| Author | Message | |||
     cbrack Rating: N/A |
Proposition: Locate and Distribute a 15" and 17" LCD Lenticular Kit for Autostereo 3D. it would basically be the same as this from iart3d.com: http://www.iart3d.com/Products/Autostereoscopic%20Display/Auto3D%20Display_Eng.htm see disucssion about the kit from this board: http://www.stereo3d.com/discus/messages/24/2332.html?1068596187 the iart product is completely insane at $449 for a 15" version. This is only the "kit" which includes a lenticular sheet to place over your LCD and a direct 3d driver (nvidia can do the same thing for free). I am so outraged at this cost that i would like to make my own for around $10 and give instructions on how to set it up w/ nvidia interlaced 3D. I need your help in locating and selecting the exact type of lentiuclar sheet designed to work with standard LCDs. There would be two models, one for LCDs with a dot pitch of 0.297 (most 15" LCDs) and the other for larger LCDs, (0.28 dpi?). There are more than enough companies selling custom lenticular sheets with varying LPI, but most are only sold in quantities of 100+, plus the extra shipping charges for overseas. I would be willing to make a bulk purchase and distribute the kits from the states if people are willing to buy them. Don't worry, i'm not out to make a profit, i'm just frustrated that the technology is there to make a good LCD Autostereo kit on the cheap and there are no vendors filling the void. Some issues that I would like to discuss before I make a bulk purchase include: 1. How many of you would be interested in buying a lenticular kit if the price is below $15, shipping included? 2. What must the LPI (lenticules per inch) be to exactly correspond to the dpi of the monitor? I have a few equations below that might be helpful. 3. Desired Viewing Angle (narrow vs. wide) and thickness (48-150 mils)? 4. Type of plastic (PETG, PVC, Acrylic) 5. Best method of attaching lens to LCD monitor 6. Best way of fine tuning/calibrating lens placement The method I currently have for determining LPI is: 1. assume a native resolution of 1024x768 for the 15" monitor. 2. assume a dpi value of 0.297mm as per the manufacture's specs. 3. calculate monitor width and height width: (0.297 * 1024 = 304.128mm) height: (0.297 * 768 = 228.096mm) 4. convert to inches width: (304.128 / 25.4 = 11.9735 inches) height: (228.096 / 25.4 = 8.9801 inches) 5. this should be the exact dimensions of your LCD monitor, verify w/ a ruler. 6. here's where i'm not sure...calculate LPI: (monitor width in inches / native resolution width / 2?) (11.9735 / 1024 / 2 = 42.76094 LPI) i'm not sure where i picked up the 'divided by 2' - i must have read that somewhere and included it in the equation. please someone correct me if this is wrong. background: I successfully made a barrier kit from John's great instructions and it worked great, but would like to get a better solution for autostereo gaming in landscape mode (the barrier kit is designed for portrait mode, and thus not sutiable for gaming). questions/comments: please email me directly if you are interested so i can get an idea of how many people would buy a custom lenticular sheet for their LCDs. my email is: cbrack at myrealbox dot com references/resources: http://www.humaneyes.com/resources/lenticular-basics.php#1 vendors: http://www.iart3d.com/Products/Lenticular/Lenticular_Eng.htm http://www.depthography.com/screen.html (in state shipping!) | |||
| John Billingham Rating: N/A |
I would be happy to "review" your "kit" Please send complimentary sample! I would still rather see a lenticular screen that EXACTLY matched the pixel pitch of an LCD,that it be used with two images only (hence narrow "sweet spot") and again, "portrait mode" will avoid color problems. There are a very few LCDs that have horizontal "color bars" comprising their pixels THESE could be used for "landscape". Best, John Billingham | |||
| clyde Rating: N/A |
count me in for a couple.. actually much if the first one works out. What the heck, Im a content creator for x3d displays, might as well re-purpose that and kick start a new market ;) regards Clyde | |||
| cbrack Rating: N/A |
complimentary samples for everyone who helps me find the perfect lenticular sheet  John, do you think i'll run into the same problem that the barrier grid has with landscape mode (pixel bleeding)? i'm hoping that the lenticules will be better at redirecting the L and R eyes to the image. Also, there's no way to avoid resolution halving w/ the interlaced method, so gaming will suffer on the 15' LCDs. I agree, we need to find a company that can exactly match the pixel pitch. anyone ever run across a custom LPI vendor? Clyde, do the x3d displays have horizontal color bars? Is that the main 'trick' that the autostereo LCD vendors use to sell an out-of-the-box 3D monitor? cbrack | |||
| clyde Rating: N/A |
cbrack, x3d dosnt use lenticular technology. They use something (i forget the exact name) thats a color filter array.. that displaces the individual rgb pixels to different points in space. This is a custom designed filter by/for them and wont be available in the market. Stereographics use a lenticular sheet and its placed digonally on the screen to overcome the artifact/moire effects Regards Clyde | |||
| john billingham Rating: N/A |
Yo, CB, I stongly recommend vertical format. Have you tried "games" with some of the "windows rotating" software that comes with many of the LCD monitors?? John Billingham | |||
| cbrack Rating: N/A |
hey john, ya, when i made a good barrier for one of my LCDs I tried a couple different rotating scenarios. i even got so fed up w/ it not working that i put a post that never got any bites: http://www.stereo3d.com/discus/messages/21/2586.html?#POST12352 problem is, that the nvidia driver seems to only work through direct 3D, which has to be in full screen mode. maybe the latest nvidia stereo SDK allows for some tweaking to allow vertical interlaced. clyde, never heard of diagonal lenticular method, any good resources on how this actually works - thought i was up to date on lent  -cbrack | |||
| Puppet Kite Kid Rating: N/A |
Vertical interlacing is easy to do to stereoscopic animations, movies and images... at least you can test your autostereo display this way. You can do it with VirtualDub at any (one)resolution, StereoMovie Player (at fullscreen view only, I think), Stereoscopic Player (which I think actually resizes the image on-the-fly), or Adobe After Effects at any (one) resolution (with a few separate renders, using separate perspectives as the source). PKK | |||
| Puppet Kite Kid Rating: N/A |
...and maybe AviSynth, using one script, but I haven't actually tried this. | |||
| cbrack Rating: N/A |
