I just noticed that if you search for "Parallax Barrier" on YouTube, a total of 11 of my test videos come up on the first page. The main $5 video is the first one, followed by the last big test I did in 2009. Out of the top 10, 5 are from my DIY solution. If you do the same search on Google 2 of my videos are featured for the video widget. Pretty impressive if you ask me. Though I am not so sure how many people search for the words "parallax barrier". Do you guys have any ideas for keywords I can use for my next video? I probably want to do something involving the Nintendo 3DS, since that will surely be popular in the coming months. Any other tips on how to improve the quality of the video would be helpful as well.

I think an head-tracking SW could solve this issue.



Could a fotography of the screen with a ruler placed on it help to determine parameters to create a suitable parallax barrier for remote users?
The user could also place on its screen a small, generic parallax barrier implementing different parameters, to be used as a reference by who will actually create the final p. barrier.

Yeah, something like that could help, but you have to realize even a mm of inaccuracy can destroy the effect. So it is almost impossible to do this remotely (I would need the monitor in my hands to test). I still have some plans to make a product out of this, but I have to figure some things out first. I will probably just end up buying a bunch of cheap monitors to test with. So I can support all the popular models. Not sure when this will happen, but hopefully sooner rather than later.

This is a fantastic project, actually I found this site via Youtube, searching Parallax Barrier!

I'm amazed how well its worked out, has anyone tried this on an iPhone? if so please email me saj_12@hotmail.co.uk

You can buy a parallax barrier specifically designed for the iPhone here :
http://grilli3d.com/

Hm, is there no way to create a "Parallax Barrier Creator Tool", where you input your LCD and printer geometry (maybe with tests), then print out some test patterns with different sub-mm adjustments, etc., etc. until the perfect one is found?

Yes, that is possible. Some members actually created tools like that (posted on the past 2 pages maybe). I didn't verify them, but they might prove helpful to you.

Ah, found it. Thank you!

I just made an online calculator for calculating the document and barrier widths, based on the supplied variables.
http://tinyurl.com/3we38kq

cybereality, could you try to make some sort of guide around this? Basically you'd create a photoshop document at your printer's DPI, supply your screen height and this calculated width, and then fill it with an interlace pattern where the number of white columns would be the same as your horizontal resolution, and black columns the same PLUS ONE, so the first and last columns are both black. I have no idea of how you could do this, apart from drawing every line manually which is probably out of the question.

I've included the width of the barriers/gaps in pixels, don't know if that helps or not.

My original version actually used a Photoshop script that would create the barrier for you. Later I ended up doing it manually, just to have a little more control, but I could easily update the script. Just been really busy lately. I want to create a new tutorial that is much easier to do. I think the method I used is just too confusing for anyone to follow.

Nice tutorial... i will try to build this. I have a Notebook Acer 7720G with maximum resolution of 1920x1200 pixel and hope it will work. Too bad it has the downside of cutting the resolution ingame into half but it seems all 3d-solutions has a problem of some kind at the moment.

I have a laserprinter with 1200x600 dpi maximum but it can only print maximum A4 but I will try it first with it.
At ebay i found transparent film coming from a business wind-up and so I bought 200 sheets for 10,20Euro. Cheap I think.

I was wondering what transparent tape is clear enough to see through clearly. I tested one at a window and... no good result. So I asked and the tape to look for is Tesa or Scotch "Crystal clear"... this tape is not the normal transparent tape so I was told... so I bought a pack with the scotch version and will see how it goes with it...

At last I bought a steel rule. Normally used by craftsmans but I bought it to use it together with a cutter. I hope to get a clear cut with this setting because I dont have a A3-Printer.

Lets see what I can get with this...

I was wondering what kind of printer makes the sharpest edge. I thought it should be a laser printer but you wrote the laser wasnt very good... So that means dpi is everything? I thought the ink wouldnt make such a good edge because the ink wouldnt remain strictly at his place... what about plotter? Or other techniques if any?

