Wireless monitor design

[HiME]

Hey all, I am currently a EE student and had recently formed a 4th year project team. We came up with the idea of developing a wireless monitor and would like some input on the usability and the technical implementation. =)

The idea is that the system will connect directly to the output of a digital signal of the video card, convert the signal into a wireless signal, and we'll have a receiver on the other end that converts the signal back to a video signal. The endpoint connection will be connected as a point to point connection. We are also seriously considering using powerline instead of wireless to transmit the data.

digital signal from video card -> microcontrollor device -> wireless/powerline medium -> receiver -> port on TV/ monitor

The advantage of this is that you no longer need any more of your bulky VGA/DVI/displayport cable!

Several problems we had brainstormed:
- Wireless channel, are we limited to the 2.4 and 5GHz bands?
- Will it be a better implementation to transmit the video signal over powerline?
- Prototyping: an FPGA broad like the Altera DE2/DE3 be sufficient?
- Bandwidth and latency constraints

Any input, comments on the general idea are welcomed 🙂 even if you have no technical knowledge, feel free to comment on the idea itself!!

Cheers,
HiME

 

[Ninjahedge]

Powerline might be better.

They have already done that with networking. The disadvantage being, if it was a general signal (not encrypted, just a new standard) then you have the problems of security and interference.

But think of it this way, being able to back wire through your machines PS, then hooked through the same power strip to your monitor. It might work...


Wireless? I would hazard a guess that there would be two problems with that.

1. Bandwidth. You may be able to get that 3rd screen (like a menu screen, or your toolbar for common apps) at a lower resolution (say 1024x768), but you may have problems getting enough to transmit a live signal through to a high res monitor, or pair of monitors.

2. Interference. There are SO MANY things that already work in that range, you might just jam them all. YMMV.

Hope the input helps!

GL!

 

[TecHNooB]

it may be very hard to debug a project like this if something goes wrong. do you have access to a high speed oscilloscope?

 

[PottedMeat]

you're probably going to be limited to the ISM bands - so yeah ~900M/2.4G/5.2

it's one thing to convert already high bandwidth parallel signals into a single signal suitable for wireless, it's another to TX/RX it. i wonder if you could just dump it into a serdes and dump the resulting signal into an off the shelf transmitter haha.

...hell if it fails say you were really designing a jammer 🙂

 

[Modelworks]

I would capture the video out off the pc then compress it using an off the shelf codec like h.264 and stream that over wifi to the receiver where it can be decompressed and displayed.

I just got a development kit from redpine signals for their wifi poduct and it would be ideal for something like this. The module is self contained , easy to interface and can handle 65Mbps if you only want to stream content one way. The module is already FCC approved so it removes all the certification issues you would have with any radio transmitter. Control of the module is done through simple serial port commands and the data transfers over a spi bus.

http://www.redpinesignals.com/Produ...-n/Wireless_Device_Server/RS9110-N-11-22.html

For the encoder it gets a bit more complicated but you can find code examples for FPGA encoders and decoders on the web.
Then all you need is a receiver module and a display module for the lcd panel.

 

[Ninjahedge]

Model, would 65 work on a 1600x1200 60hz signal with 100% refresh?

Compression works well when you have large contiguous and/or static fields of color, but in things like fast moving video or gaming, you might find it hard to get the necessary BW to get everything across (compression goes to nil in some cases or you get blur/blocks).

I am not contesting the suggestion, just curious on its viability....

 

[Modelworks]

Model, would 65 work on a 1600x1200 60hz signal with 100% refresh?

Compression works well when you have large contiguous and/or static fields of color, but in things like fast moving video or gaming, you might find it hard to get the necessary BW to get everything across (compression goes to nil in some cases or you get blur/blocks).

I am not contesting the suggestion, just curious on its viability....

65Mbit is enough to play back 1920x1080 @ 30fps content at full resolution using h.264 so no problem there.