pkk, i tried stereoMovie player on my vertical barrier grid and you are correct, it works great for images, movies, and animations. just was never successful w/ 3D gaming - which is kind of weird in portrait mode anyway even if the nvidia drivers supported it, i would probably not play a game in full portrait mode, but instead play games in 4:3 or maybe an equal ratio with windowed mode. so the question would be, is autostereo worth it if you have to play games w/ halved resolution (due to interlacing) and w/ reduced screen size (due to required LCD rotation). -cbrack | |||
| clyde Rating: N/A |
damn! I typed a long message and then lost it. in short Cbrack.. go to www.stereographics.com and then see th e white papers section for valuable info regards Clyde | |||
| Hd Rating: N/A |
Yeahh chrack, you're definitaly my new hero!!! I would buy 5 or 10 of them and spend you 100$ extra for your effort to make this happen!!! really cool;-) | |||
| Anonymous Rating: N/A |
Funny, I just found out that there is a big big industry producing this materials for 'lenticular postcards' and such stuff! | |||
| Puppet Kite Kid Rating: N/A |
The "half resolution" viewing with vertical interlacing on an autostereo display is identical to line-blanking (or "interlaced" or "interleaved") with shutterglasses and a CRT, _but_ a good LCD display should be brighter and more clear than a CRT, so I think the autostereo display would look *better*, IMO. I personally have never like the idea of rotating the monitor 90 degrees, though. I'm not sure I would get used to that. I guess my mind is "in tune" with wide-screens... even a 4:3 display seems to horizontally crop the imagery too much, IMO. Rotating a 4:3 display to even a more "cropped" format would kill it for me. Heck, I want an autostereo wide-screen :-) Now _that_ would really be something, and 16:9 3D movies would be total killer :-) I would love one of your kits, too, if you can somehow get the job done without rotating the screen :-) PKK | |||
| dep3d Rating: N/A |
tooling charge for an acurate lenticular is $30-50,000. even the pressing charge is more than $15. you are are going to have to sell alot for $15 to get back your investment. I am working on a two-view 17" 3d monitor with a precision barrier installed and later a removeable barrier kit that should sell for a few hundred dollars. may even have some nvidia drivers (not sharp) to go with it. | |||
| cbrack Rating: N/A |
hey dep3d, couple of questions... have you been running into problems with your removeable barrier kit w/ the LCD in the regular landscape position and pixel bleeding? if you've already researched the lenticular angle and decided to go w/ a barrier grid instead, what are the main reasons? if the tooling charge is the main obsticle, then i believe that you can work with a lenticular sheet that has a close match to the right LPI and make minor adjustments in the distance between the lenticular and the LCD (maybe w/ an exact spaced plexi glass between the two). You can calculate this distance for barrier grids w/ the help of this site: http://www.ponies.me.uk/lentikit/wiki/index.php?title=Barrier%20Screens i imagine the same math will apply to a lenticular sheet. -cbrack | |||
| cbrack Rating: N/A |
clyde, thanks for the link. need to go back to school for a lot those white papers, but i did find what you were mentioning from this paper: http://www.stereographics.com/x_downloads/whitepapers/synthagram/SynthaGram_LL_031904.pdf particularly here: "Winnek first realized that angling the lenticules so they were not verticalgoing, but rather tipped at some angle, had important advantages in terms of reducing moiré, and producing a sharper image. He did it in the context of being able to make photographic duplicates from lenticular images and he did this decades before there were flat panel displays. His work can be directly applied to flat panels because it reduces or eliminates the optical moiré produced by imaging sub-pixels through a vertical-going lenticular sheet. When placing a lenticular sheet on top of a flat panel color banding and the horizontal magnification of the pixel interstices becomes overwhelming." doubt i can find a vendor selling angled lenticulars, probably have to stick w/ portrait mode. this is getting interesting... -cbrack | |||
| Hd Rating: N/A |
by the way.. !!! W I D E S C R E E N S !!! ..mostly of all existing HDTV (1920 x 1080) -TFT's (including Benq, Nec, Hp, Sony, Apple 23" & 30" etc.) have a dotpitch of 0.258 mm !!! Let's try to find a lenticluar screen manufacturer that can handle 49,224806201550387596899224806202 LPI;) (25,4 mm : 0.258 mm dotpitch : 2 = 49,22.. LPI) ?? | |||
| Hd Rating: N/A |
http://www.research.philips.com/technologies/ display/3d/downloads/spie96-1463.pdf 1/8 Multiview 3D-LCD, found on google! | |||
| Anonymous Rating: N/A |
http://www.research.philips.com/technologies/display/3d/downloads/spie96-1463.pdf | |||
| clyde Rating: N/A |
Cbrack! you dont have to ask the lent manufacturer to make diagonal sheets .. You buy regular one and just cut them at the angle sure there will be wastage involved, but what the heck .. @DEP .. Here's some ideas that im sharing, due to lack of finance to do them myself.. but would like a piece of the pie if you think its edible .. I was thinking.. using an etched LCD "barrier" is the best bet. this was your paving the option for true 3D broadcast television. You know.. you insert a think film lcd sheet with Liwuid crystal Black lines. This should be made to be "turned /off/on" by some software on a chip form. In the near future, advertisers could "send" their advert preceeded by a "switch on" signal in the broadcast itself, this would put the display into 3d mode and have products floating in your living room. after the advert is over, swith off send the display into normal 2D mode. This same "addressability" of the barrier can be used in conventional autostereo content, whenever fine resolution is needed ...for viewing text for instance Let me know what you think ;) regards Clyde | |||
| clyde Rating: N/A |
errrm .. typo .. think film lcd = thin film lcd.. Clyde | |||
| Hd Rating: N/A |
Ugh, just using another LCD above as a barrier grid is a really cool idea, clyde;) | |||
| Hd Rating: N/A |
+ I got question now: I just thought if these lenticluar-postcards material will handle 2 eye stereo? I've never seen such a lenticular postcard with a real 'stereo' effect.. just with flipping viewangle.. ? | |||
| clyde Rating: N/A |
there are Stereo 3d lenticular postcards Hd. lat one i saw was the cover of the dvd in the local video store for the movie "day after tomorrow" (i think thats the name of the movie.. the one where the flood takes out newyork and the rest of the world. Regards clyde | |||