If the result isnt very good I will look after canon pixma... they have printers that print up to 9600x2400dpi in black and white. And the ones that can handle A4 but no A3 are sold for 40Euros if they are used. So this would be my next move I think when the result isnt pleasing.

Is the image darker with a parallax barrier because of the black? If so... wouldnt it be possible to use a light color next to white? White itself wouldnt be possible I think. But the color would block the pixle the same way I think. And its often the case that printers can print higher resolutions in color than in black and white.

What Im wondering is how to find out the pixel pitch? I cant find it for my notebook...

How about the difference between the parallax barrier and the pixel? Its not taken into account in all formulas except the javasoftware. But it looks to me like an important point isnt it? I mean 1 mm more or less could mean that at the edge of the screen the pixel are seen completely different. So when not taking this into account and directly putting the film with the ink-side to the screen will it work perfectly or does it mean there is enough space between so that at the edge you will see different pixel even though you calculated the rest of the values correctly?

A laser printer at 1200dpi should work. You would think laser would be better than inkjet (maybe it is) but with my particular Samsung laser printer, the printouts were inconsistent. Meaning I would print the same pattern twice, and it would come out slightly different. It was very difficult to work with (although it was enough to get a proof-of-concept). Printing in color might work, I never tried it. That could be interesting. With a light gray it may make the image appear brighter. The parallax barrier line should be about the same size as 1 pixel (or the dot pitch). In practice, it may have to do slightly smaller due to the perspective caused by the glass on the screen. Turning the sheet to face the monitor can also help (and was what I did). But the pattern is always uniform. Once you get the right number, it will work for the whole screen (or a part of it, if you are printing multiple sheets).

Inconsistent? Maybe it was because the rollers that drive the paper were not working together correctly so the paper was driven more at one side? I hope i wont have such a problem...

I dont know if a light gray would be a good idea... because wouldnt that only mean that some dots arent printed to get some more white into it? If so I think it must be a real color so that blue, red and/or yellow is needed to build the color. And probably it cant be too light because that maybe would result in not printed dots too? But then the maximum that could be done would be to use yellow, red and blue together. Not at one point because that would mean to get black but near together. But I dont know what color that would give. Maybe even black too?

I agree with you SebastianJu, it was a nice tutorial. I do print my stuff with a LaserJet though. I think that it prints great quality images. Just make sure that you have it set to 1200dpi. I'm pretty sure that this is what magazine printing companies use when printing too.

white won't work, since you can't actually print white. any color other than black will give a poor result.

The barrier width is crucial. get it wrong and it simply won't work except for a narrow column somewhere on the screen. The air gap between the screen and the barrier is equally important, but that's not a factor when printing the image, so if you have the right barrier you can just trial and error the correct distance.

Hi all.. it's very nice this work and i tried but i have a problem to get the correct pixel pitch of zalman zm-m240w.... the 22'' is 0.282 but this is the 24'' version and in many site the value start from 0.270 to 0.285... i tried 272, 276, 282 without fortune... anyone have tryed? It's possible to calcolate the pixel pitch? (this also in case of make a test for my notebook asus where the pixel pitch of the monitor isn't simple to find)... thanks thousand

Yes you can calculate the pixel pitch by simply measuring the width of your display. The formula is then :
dot pitch in mm = width in mm / horizontal resolution.

For the 22" (16:10) the width in mm is 22 * 2.54 * 10 * 16 / sqrt(356) = 473.86 mm, this gives a pixel pitch of 473.86 / 1680 = 0.2820 mm.

For the 24" (16:9) the width is 24 * 2.54 * 10 * 16 / sqrt(337) = 531.31 mm, which gives a pixel pitch of 531.31 / 1920 = 0.2767 mm. If they didn't lie about the real diagonal size (which happens sometimes with some manufacturers), then you could probably test 0.277 which is closer to the real value. If this doesn't work, then the problem with your barrier should probably lie elsewhere.

If you do measure the actual width you will find out if 531.31 mm is the correct value, an error of 1 mm in the measure won't have much impact on the final result (0.005 mm in this case).