The problem is encoding the video fast enough for something like real time. They make real time encoding DSP chips but they are not that cheap, tv uses them for sporting events, speeches, etc. I'm not sure what the current fpga coded versions are capable of doing. TI has the davinici processors for video that can do two streams both ways at the same time. Other options if you can afford the SDK and are willing to sign NDA is sigma designs. The SMP865x series can handle encoding and decoding of 1080P video streams in real time. There are solutions to the problem but it comes down to budget.

 

[Mark R]

It's already been invented, and is shipping in consumer products right now.
Here you go

Full, uncompressed 1080p - with 3D support coming soon.
1.5 Gbps using a 40 MHz channel in the 5 GHz band

Note that in order to get sufficient SNR, this makes use of numerous advanced techniques, including high performance pre-coding, and highly-parallel MIMO techniques. Most commercial offerings use 5 Tx and 5Rx antennas for spatial multiplexing.

(Sadly, the specs are protected by NDA, so there isn't much info easily available)

 

[DanDaManJC]

I'd also suggest looking into this as well:

http://wirelessgigabitalliance.org/

specifically their plans to standardize wireless displayport:
http://wirelessgigabitalliance.org/...rate-on-next-generation-wireless-displayport/

I think in general the project would be very cool... but the issue is really just money. You're going to need some beefy microcontrollers and fpgas to process some of these signals... dev boards won't be cheap. Or you could design a board -- but chips this complex will likely have more than just 2 layers -- so add up more cost for a complex board. Then how would you solder those chips on the board? Most likely the chips would be BGA -- not exactly the easiest package to hand solder.

I'd imagine your professor won't let you just buy 1 giant dev board either and turn this embedded hardware project into a 100% software project -- so you're going to be required to at least interface a couple different dev boards and/or transmitter/receiver packages.

How about doing the same project, but xmiting a smaller video signal? As a proof of concept. This way you can use cheaper hardware -- and more importantly -- hardware that you'll actually be able to build up by hand.

On another note, how long do you have to do this project? A full year next year? Or is it a quarter-long capstone? If just a quarter, I'd definitely stick to the proof of concept idea. That alone will be a hard enough project. If you have a full year next year, you could potentially put together a proper business proposal and fish for funds from industry. Would also give you guys time to figure out how to solder some of the hard-to-hand-solder chips (again, those BGA microcontrollers come to mind).

 

[HiME]

Sorry for the late reply guys. After several discussions, we had agreed on going for the idea of a powerline monitor. And the choice of standard for powerline communication would be HomePlug AV if we decide to pursue this route.

The reason for doing so is that wireless monitor standard had already been implemented as mentioned above. The market already have plenty amount of devices that can do the job wirelessly.

A similar thing can be said with powerline video transmission. At the end, all we have to do is to take the digital signal, modulate/ demodulate using the transmitter and receivers that are already on the market. Therefore, we are currently trying to come up with better ideas as to how we can extend the scope of this project. Most being motivated with the fact that our project proposal on the powerline monitor is simply dealing on the physical layer of the signal only, and it might be a little too simple.

Of course, the bandwidth issue is also a big concern. The transfer of uncompressed video is a huge burden on the transmission line and cannot be done even if we can utilize multiple frequencies during the transmission.

The project itself is one year long and we have a group of 4 people working on it.

Heck, my teammates were discussing about a hybrid solution an hour ago, auto load balance the data between the wireless 5 Ghz spectrum and the power line medium. What do you guys think about it? On the wireless side we'll have the standard upcoming Wireless Home Digital Interface, and on the power line side we'll use the Homeplug AV standard. The idea will add a larger buffer to the solution, along with the use of sequencing numbers to re-assemble the data that'd been sent over the two links.

Thanks for all the feedback 🙂

 

[DeathRayLoveMachine]

I just wanted to say, you are NOT limited to 900/2400/5000MHz. There are many unregulated bands in the IR spectrum, and they would probably be better suited to this project. (900/2.4/5 transmissions are limited to the equivalent of 8 watts isotropically radiated, unless you have a HAM license. Also there is MUCH more bandwidth available in the IR spectrum. Obviously you have LOS limits, though.)