Panda community,
Q from the meego list: http://marc.info/?t=129258766900004&r=1&w=2
Could someone help suggest potential solution for Panda from h/w
perspective?
--
Regards,
Nishanth Menon
>
> i'm not sure why anyone would want scart output, but it would be
> doable too with some work.
>
> as to the "conditional access slot", i'm not sure what this is
> referencing. possible a "SIM" card of some sort. i'd need more details
> to answer this one.
Looping in brendan as well.
Regards,
Nishanth Menon
So long as you have the LCD bus, yes. It'll be expensive, and Not A Real Set
Top(tm) , but you can do it - the components can all be bought off Digikey.
We did something similar for a demo with a Beagle C3 a year or two back.
Better would be for TI (or someone) to develop an STB reference design and
you can use that instead.
You probably want your IR done by an MSP430 or some other microcontroller;
saves spurious interrupts and if you're going this way, cost isn't an
issue. Or (*fx: blows own trumpet*) use <http://www.kynesim.co.uk/products.html>
or one of its many clones and avoid the problem completely.
You'll need SPDIF out of McASP rather than McBSP; don't know what's on
the expansion connector off-hand. Otherwise you're into a rather expensive
SPDIF out codec - Wolfson do some.
You can't have component unless you have the LCD bus exposed - only S/Video
(that said, component is a pain in most ways anyway).
Conditional access means many things to many people; in Europe, Dominique
probably wants CI (<http://en.wikipedia.org/wiki/Common_Interface>) which is
basically PCMCIA. In the US you might (?!) want Cablecard, which is essentially
identical but works differently. You would probably want to slap a Xilinx down to
do the voltage translation and use that for your control micro and IR whilst you
were at it.
Though, frankly, no-one is going to licence your CA/DRM implementation anyway
so an access card implementation is supremely useless to hobbyists (though if
you feel like working for STB companies for free, they probably won't turn
you down).
Being firmly in the Android camp, I happen to think the MeeGo guys are basically
chasing up an increasingly blind alley (albeit a blind alley whose upstreams I
will gladly nick :-)), but if they want a decent, standard, reference set-top, an
OMAP4-based one seems perfectly practical to me - and it has the advantage that TI
are happy to support the community; the only question is whether someone will design
and manufacture one - and, of course, given Dominique's position, Intel should
really be providing free reference hardware and software themselves!
If the business case turns out to be breakeven-able, I can ask around and see
if a suitable reference box can be designed/otherwise made available?
Richard.
What about this solution:
http://wiki.neurostechnology.com/index.php/OSD3.0
I believe this is based on the TMS320DM8168 video processor from TI.
Kind Regards
John
>
>
> Richard.
I don't see the OMAP4 as a STB chip at all, you need to add too many things to it
as this thread has proved.
Starting from something like the Netra aka DM8168 might be better, but then again
this one also lacks stuff like CA.
For a commercial STB product, lack of builtin (and thus "secure") CA is a big
showstopper.
It really isn't. However, it is not too expensive, it is well-understood and has good
software support and if you're worried about actually getting a product to market in a
reasonable timeframe at a reasonable development cost, it is (IMO) much better to pick
a well-understood device which is easy to engineer with and build that up than it is to
throw a bunch of IP together on a SoC, cross fingers, and spend your time trying to
debug the hardware.
By extension, if you're planning to get a bunch of random hackers together to support
a platform, starting with something that barely runs Linux is handicapping yourself
before you start.
- or in other words, for MeeGo (or indeed Android) Connected TV I'd suggest focusing
on the functionality. Let the commercial STB companies spend 5 man-years fighting STAPI.
>
> Starting from something like the Netra aka DM8168 might be better, but then again
> this one also lacks stuff like CA.
You mean Integra, and yes; they are better suited to the job in hand but
far more expensive than OMAP + addons, they have much less mature software support,
and I don't know of a cheaply available dev board for them.
If you're really looking at an STB-targetted device, ST7105 seems to be the thing
to use; I hear ST are selling dev boards for it. But you'll spend most of your
time trying to persuade it to start up rather than doing useful engineering.
Trident are rumoured to have a dual-core Cortex-A9 STB device, but it seems to exist
so far only in press briefs.
>
> For a commercial STB product, lack of builtin (and thus "secure") CA is a big
> showstopper.
Correct. And anyone building a commercial STB project will have to write some.
I'm just pointing out that since no-one in the hacker community is going to get
certification for their CA, (a) it will be hard to test, and (b) it will be
useless except to people who intend to take your code and make their money off it.
If you're happy with that, a free CI/CableCard stack would be very welcome. I just
don't see the incentive for anyone to write one - perhaps one of the people who have
one will see an adoption advantage in making theirs open-source, though.
(and I will personally buy a pint for anyone who can come up with a standards-level
MHEG stack)
Richard.
No he means netra, which is the family name. Integra (AM389x) is the spin without the interesting IVAHD and DSP bits. C6A816x has the DSP as extra and DM816x is the full package.
regards,
Koen