertical. It is definitely a triple setup though.
But...he explains it all in his previous post
Found a nice calculator for optimum projector distance from screen.
In there, you put maker and type of projector.
Then you put width of screen, or height, or diagonal and it shows throw (optimal) distance of projector to screen.
I chose 90cm height, to 1.60cm width. Note, that is the projections width for 16:9 HD.
If screen is curved, actual width of screen for 180 degrees, is πd/2 i.e. 3.14 X 1.60 / 2 = 2.50m width and 0.90m high, which is quite adequate.
Optimal throw distance between projector and screen is 2.62m which is ok by me.
Interesting thing is, that the calculator gives a throw range between 2.4m and 2.88 which means that things stay focused between a distance of 2.88-2.4=0.5m , which is a number that should be taken seriously at the positioning of projector.
Finally, luminosity of screen at that range is very good, up to 207nits.
No it is a curved screen, the vertcal lines mark the transition between the portion of the screen that is backed by plexiglass and the portion that only has projection screen without plexiglass.
Well, its not really a curved screen honestly, but nearly. A curved screen is ONE screen with a steady and smooth curve. Your flat panel in the middle has no curve at all. Its really like a tripplescreen setup.
Have you tried the one ive posted , cause that would give you the way better immersion. I know remco tried curved screens before but he says it went nowhere near a screen 40mm infront of you.
I do also think you loose focus and light on the sidepanels.
Separate names with a comma.