[mythtv-users] It's not the TiVo that she loved...

David Schmidt david.schmidt.in.dallas at gmail.com
Thu Apr 6 13:18:05 UTC 2006


On 4/5/06, Brian Wood <beww at beww.org> wrote:
> I think several factors work against the 350, or at least the mpeg-2
> decoder in it.
>
> I think it is agreed that except for that mpeg decoder the350s output
> is in every way inferior to a $40 5200 card, so I am referring to its
> decoder use and not its use as a standard framebuffer.

While the dropping of 350 support is a big loss for me personally (it
means I'll be on 0.18 until either my 350 and gets replaced or the
frontend it eventually serves goes HD), I think it is only a loss to
the Myth community as it helps to bring in users nervous about
practicality of committing to Myth before trying it out.

My current system is prototype only, though I'm using it day-to-day
for all my TV (and about 1/2 my wife's).  It's a 350 in my old P3/450
with 384M RAM, 1/2 of a 20G drive for everything but video/audio and
another whole 20G drive for /video.  It has a Voodoo 3 3000 unused by
Myth.  All except the 350 I owned previously.

At the time, no posts I could find indicated xv/xvmc worked
sufficiently with the Voodoo card to decode SD with the 450 MHz.  Most
also seemed to hint even with an nvidia or other "standard" Myth card,
the 450 would be borderline.  With the 350, I can record/playback with
less than 20% cpu--sometimes <10%.  For under $200US investment.  I'm
not sure I'd have tried it, considering the frustration a lot of users
go through and eventually give up, if I had to spend more initially.

The prototype is going great!!  In a month or two, barring
complications, I hope to build a proper combo back/frontend and remote
frontend that wouldn't need the 350.  I'll still use it, since I have
it, as long as it works.

I've been a C++/UNIX programmer professionally for over 18 years. 
Once my production system is up, I hope to contribute back lots (I've
got some ideas--especially for MythRecipe), but I'd probably never be
here if not for the 350 allowing me to prove to myself and my wife it
was do-able.  I've convinced a co-worker (also a UNIX/Linux hacker) to
try out Myth and he also initially went 350 because of the minimum
cost and cpu impact on his existing Linux box.

I'd hate to miss recruiting talent because of this, but understand why
those with working, more modern systems don't see the need.

>
> HD will be a bigger and bigger factor as time goes on, and the 350
> can't handle that.

People will likely have external converter boxes that output
downsampled SD to their older equipment for decades.

>
> More and more people will go mpeg-4, as CPUs get faster and storage
> requirements for HD get bigger, and as direct to mpeg4 capture
> devices become more common.

Many will.  Many won't.

>
> I hear that Myth will be going to an Open-GL based UI, which would be
> a problem for the 350.

But from what *I* hear, though the OpenGL painter is the default, the
older painter will still be command-line selectable.

Not trying to argue (honest).  Just providing another datapoint/view. 
If I were a developer with beefy enough hardware or all HD so I no
longer needed it, I wouldn't see a need to continue supporting it
either.  It's the nature of the open source beast.

Hopefully for beginners on the low end, I or someone else will get up
to speed enough to carry on support before the code actually get
ripped out/disabled.

Gratuitous thanks to all those who contributed to an excellent, mature
product.  I hope to someday have the honor of being among you!


More information about the mythtv-users mailing list