EDIT : corrected, calculations went badly the first time :/.

in this case, 1mm is NOT an acceptable error. 1mm is the width of almost four pixels. There is literally no margin for error with this

The 1 mm error is divided by 1920 in the end, so as I said it gives an error of max 0.005 mm for the dot pitch calculation.

The dot pitch is only needed to get in the ballpark. You still end up needing to do trial and error at some point, because there are a lot of variables at play. So even without the dot pitch, you should be able to make a decent estimate just by measuring the screen and using some simple math.

I hope someone can help me to make a barrier film for my 17inch monitor (sync master 732) to use as autostereoscopic display.I wish to get how can i make the transperent film for this and how can I prepare image for the same.

Did you see my tutorial:
viewtopic.php?p=55961#p55961

I include a test image you could use, but it also works with the iz3D driver.

Tnx for the answer ... 276 ... i tryed but probably i have to set 0.2767 then.. ok...
I'm waiting the new trasparent film for try... (When i have time ........... uff)

no really, if your measurements are 1mm off, you won't get a usable barrier. that one millimeter is almost four pixels, so if you position the barrier correctly on one edge of the screen, the other side WILL be off by four pixels.

sure, the 1 millimeter error is divided by 1920(so it's actually 0,0005mm), but since you're printing at least 1921 lines, that half micrometer is still a millimeter in the end.

so, if you want to calculate your dot pitch for any reason other than this, half a micrometer is probably considered acceptable but for this, you want it as accurate as possible.

Thanks for the correction, as you said it's actually a 0.0005 mm error margin for the dot pitch (1 / 1920 = 0.00052) and not 0.005 mm as I said. Stupid me...

It's in fact even better than what I tought because a 2400 dpi printer has a dot pitch of 25.4 mm / 2400 = 0.01058 mm, which is a lot less precise than this error margin. And the given dot pitch of monitors is generally rounded to max 3 decimals (which is the case for the Zalman ones), which gives a 0.001 mm error margin. Still a lot worse than what you would obtain by measuring the width of the display yourself with a 1 mm error margin.

So in only two tries - by selecting your lower and higher bound for the measure of the width of the display - you should theoretically be able to print the most perfect possible parallax barrier printable with any printer.

But as cybereality said the dot pitch will only get you started with a first approximation, because in the end the precision of your printer won't be good enough whatever its dpi is. So you'll probably have to use uneven values for the tickness of both transparent and black bars, it's probably the reason why using 600 dpi printers can give acceptable results when printing at a 2400 dpi resolution.

And you should also account for the fact that the thickness of your bars shouldn't be exactly equal to the dot pitch. For the effect to work the parallax barrier is not placed exactly on the surface of the pixels, but some millimeters away from them. You don't have the choice anyway because of the thickness of the different strates between the pixels and the surface of the display.

As illustrated in this schema, you can see that a parallax barrier with bars of the same thickness than the pixel pitch won't work, even if this illustration is not at the correct scale :



Unfortunately I don't think there is any mean to know or measure the thickness of the surface of a display. And even when knowing this thickness, the fact that it's made of glass and other materials has an incidence in the way the light travels it. So you'd also need to know it's refractive index, and then use the Snell law to make the calculations for your parallax barrier.

In the end it will be much more simple to do some trial and error, starting with an acceptable approximation and going down till it's correct. Be prepared to consume some transparent paper...

What I found to be helpful (and also save paper and ink) was to print 5 or 6 patterns on one sheet, each about 2 inches high and 8 inches wide. This way I could test a few different options without wasting a whole sheet (and more importantly all that ink). You don't have to print all 5 patterns at once. You can just print the top one, and then the others in order as you adjust via trial and error.

You're missing something. The exact width of each individual barrier isn't that important. It's just the average distance between them that has to be very accurate.

In effect, you'd have to divide the printer error margin by 1920 to make your comparison valid.