| Peter Wimmer Rating: N/A |
NVIDIA's stereo driver can be used to create interlaced patterns both in rotated and non-rotated format. You have to apply the following registry hack to enable all stereo types: ------------------------ REGEDIT4 [HKEY_LOCAL_MACHINE\SOFTWARE\NVIDIA Corporation\Global\Stereo3D] "ShowAllViewerTypes"=dword:00000001 ------------------------ Then, you can select both horizontal and vertical interleaved format:   Both stereo types are compatible with the driver's rotate functionality. The only problem is that some application except 4:3 instead of 3:4 resolutions, including the medical test image and the test application. These don't work. Other applications work perfectly, for example Stereoscopic Player in NVIDIA mode (View->Viewing Method->NVIDIA Stereo Driver), giving best hardware accelerated video performance on your autostereo display!  | |||
| clyde Rating: N/A |
The more I think of it ... the more I realize.. Peter's the one whos eventually gonna make a multiview workable option for auto stereo displays ;) | |||
| cbrack Rating: N/A |
peter, huge thanks for verifying that portrait interlaced is actually working w/ nvidia - thought i had tried everything, but now i'm going to break out my barrier grid and see how enjoyable the gaming experience is. next item to experiment is: can i get my LCD (while in portrait mode and set to native resolution 1280x1024), force the ratio to 4:3. in other words, the portrait mode is actually 1024x1280 - 1024 pixels wide, and if i wanted a traditional gaming ratio, i would have to be content with 1024x768, effectively cutting off 256 pixels from the top and 256 from the bottom. i'll see if my LCD supports this.... -cbrack | |||
| Hd Rating: N/A |
Really cool, it's going further here!!! Anyone know a manufacturer capable of 49,224 LPI for my BenQ FP231W HD display ?? Can't wait for playing HL2 in full stereo HD;P (it's the best and cheapest;)) http://www.pricescan.com/items/item160072.asp | |||
| cbrack Rating: N/A |
clyde, nice idea on the etched LCD barrier. can you give me an idea on what "cut them at an angle" means for the lenticular sheets. do you mean rotate the square sheet at some angle and then trim off the excess? sorry, bit confused -cbrack | |||
| cbrack Rating: N/A |
hd, wish i had one of those widesreens. iart3d.com sells a lent that might work if minor distance adjustments are made: 20"x15.7" Landscape LPI: 43 thickness:2.0mm heck, has anyone tried using a lenticular lens that closely matches LCD dpi? -cbrack | |||
| Hd Rating: N/A |
Ahmm.. minor distance adjustments?? 43 vs 49,22 LPI are some really big adjustments.. and actually I thought your topic was about 'finding a cheap plastic';? something I found: http://www.starosta.com/3dshowcase/3dartbits/lenticular.jpg | |||
| Hd Rating: N/A |
..anyone know a link with a comparison of the different lenticluar techniques? Actually I would like to know the 'physical' difference between a lenticluar sheet for 'stereo' vs. 'multiview animation' Can anyone explain this to me? | |||
| Anonymous Rating: N/A |
..wrong link.. www.lpc-world.com | |||
| Hd Rating: N/A |
Just found a got comparison of the existing techniques; http://www.langegraphics.com/Version2.0/FilePrep/Lenticular_Prep.htm | |||
| Hd Rating: N/A |
+ another manufacturer with 10, 15, 20, 30, 40, 50, 60, 62, 75, 77, 90, 100, 120,160 LPI and custom sizes available: http://www.skywings.hk/products.htm | |||
| clyde Rating: N/A |
cbrack.. thats exactly what i meant.. rotate the sheet ..and trim off the excess Cheers Clyde | |||
| Hd Rating: N/A |
Ok, here a overview of the most common used dotpitch's from current TFT displays available: ----- (30") / 2560 x 1600 / 0,250mm = 50,80/0,258(?) (23") / 1920 x 1080 / 0,258mm = 49,224 LPI (20.1")/ 1600 x 1200 / 0,255mm = 49,803 LPI (19") / 1280 x 1024 / 0,294mm = 43,197 LPI (17") / 1280 x 1024 / 0,264mm = 48,106 LPI (15") / 1024 x 768 / 0,297mm = 42,760 LPI ..Looks like (if the Apple specs of 0,250mm are correct) the Apple 30" is the only display that would work with a sheet of 50 LPI with some distance adjustments, or maybe it matches exactly through the outer viewangle? Anyone here know how to calculate the exact LPI scale needed to match the viewangle with the outer pixels??? Here again the manufacturer able to deliver 50 LPI, for example: http://www.skywings.hk/products.htm !!!Let's go further with this topic!!! ------------------------------------- | |||
| Hd Rating: N/A |
'AN' overview, by the way.. | |||
| clyde Rating: N/A |
Hd wrote..."..anyone know a link with a comparison of the different lenticluar techniques? Actually I would like to know the 'physical' difference between a lenticluar sheet for 'stereo' vs. 'multiview animation' Can anyone explain this to me? "..... HD im assuming you mean "flip type or motion based lenticular graphics" when you say Multiview animation... If that is true, then: In normal lenticular graphics (its a whole industry in itself.. postcards, posters and billboards) are said to be multivie, or flip .. when instead of using a stereoscopic image, you use a sequence of images.. example .. an animation of a girl "winking" at you. if this was say 9-12 frames long each frame/picture showing progressively her eye shutting ..then when you viewd the resulting interlaced picture placed behind a lenticular lens.. depending on your head position you would see the "animation". Now in lenticular posters, Stereo means that you have 4 -9views minimum of a stereoscopic image. there is software that does a decent job of creating these in-between views from a stereo pair. Search for BAS-Relief and 3D on google. The inbetween are needed in a lenticular poster creation as they are needed for Autostereo Monitors, so that the "3d sweet spot" is larger. Of course if you use just the original 2 images for a lenticular poster, you will see it in 3d when interlaced but that will need you to be "head on" or at another "sweet spot zone" depending on the LPI of the lenticular sheet. usually for 3d in lent posters/billboards.. LPI of 15 -35 is used and for flip 40lpi and upwards. Hope this helps Cheers Clyde | |||
| Hd Rating: N/A |
Ahm ok, another (last) question: -At a 9-view 'stereo' lenticluar image/or display, how many pixels/imageparts are under a single lense? 9, where a single view stereo image has only 2?? Do you know a graphic showing this relation? ..Just havn't understand this till now;) And how big is the maximum sweetspot of a single stereo view? 2cm?? | |||
| clyde Rating: N/A |
hmm part one of yr questions a bit hard to answer by typing in reams of text.. Try googling "lenticular Graphics software" you'll come up with the illustrations. Part two .. I thinks its around 2-3 inches. Regards Clyde | |||
