Making A Videowall

Ur@eus writes "Zeeshan Ali Khattak has made a videowall using Red Hat Linux, GStreamer and commodity hardware. The solution was made based on the need to create a flexible and cheap solution for use in Pakistani Schools and Universities using commodity hardware. To find out how this was done and some more details, and of course some cool pictures, check out the Video Whale project homepage."

Comments

Cropping needs improving

[fruey]

Great idea, looks good... but in most video walls you have to allow for the space between each monitor when you crop, or reduce the physical space between each. Here, there's way too much space between each screen (both vertically and horizontally) and the images look strange because the cropping doesn't allow for it.

Re:Cropping needs improving

[helmutjd]

I think in this case they would've been better off with a different brand/type of monitor altogether... those ones have far too much plastic trim around them. Cropping that much out of the movie would really kill the image, and the only other alternative would be to remove all the monitors from their mouldings to get 'em closer together (not fun). Sweet project, though.

Wouldn't four quadheads be more usefull

[baschie]

It seems that four computers with four pci videocards would cost about the same as one computer with 4 Matrox G200 MMS Quadhead videocards ($699). You would not need all kinds of software distributing the videosource over four computers and it would make administrating and moving the whole system much easier.

Re:Wouldn't four quadheads be more usefull

[Quixote]

But what about the bandwidth issues? Will the PCI bus handle 16 video streams flowing through it (e.g. a movie, as the wall in the article shows) ?

16 monitors vs projector

[Anonymous Coward]

Hmmm. 16 x 17" monitors @ $200.00 each is $3200.00... Hell, even if they were only $100.00 each you can get a very nice projector for less which will blow away the functionality of this system.

The whole idea behind a video wall is that you can display the same, different, or transitional information across the monitors. They (the monitors) can be ganged together for a single display, split into sub-displays, or data can be moving across them. The system that they describe has very limited use and will not be able to do what a videowall is meant to do.

Videowalls are quickly being killed these days by projectors except in the instance where you have limited installation depth and do not have room for the minimum throw of the projector. Or I guess where the ambient light is too high but even then you would want non-glare screens on your video wall for the same reason and the lamps in projectors are getting quite bright these days.

Even if you take into account the annual cost of lamps for the projectors you would have to balance this out against the maintentance cost in parts and man hours of the system he has built. bet it works out pretty close.. Just getting all 16 monitors to calibrate equally is going to be a nightmare.

Re:16 monitors vs projector

[FyRE666]

Ok, what if they used a cluster of 16 projectors? ;-)

simulated

[SubtleNuance]

For god sakes man

take those monitors out

of their cases. Put

Them closer together and

kill the funky spacing.

Listen to you!

[Matey-O]

Jeebus Cripes! These guys whip up a Cool Hack(tm) with scrounged materials, make it work and add to the collective abilities of Gstreamer and all you have to say is 'it looks crappy'.

What a bunch of hypocrites. The fact that the monitors can be swapped out after a proof of concept, and that you've got the power of four CPU's available (Beo-mumble) is completely lost on you guys.

And I've figured there'd be at least ONE MPAA crack from somebody.

Cost-effective ?

[tmark]

the need to create a flexible and cheap solution for use in Pakistani Schools and Universities using commodity hardware

I just can't believe this is cost-effective for more than a 4-screen display. With quality video projectors costing less than $2-3000 USD, this solution doesn't save much money, and is far less convenient in terms of portability - how would you even move around an 8x8 grid of monitors -, which would seem to be key for application in schools and universities. Also, the whole array is visually distracting due to the breaks between the monitors.

Sometimes people get distracted by technology and forget about the constraints of the problem to be solved.

Re:Why not use old LCDs?

[kubla2000]

You could salvage LCDs from old laptops.

Not so easily done: http://www.eio.com/lcdconnect.htm as far as I can tell, you'd have to get into some pretty heavy-duty electronics and buy convertor cards to handle the different input expected by a laptop display from that which is delivered by a vga card.

@#$@ the warl

[Nobley]

Wire this thing poorly and you might just get yourself a Firewall.

eliminating the spaces...

[rusty0101]

I have seen several notes deriding the spaces between the displays. The "fixes" suggested for this include using LCDs (which I suspect are outside of the budget), or disassembling the monitors and bringing the CRTs closer together. (anyone want to discuss the safty issues of pulling one of the center displays out to replace it?)

I suspect it would be far cheaper, to use fresnel lenses in front of the CRT's with modifications to the rack they built to center the CRT on the fresnel, and mask off the power light for the monitor.

Will it be perfect? No, but I think it will be more flexiable.

-Rusty

Re:movie.* = comedy

[SCHecklerX]

Any movie watched on that wall will make it a funny comedy. I mean, just watch the head on this dude!

So, basically it turns all actors into canadians.

"What's wrong with their heads?? It's ok, their canadian." (or something like that :)

A creative solution to the problem of...

[Spurious Growth]

No Funding! I think the point that people are missing here is that the project does not have a budget. There is no money to spend on a video projector. The entire system relies on hardware which is, on a regular basis, serving an entirely different purpose. This solution allows them to create a large display when it is necessary, out of components at hand. Almost any computer lab can generate a 4' x 5' display on demand.

Also, what is the effective resolution of such a screen? It sounds remarkably similar to the IBM ultra high resolution LCD we heard about a while back.

Spurious

Once again, did anyone read

[JWhitlock]

...or just look at the pictures. As the author states:

I rather quick realized that the only extra component we needed in Gstreamer was one that that did the cropping for us. At the same time I also saw that my knowledge about multimedia and GStreamer was not good enough to allow me to write this element. So I tried asking my friend Wim Taymans if he would be willing to take on the task of writing such a plugin. He was kind enough to do that not only because of my need, but also because he saw it as another nice feature of Gstreamer that would be needed by many others.

So, they know that cropping is a problem. But that's just software (and maybe a bit of hardware, with those huge borders). But that isn't their biggest hardware problem.

It turned out however that our biggest problem was finding PCI video cards in Pakistan whose XFree86 drivers could do XVideo, this in a situation when its hard to even find PCI video cards at all in the market. Solving this problem of lacking parts took us 3 months and at one point we even considered abondoning the project. We still need more cards because we do not have more than these 16 s3virge/DX cards. If any of the cards stop functioning, we are out of business.

This isn't the U.S. - it's Pakistan. They can't go to eBay and find a 4-head card, or even pick and chose cards (or even monitors) based on requirements. They have to use what they can find, and I for one am impressed. How many of you would give up if it took three months to just find the hardware?

This is a great engineering story, of folks working with what they have, and a great Free Software story - they could have tried some pirated copy of commerical software, but instead they decided to use open source components, stretching what is possible. Could it have been done with a projector? Sure, if one was availible. But now the state of multi-monitor free software has been advanced a little, which may benefit you or me some day.

I hope that there were some other people who saw how cool this was, who are contacting the authors with useful suggestions about removing the shells and mounting the tubes closer together, that are looking at the GStreamer source and thinking about how to add cropping, and how to make cropping easy, and hopefully a few people that are thinking about donating equipment, and realizing how lucky they are to live in a world where you can order a projector from Amazon and have it delivered in days.

More importantly, what will they call it?

[commodoresloat]

What are they going to call a wall from the video whale project -- the Whaling Wall?