I do agree though, dot pitch is a bad way to do this because it needs to be so very accurate. I know you can find dimensions of the actual active display area for some displays, down to 0.1mm. If at all possible, you should use that rather than dot pitch.

That's what I said in my previous message when I talked about the resolution of a 2400 dpi printer that was too limited for the task, hence the need of uneven thickness for the parallax barrier.

There is no reason to compare the error margin of the dot pitch with the error margin of the printer dpi. If the dot pitch error margin is inferior to the printer precision then the weak link is the printer, and the dot pitch estimation has enough precision.

As I've explained the dot pitch estimation doesn't have to be that precise since it's not this exact value that will be used for the parallax barrier, it's only a starting point. And as I also said, this starting point will be more precise than the official dot pitch you'll get from manufacturers if you can measure the screen width up to 1mm error. I didn't know manufacturers gave display widths at a 0.1mm precision, if you can find that all the better.

I was happy, because I thought I had that exact monitor, then I was sad because I realized I had the 215 one, then I was meh because there wasn't something purchaseable premade yet anyway, and then I went to check my printers DPIs and I have two 1200x1200s and one 2400x1200 and was happy because I thought I might be able to do this, and then I was sad again because none of them could handle a sheet big enough to do it in one solid piece and I don't have ink anyway. x3

Still, keeping an eye on this. o3o

I just stumbled upon this thread and I am looking to get some help. First off my printer is a Canon Pixma MP210, and it says that it will do 1200 dpi. My monitor is a 22" with a 0.248 pixel pitch that brought me to a ppb of ~11.71. I have printed about 15 interpolated tests and none come closer to real results than a 12 opaque pixels and 11 transparent pixels template that I made, which still seems to do nothing when I look at the test image provided(opening with gimp at setting the view to 100%). Using that barrier I can move it slightly and see the image swap happening, so I know its on the right track. Nothing else seems to even almost show any 3d effect. Any help would be greatly appreciated. My monitor is a LG W2240T, and thanks in advance.

If you use the color charts as a reference, you should be able to tell if the pattern is too big or too small. That will at least help you see which way to go.



That printer says its 1200dpi, which should be enough. Its probably just a matter of finding the correct pattern. I had a lot of trouble with this myself, and mostly it was a trial and error procedure (over like 2 years). Make sure to also try changing your distance from the monitor (move forward or back) and try flipping the sheet around (so the printed side faces the monitor or vice versa). I am going to try to see if I can find an easier method. Some people here have posted some formulas and programs that might help you find a better pattern (just check the last 4 or 5 pages of this thread). You might want to try some of the stuff they suggest.

i just had an opportunity to try this out on my netbook, and i got a working pattern in just 3 prints(thanks cyber, those color charts really helped )! I really need some kind of spacer behind it though, because i have to get it 2-3 mm up off the LCD surface, and it's just a pain trying to get it flat by taping it to the bezel. I guess cyber's monitor is better for this, since he can just put it straight onto the display surface, and the internal distance to the actual LCD panel just magically makes it work.

I'm getting terrible ghosting right now btw, but that's mostly because my ink cartridge is running on fumes. The first few prints were better, but back then it just said "Ink Low". Now it says "Cartridge Empty" and has to be sweet-talked into printing a page. Still works though

Awesome man! Maybe I have been lucky with the Zalman. I have not tried it with any other monitor yet. You can try inserting a blank transparency sheet in between the barrier and the monitor, but you will lose some quality.

Well, considering I'd need at least 10 sheets, a solid spacer is more realistic if i want to see just about anything through it. right now the bezel gives me the perfect viewing distance, so I'll probably try a 3mm thick piece of acrylic or something.

I was looking into building a more permanent fixture for the sheet, and acrylic seemed like the way to go. Its pretty affordable.

Where i'm working now, we have printers that take 36" rolls.. Perfect for a 72" 16/9 screen!

Nice!

Hi , is this possible to buy a "Parallax Barriers"

At the moment, no. But I am looking into options for turning this into a product.

Thanks for this exelent information