| cbrack Rating: N/A |
hd, dep3d shot down the idea of precision LPI creation, citing that the tooling charge for an acurate lenticular is $30-50,000. i can tell he knows more than i do about this, so i currently believe him. will check w/ vendors once our "vendors to call" list is complete. Thanks by the way for finding more vendors - think our count is up to seven or eight that I will seriously call and discuss prices with. back to the LPI problem, i believe that the distance between the lenticular sheet and the LCD (i.e. barrier distance) might enable us to use a close match lent screen. check out the calculations from this site: http://www.ponies.me.uk/lentikit/wiki/index.php?title=Barrier%20Screens i'm going to do some verifications w/ the barrier distance formula. i'll start by printing a barrier grid set to some of the common LPI values seen on the vendor sites, and then see how large the barrier distance needs to be to get a proper line up. just need to get some very thin sheets of plexiglass for the experiment. will post results soon... -cbrack | |||
| cbrack Rating: N/A |
hd, nice list of LCD DPIs. quick question - was wondering if we could get away w/ using a smaller resolution than the native in order to 'fit' the pixels into a standard LPI. this would be similar to lenticular photography where you start w/ a fixed LPI lent sheet and then resize your picture to exactly fit the LPI. in other words, if you resize a native 1600x1200 lcd (.255dpi) to 1024x768, you have essentially just changed your DPI to .397mm and the needed LPI then changes as well from 49.80 to 31.96. i tried a few calcs on the above lcd list but didn't get any exact even matches (40,50,60, etc.) is this line of reasoning even correct? i don't fully understand how resolution resizing occurs at the pixel level. -cbrack | |||
| John Billingham Rating: N/A |
A goofy idea I haven't tried yet: Depending on the type of plastic that it is made of, it may be possible to heat-shrink or heat and stretch a commercial lent slightly. And If you look at the corelation between LCD pitches, and available lents, you'll see it's not TOO much of a "stretch" at that! (I'd start with hot water) Next "not too crazy" idea: since I actually made a "ruling engine" which was accurate enough to cut "micropolarizer" from self adhesive "half wave retarder" sheet for an LCD. It should be possible to cut "lenticules" with such a device, to whatever pitch you desire. If you look at a cross section of many lent materials, you will see how simple they are, often just a triangular "notch" every space. (yes, if you want toget fancy you COULD make a mini curved scraper, to cut actual radiuses, or , god forbid, parabolas!) Please, one of you "young and industrious types" try this, I'm too old and tired right now! Best, John Billingham | |||
| Hd Rating: N/A |
Hi cbrack, I really appreciate your ongoing barrier-tests regarding the distance adjustments, but I just guess a more far distance would lower the sweetspot..? Also changing the native resolution of a LCD doesn't change the fixed dotpitch, because the resolution is just interpolated(?) if the non-native size isn't divisibly by 2 (for example) so we're talking about 1 & 1/2 pixels (or anything), and to prevent ghosting etc. and prezise pixelmatching is really needed, I think.. the needed LPI is easy to calculate by dividing 25,4mm : dotpitch : 2 ..I Don't reallny know about the materials, but maybe the 'shrink-idea' from clyde would work anyway?? to shrink a 50 LPI to 49,803 seems to be not that much;) | |||
| Hd Rating: N/A |
AHm, I've meant john, but clyde also had this shrink idea.. funny isn't it;? | |||
| clyde Rating: N/A |
Na all credit to shrinking goes to JB Top of my head im just thinking tho, It would be hard to do "even" expansion or shrinking of a lenticular sheet wouldnt it.. If it turned out un-even the 3d effect would be out of phoase across the width of the lcd screen. Bet bet seems to be making a silk-screen barrier on clear acetate and using a transparent thin plexiglass spacer. The advantage this has is, your actually "making" your own lenticular sheet at an "exact" dot pitch of your monitor. JB has already done such a "kit". Sure brightness is reduced in half by a barrier, but lcd's are getting brighter by the day.. Clyde | |||
| Anonymous Rating: N/A |
What is the difference in sweetsport between barrier and lenticular sheet? | |||
| Anonymous Rating: N/A |
why this thread went dead again? | |||
| John Billingham Rating: N/A |
The "Sweetspot" is adjustable with BOTH methods, on a barrier by changing the thickness of the spacer between the barrier and the interlaced image, and by varying the thickness of the barrier "lines" On a lent, the "lenselike" curvature of the lenticules affects viewing angle, also the thickness of the material (or lent material plus spacer) has an effect. My "laptop barriers" try to stay at about a 30" viewing distance, and , I'd guess, a lateral "sweeetspot" of 1/2" or so. Best Wishes, John Billingham To keep a thread going, post a goofy message! | |||
| cbrack Rating: N/A |
hey everyone, been stuck at a ranch for thanksgiving - no land line. JB, really like the "ruling engine" idea. what type of monetary investment would be required to setup something similar? hd, here's the calcs that i have for a working barrier distance. starting w/ a 0.2555 dpi LCD, assume the following: dpi: 0.2555 eye separation: 60mm distance from screen: 355mm lpi: 50 barrier distance: 1.5mm using the calcs on http://www.ponies.me.uk/lentikit/wiki/index.php?title=Barrier%20Screens these figures seemed to check out. reading through the formulas, i noticed that barrier distance does not depend on lpi (barrier seperation is the term he uses on the site), but it does depend on eye distance from the LCD if you do change the barrier distance. this is interesting, since we can theoretically decide how far we want to be from the LCD and then make small changes to the barrier distance value to achieve a sweet spot. i'm going to go ahead and order a lent sheet for my 15" monitor and see what the results are. i would suggest setting up an excel spreadsheet w/ some of the equations and seeing if you come up w/ the same results... -cbrack | |||
| Andrew Rowbottom Rating: N/A |
I've been reading with interest, and have to but-in ... before you go ahead and order a lenticular sheet consider: Lenticular sheets are usually designed so they focus at the back of the sheet (where the paper is usually placed). If you pull them away from the paper (screen in your case) the focus is lost, barrier screens don't focus so don't have this problem. I've no idea how badly 1.5mm will affect the focus. On the other hand playing around with a real piece of plastic teaches you far more than scribbling numbers on a piece of paper. I think you stand a better chance of getting something working because your screen is almost exactly 100 dpi. Have fun and don't spend more than you can afford to throw away! Andrew Rowbottom | |||
| Anonymous Rating: N/A |
Hi,is there any physical difference (lense forms) betweens stereo-3d, multiview and other lenticular materials? does they have a different kind of lenses for diffent kind of lenticular effect? | |||
| A. Petersik Rating: N/A |
You should read through the papers at: http://www.research.philips.com/technologies/display/3d/paperspatents.html I guess that it is not important that the lenticular material must exactly fit the pixel size if you use the slanted idea. What we need is software which puts the pixel of the matching view at the correct position. These values can be calculated with help of the formulas in the philips papers. I hope, I can start some experiments next week. A. Petersik | |||
| John Billingham Rating: N/A |
Does anyone else have a copy of Philips' "Octopus for 3D LCD demo software??? It was available a few years ago, and compiled images to be "read" with angled lenticulars. I never really got the hang of it, but knew what it's intent was! Best, John Billingham | |||
| Anonymous Rating: N/A |
http://www.research.philips.com/technologies/display/3d/software.html | |||
| Ico Rating: N/A |
hi All, I found your thread today, here's my two cents : I did some experimenting today with a barrier printed on a transparent overhead sheet. My findings so far: - Ive picked up my math, and made some calculations that take the distance between the eyes, the distance from the eyes to the display, and the display's DPI as inputs, and results in the distance between the grid and the display, and the grid's interval. If anybody is interested, drop me a mail at 3ddisp@zevv.nl - The printer resolution is not a nice multiple of the display's resolution; so it's not possible to print a uniform raster which exactly matches the display. The raster has to be irregular, which results in brightness differences. - The distance between the display and the raster is critical; very small errors will result in huge moirre effects. Seems to be not so easy to attach a piece of sheet to a display at a distance accurate to tenths (hundreds?) of a mm. Anyway, the results are very very poor up to now; I am not able to attach the sheet to the display in a proper way | |||
| PeterPan Rating: N/A |
Hi, I read through the philips papers and there is something I don't understand.. maybe anyone can explain: They say '9-views' and their software also let you import '9' views, but does this mean 9 'stereo-views'? Otherwise I count 10 camera perspectives needed to have 9 stereo-views like L1+R1, (R1->=L2)+R2, (R2->=L3)+R3 ... anyone can explain this? | |||
| PeterPan Rating: N/A |
like this: ___L++R__L++R__L++R__L++R L++R__L++R__L++R__L++R__L++R 1__2__3__4__5__6__7__8__9__10 ? | |||
| clyde Rating: N/A |
hi PeterPan, Formulas thow me off.. but in essence.. 9 views afaik... means 7 "interpolated" views between the 2 outer conventional stereo views. meaning Left view ...L2(interpolated)..L3...L4..r5...r6...r7..r8..RightView This can be captured by 9phtsical or virtual cameras lined side-by-side , or using conventional stereo views comprised of Left/Right images and creating in-between views in software such as BASRelief etc... Cheers Clyde | |||
| PeterPan Rating: N/A |
hm, any formula in what distance and relation these 9 virtual cameras have to be placed? I guess it won't work with the wrong setup and view-distance... | |||
| clyde Rating: N/A |
well "ideally" they two outer cameras should be spaced at the same distance as our own eyes, with the in-between cameras ..err , in-between them. Of course this is going to be real cumbersome in real life.. so just like as in normal stereo photography.. you can "increase" main stereo-base to make room for in-between cameras (physical or in the case of 3dCG, virtual cameras) Cheers Clyde | |||
| PeterPan Rating: N/A |
Hi, it's just that there are so many different factors like distance, focal lenght and space between etc.. is there any formula that says, for example, the cameras have to be 6cm to eachother to match the human eye-distace? And what about viewangle etc.? Any casestudy or paper with the topic? | |||
| clyde Rating: N/A |
errmm.. a bit out of my league to be honest.. all i can say is ..logonto www.stereographics.com and download thier content creation kit. then if you know 3dsmax, you can actually 'see' the virtual cameras the plugin creates. Maybe you can study clues from there, incluing how factors like focal distance, DOF and targetting (focus) would work out. Cheers Clyde | |||
| PanPeter Rating: N/A |
errmmm, last week I found nothing there, but now it'S on the front page... funny isn't it?;) | |||
| clyde Rating: N/A |
I spoke to them,,, heeheh ;D (kidding) Clyde | |||
| PanPeter Rating: N/A |
no, you are one of them;P (kidding) | |||
| Anonymous Rating: N/A |
Guys, PLEASE, you gotta stop this hilarity!!! I'm ROTFL, and have split my pants! | |||
| Larry Elie Rating: N/A |
John suggested back in Nov that one might rule your own platic. I have acess to one of the best machine shops in the world, and I did some digging. That isn't the right apporach; you want to make a tool that can rule the plastic. If I have an accurate profile in 2D (both pitch and profile with depth) I can have an EDM sunk probe made that could rule plastic sheet in any width in one pass. This is time, not money, and it would mean I have to call out some favors, but once the rule cutter is made, one can cut all the sheet you want with the cutter. I might end up warming the cutter depending on the plastic. I only want to do this once. If someone is really serious, give me real dimensions and suggest a plastic. I am also thinking this would be a dandy opertunity to have one made for 8.5x11 overhead sheet to make a poor mans lenticular stock for prints. If you wait for me to get around to doing the calculations, it won't happen; I do science for a living and I have another very busy hobbby. If I had acess to a good monitor, I would just take accurate measurements using a molding material we reverse engineer with. Anyone want to do a few-square inch mold off a working monitor? Why am I now interested? I just burned out a monitor last night and I'm tired of waiting for Phillips to get the price down. | |||
| John Billingham Rating: N/A |
Hi, all, I haven't measured the current batch of LCD monitors and laptops. I'm sure there are at least a few "standard" pixel dimensions!Which is why I suggested the single cutter/ruling engine approach. BTW, I still VOTE for VERTICAL orientation of the screen, to avoid color distortions. Or are you thinking "slanted lenticular", Larry? I think the Philips software allowed one to "set" the angle of slant, so potentially, a close to correct pitch could work? Boy, there is NOTHING like goading someone else into doing all the hard work, is there? reminds me of the "Little Red Hen", huh? Now who will help me EAT the bread???? Best Wishes, John Billingham | |||
| John Billingham Rating: N/A |
HMMMMM! The two LCds I've been "looking at"(I'm broke now) both have a .297mm pixel. (poor people can only dream 15"!!!, but, then, thats about the max we'd get decent interlaced video running) So,If we do "two image per lent" "sweet spot" autostereo, we need a 42.76 lenticule per in lens sheet. Please don't forget Andrew Rowbottom's important reminder that the "focal point" of the lenticules will be a bit further back than "their" back (allow for the thickmess of the LCD screen protective layer(s)) Best, John Billingham (Henny Penny was another famous chicken) | |||
| hd Rating: N/A |
Hi John and friends, nice to see this discussion taking more shape.. finally;P I'm also at the point that multiview isn't really needed, even if multiview/slented lenticluar will bring 3d into the mass market.. So of course we should try to find a solution that deliviers a "single" stereoview with maximum sweetspot at one of the common dotpitches listed: (30") / 2560 x 1600 / 0,250mm = 50,80/0,258(?) (23") / 1920 x 1080 / 0,258mm = 49,224 LPI (20.1")/ 1600 x 1200 / 0,255mm = 49,803 LPI (19") / 1280 x 1024 / 0,294mm = 43,197 LPI (17") / 1280 x 1024 / 0,264mm = 48,106 LPI (15") / 1024 x 768 / 0,297mm = 42,760 LPI ..And it should work anyway with nvidias stereo 3d driver for 3d-gaming, which means in fact vertcial interlacing or something.. I don't know too if the vertical subpixel-orientation of current lcd's is really that big problem if the lense wouldn't have a specific focus on subpixel basis (like the slented technqiue) so if the lense has a minor distance, is should work anyway with a vertical lens orientation and vertical interlacing.. And now, I'm comming back to the point where a few bigger LPI (like 49,80 real LPI vs. 50LPI sheet) probably fixes the viewangle problem with the outer pixels... I don't really know, I will see it if I get an cheap 0,255mm dotpitch LCD on ebay and try it by my own with a 50 LPI sheets.. PS: I'm really broke too | |||
| Larry Elie Rating: N/A |
I still need a profile. I'm assuming it a parabolic cross-section, and if they did their math is is already multi-view. I am describing a extrusion (or better, pulltrusion) process, so I would shoot for one size and it would be basically the whole monitor width. If anyone has acess to a 15", 17" or bigger stereo set, I can get you mold stock. All I need is a few square inches... it comes right off after drys in a few minutes. Then I have it profiled as it's too fine for a CMM probe. If no one can find one, I will. This is a big company, and someone must have one by now; it's just a matter of time. You can find the pitches of the LCD in a table, or I can just count them off a close-up with a reticle. Yes, I'm serious. I can call out the profile to microns, so let's do this right. | |||
| John Billingham Rating: N/A |
I'd love to hear the details! Walk into an electronics store, gob some mold material onto the demo 3D lcd, "oops, EXCUSE ME!, Or my "cheap bastard" technique: Order one, take a casting, return it..... "This P.O.S is NO F---ing good! Gimme my money back!" (make sure to use a credit card that will stand up for you!) Best Devious Wishes, John Billingham | |||
| hd Rating: N/A |
Hi, I will probably buy a 20.1" LCD with 0,255mm dotpitch (49,803 LPI) on x-mass and try it with a 50 LPI sheet and nvdiia stereo driver.. Anyone here who can tell me if this could work anyway? | |||
| cbrack Rating: N/A |
guys, sorry for the delay - i'm waiting on a 50 lpi sheet to arrive for further testing. will post results when it arrives. larry, thanks for helping in the project - its amazing that you have access to those resources! help me understand exactly what you need so i can get it out to you ... since your first run will be geared to a .297mm pixel 15" lcd, we need to get you an interlaced test image for printing (8.5x11). after you get the image you will then print it out (100% scaling) and test your fab'ed 42.76 lpi sheet. is this correct? as far as type of plastic, i would suggest a heavy plastic in order to stabalize the sheet against the monitor. and a 1.75mm thickness. last time i did the calcs for .297dpi i came up with these figures: dpi: 0.297mm eye seperation: 60mm eye distance from screen: 355mm lenticular thickness: 1.75mm lpi: 42.97 important note: the lpi figure is different than the ones suggested above (42.76). this is because i'm using the calcs that take into consideration eye seperation found on: http://www.ponies.me.uk/lentikit/wiki/index.php?title=Barrier%20Screens if we simply want an exact match we can start with the 42.76 figure. -cbrack | |||
| Larry Elie Rating: N/A |
I can pull favors to make ONE extrusion die. I don't much are if it's 15", 17" or 21". I need to know the material for the film/coat (the 8.5x11 was for a parallel project for PRINTING 3D lenticular on the cheap... another project) and I need the full profile. Thanks for the link, but since I can only do this once, I would rather go the whole distance to mult-view, which is probably parabolic or multi-prismatic and not simple prismatic. Since I expect that is what the 'bigger' companies have done, I was looking for the Sharp or Phillips has profile. I will read the full link, and if I can't find a multi-view, that's what I will call out. BTW, I expect Phillips to hit the market within a year with something very real for consumers. I also expect someone, maybe Phillips to make a display with a flat phospher striped monitor as opposed to LCD and the appropriatly scaled lenticular face to do away with LCD problems but retain the simple 3D. That's harder, but probably going to happen in time. The final will probably be an OLED deposited on the back of the lenticular sheet itself, but that's probably 5 years away (for a large face). Larry | |||
| hd Rating: N/A |
Hi cbrack, nice to hear that I'm not the only one trying this with 50 LPI over christmas;) I was sure the exact LPI match can't work thorugh the eye-separation, so can you do the calcs also for 0,255 & 0,258 dotpitch? I'm a nobrainer in math.. | |||
| Larry Elie Rating: N/A |
Check out Phillips (lots of general info) http://www.research.philips.com/technologies/display/3d/technical.html to show how 'multi-view' could work. BIt isn't just a normal lenticular sheet anymore, and size does matter. TW, thick plastic will probably bias the positioning just as badly as the glass on a CRT, for the same problems. I also have to work on a polishing step. What you really want is a full lens over each pixel... oh well... Larry Elie | |||
| hd Rating: N/A |
Larry, the philips resource was also posted above, but I think multiview isn't really needed for our intended purpose, to get a most high quality stereo view.. The biggest problem with multiview is actually that the quality is low and there is not much video-content, a 3d-driver or hardware-power available that handles 7, 8, or 9-views in realtime.. That's the reason why we should put our focus on a 'single-view' stereo solution that works with the current 2-view based content.. by the way, what we want is 'a full lense over each TWO pixels!' ..that's the way how lenticular 3d-works;) | |||
| John Billingham Rating: N/A |
Remember, with "multi-view" you always have LESS "apparent depth" than with 2 image. You pay slightly with a narrow "sweet spot". I also believe, that "optimally" the lent pitch would be a "hair" smaller than the pixel pitch. (imagine a line from your eyes to a monitor 30" away, in order to pass through both the center of the outermost lenticule, AND the outermost "interleaved" columns(on the lcd), the center of the Lenticule would have to be a bit nearer the center of the screen, than the image under it. John Billingham (how can I stick a drawing here?, Peter obviously figured it out (of course, he's a lot brighter than I!)) | |||
| Larry Elie Rating: N/A |
You are right about each 2 pixels; I type faster than I think. I'm not so sure about the other; a 7 or 9 view should be able to do 2 view just fine; you just replicate, then when the hardware/software is available, you can still use the monitor. I have not seen the quality. The first lenticular I saw was B & W some 10+ years ago; the 'sweet-spot' was your head within perhaps 1" on 3 axes, and the images were pretty poor to boot. Very tiring. It sounds like most people here are thinking in terms of gaming, and I am not. When I say I can only call out this favor once, I mean it; it'll take me a year to get people to owe me enough to pull it off again.... nor am I that sure I won't opaque the plastic in the pull/roll/press step. I am leaning toward a roller because of this, but I haven't read enough about the creation of lenticular sheet at this point. If I fail, I can't iteratre the experiment for a year or so. Larry Elie | |||
| John Billingham Rating: N/A |
You really should read up on, and "interpret" this for yourself! I would NEVER trust anyone else, ESPECIALLY a supposed manufacturer of 3D monitors to provide the neccesary "specs" There are too many different "understandings" of lenticular screen usage! One interesting source is "N.A. Valyus, Stereoscopy" (an english translation of a Russian book from the 60s.) And , before you "call in" all those favors, you can print yourself a barrier with an inkjet and try things out, who knows, you may discover how good "barriers" can be. (the previous is a "not for profit" promotion of "Autostereoscopic barrier imaging") Best Wishes, John Billingham | |||
| Larry Elie Rating: N/A |
You are right about the cost to call out the favors! Do you always make one layer barriers? Have you thought about a stack of barriers? Registration would be an issue.... | |||
| John Billingham Rating: N/A |
Greetings, Lenticuloids! Dr. Sergio Baldissara has very kindly made a .zip file out of the "raster" chapter of "N.A. Valyus, Stereoscopy" and placed it in the files section (raster.zip) of one of his subversive groups. http://groups.yahoo.com/group/3dV2/files/raster.zip You will have to "join" to access, but all comers are accepted. (and immediately linked to the Office of Homeland Security database) The Soviets did some interesting work with lenticular and barrier autostereo! Grab it quick! before it leaves! Best Wishes, John Billingham | |||
| Anonymous Rating: N/A |
damn you mr. billingham, what we get from this crap? there are tons of historic papers like this on google!! | |||
| Mr. Billingham Rating: N/A |
Why,you're most welcome!, "anonymous" Yes, you are correct, there ARE "tons" of historic papers on "Google" Perhaps if you read some of them, you might discover, as I have, that the "ancients" understood a lot of this "stuff" way better than we can ever hope to? Best Wishes, "Mr.Billingham" | |||
| cbrack Rating: N/A |
update, i've ordered small quantities of many different types of lenticular sheets (40,50,60lpi) as well as 42.333lpi. when they arrive i can begin some better experiments... also, i want us to agree one the exact LPI size to give to larry. here are my thoughts... john, you are correct that the 'lent pitch would be a "hair" smaller than the pixel pitch'. based on my formulas, this value is determined by the distance of your eyes from the screen. i.e. if you are viewing the image 600mm from the monitor, then the lent pitch should be 0.2965 and if you are 300mm then the lent pitch should be 0.296. These values also assume a 1mm lent sheet thickness. therefore the 'eye distance' is the key to determining the LPI value for larry. Questions: Larry, what thickness do you want to use for the sheets? Everyone else, what should be the ideal viewing distance from the screen? 1 foot, 2 feet? Right now, I'm typing this 19inches from the monitor. So assuming 1mm thickness for the sheet and a 482.6mm viewing distance, the lent pitch should be 0.2963 and the LPI should be 42.850 | |||
| John Billingham Rating: N/A |
Well, Before my eyesight started to get "longer", I had settled on 30" as an "optimal" distance from which to view barrier images on the LCD screen of a laptop (pitch similar to what you are thinking about) This distance was chosen as one at which the "barrier screen" became less obvious. Get too close, and registration becomes fussier. Good Luck playing with the samples! John | |||
| MAYAman Rating: N/A |
Would you be able to make a kit for the Dell W2600 26" LCD? I've been trying to get some 3D gaming on but I can't find a headset which will really fit what I want. I'm running a dual Athlon MP 2800 with Nvidia 6800GT Thanks | |||
| MAYAman Rating: N/A |
I forgot to mention the W2600 is widescreen 1280x768 Thanks man | |||
| Anonymous Rating: N/A |
Sure, NO PROBLEMO! in time for Christmas, too! | |||
| cbrack Rating: N/A |
Hi everyone, Still waiting on three different shipments of lenticular sheets. Will let you know when they arrive! Larry, can you send me an email so we can finalize the exact dimensions of the lent sheet. I have all the figures and would love to get the project rolling. my email is: cbrack at myrealbox dot com thanks, merry christmas everyone! -cbrack | |||
| Larry Elie Rating: N/A |
cbrack, I sent you an e-mail from home (mine is listed) and I didn't get a reply. I am still waiting of figures. BTW, I seem to recall someone in the Detroit suburbs doing 3D off a big-old rear-projection TV by having one gun run off a second signal, and the lenticular grill providing some glasses-less 3D at one position in front. That must be 5 years ago, but I would like info on the position it worked from. | |||
| Anonymous Rating: N/A |
Hey!!!!! I sent you fifteen dollars, and I haven't received a thing!!!! Maybe you 3D guys ARE as sleazy as people say! | |||
| cbrack Rating: N/A |
larry, just sent you the info - keep me posted and good luck! -cbrack | |||
| Anonymous Rating: N/A |
at photokina-at dgs(deutsche gesellschaft für stereoskopie) was shown a new-perfectly looking lenticular screen-adapted to a laptop. pls ask these people of that man from bavaria-germany(munich) what is going on with his genious product. hasnt it discussed here? i ma no 3d-expert but then man showed me all kind of 3d-possibilities. one could move in from of the screen-the image staying stable! the goal is offering this technology to laptop-producers to let them produce it at a low-price-level. isnt 3dz of uk offering similar screens? what about those? how good are they. anyone seen them somewhere? btw: 3dz-software is a german product. | |||
| cbrack Rating: N/A |
anon, thanks for adding 3dz to our list of vendors distributing a lent sheet kit, i'm curious how they do 'sub pixel' filtering with the DVI dongle. if anyone has experience w/ this kit, please post - we still have the problem with subpixelation, and waiting to see if our home-made kits will work in landscape mode. -cbrack | |||
| cbrack Rating: N/A |
hd, hope you're still reading in the forum. I finally got an order of 50lpi from HongKong and I would like to send you a sheet for testing. Also, anyone that has access to a 0.255dpi LCD can help in this project. The 0.255 dpi LCDs are normally 20.1 inches like these: http://www.amazon.com/exec/obidos/tg/detail/-/B0001DHHH0/103-1229725-5906265?v=glance http://geek.pricegrabber.com/search_getprod.php/masterid=2313337 i have to wait a few weeks before I get a dell 2001FP, but if anyone has one of like these you can participate in the experiments. based on our previous discussions and calculations, the 0.255 dpi is the only dpi that will work with prefabricated lenticular sheets, particularly 50lpi. 15",17", and 19" LCDs can't be tested at the moment until we find a distributer who can do a custom lpi job. larry and I are working on a custom tooling solution and will let you know how that is coming along later. If you are interested and have a 20.1 lcd, please email me and we'll exchange shipping info. I won't charge anything for the lent. sheet but would appreciate helping me w/ shipping costs. my email is: cbrack at myrealbox dot com. glad to get this project rolling again! -cbrack | |||
| Innar Rating: N/A |
cbrack or anybody... How did your testing go? Which one was best lenticular sheet? | |||
| Anonymous Rating: N/A |
http://www.humaneyes.com/3d-software/3d-software-overview.php | |||
| Alan Probst Rating: N/A |
Hi. Wow. You people are amazing. After quitting teaching about eight months ago, I discovered that I had no other sellable skills and so to kill time while I've been broke and unemployed, I've been dabbling with 3D. I'm going to ignore all those numbers (because math is my nemesis) and 3d lcd screens (because I'm pretty sure I don't have one) you people are talking about,. . .and because they made my eyes hurt (the numbers, that is). I just wanted to say 'hi', I appreciate what you guys are doing (I've been playing around with phantograms, but you gotta wear silly glasses for those... this lenticular stuff is much cooler sounding), and ask a really really stupid question. ... If I were to take a stereoscopic pair printed on regular paper, slice the images into thin neat little strips, and attach them to something ... cheap ..., basically making ridges like a potato chip, could I make a 3d image (if its small like a baseball card and the person were to stare straight on)? Again, you people are great. I'm not as smart as you all, but I like 3d and I'm broke, so I seem to have 2 of the 3 general qualifications going for me... alan | |||
| peter Rating: N/A |
Yes, sure.. it's just called 'lenticular printing' like that kind of 3d-images you'll find on a DVD-cover, a movie poster or a walt disnay postcard | |||
| santins Rating: N/A |
It is impressive all this. Finally, did you manage to make the DIY lenticular screen work? I would like to know the details, if possible. | |||
| Peter Swanston (Peter64) New member Username: Peter64 Post Number: 5 Registered: 2-2007 Rating: N/A |
Hey well after reading this I decided to derive a spreadsheet to help with printing barrier screens. I'll upload it incase someone else wants to do some investigating. I derived all the formulas myself you can change the values at the top to alter the outputs below. I used photoshop to design my parallax barrier which I will be printing when i get access to a 1200 laser printer and some hq transparencies If i ever print it i'll let you know the results. I was planning to test vertical and horizontal monitor positions to eliminate color fringing but I'll see what i have time for.
| |||
| Peter Swanston (Peter64) New member Username: Peter64 Post Number: 6 Registered: 2-2007 Rating: N/A |
Hmm well it did work ... however the result was less than ideal and took nearly 3 hours of messing around . Luckly 600 dpi with bars 7 pixels wide perfectly matched my monitors dot pitch... I don't think this would happen for anyone except users with 15 inch monitors and 1024 x 768. Secondly I had to print onto the OHP transparency twice to get enough block to block out white light that was coming from supposidely blocked pixels. The result is still very very poor and does not work with black on white. Additionally one needs to position ones monitor on it's side ... horizontally will not work as color banding prevented this. Additionally I attempted ideal parallax barrier widths using alternation of 7 black pixels 7 white pixels dpi so that it perfectly matched the monitors dot pitch ultimately i had to use 12 black pixels and two white pixels. This means the monitor is reallly reallly dark. If anyone else has had any experience doing this and achieved better results please let me know. I will toy around with it a little longer and see if i can get anything more ideal. |
Spreadsheet for Parallax Barrier Calculations for LCD