Wednesday, February 25, 2009

Move HVR-950q analog work

Tried to wrap up some of the outstanding issues on the analog support. Up until today it has pretty much worked for everybody, with a couple of minor cases of user error. However, today a user kindly pointed out that the driver is pretty much completely broken when trying to use mplayer, which is *BAD*.

So, I spent pretty much most of my night trying to nail down the differences in the V4L2 calls for tvtime versus mplayer.

Wow, I'm tired. Off to bed...

ATSC Patents

File this one under the mess known as "federally mandated standards dependent on stupid patents".

http://www.broadcastingcable.com/article/179758-Set_Maker_Asks_FCC_For_Help_In_Patent_Dispute.php

It basically boils down to Funai having bought some of the patents that are required to implement the ATSC PSIP protocol. The *entire* ATSC patent portfolio is licensed for around five dollars, and these guys are trying to strongarm Vizio in to paying five dollars just for their patent.

This pisses me off on three levels:
  1. The definition of RAND (Reasonable and Non-Discriminatory) suggests that the pricing should be reasonable. Charging five bucks for one patent when the entire rest of the portfolio goes for about the same price isn't exactly reasonable. Hell, the H.264 stuff is only going to add $0.25 to the price of the portfolio, and it does *WAY* more than PSIP
  2. RAND licensing prevents open source implementations from being written - which is particularly egregious when talking about federally mandated standards. The GPL isn't compatible with RAND since you cannot pass on the same rights that you received, so this basically means there is no "free" way to implement something the government has declared a standard.
  3. The patent is stupid. Anything for which I can write a relatively complete implementation of on a single Saturday afternoon probably shouldn't really be considered patent-worthy. And this isn't some hypothetical - I actually did write the PSIP parsing code for Kaffeine in one afternoon.

/Gets off soapbox

Sunday, February 22, 2009

Sunday

With the HVR-950q analog code now available for download, I reached out to the various mailing lists to solicit testers. I also did a search through the mailing list history and sent out emails to everybody who had specifically asked about support over the last eighteen months.

So far, the results are generally positive. Two people reported it was working, one is reporting a crash which I am investigating, and one pointed out that I haven't put in the hardware profile for the HVR-850 yet so he couldn't test.

Mike and Hans also did some code review and provided a series of comments over irc on #v4l, which I need to incorporate (nothing difficult though).

With this project getting wrapped up, I'm starting to think about what to work on next. I've been considering something from the following list (in no particular order):
  • Analog support for the Pinnacle 801e/801eSE (which shouldn't be *too* hard since I've got analog fresh in my mind)
  • Finish the analog support for the Pinnacle Ultimate (even though the digital side is looking less hopeful from an NDA perspective)
  • QAM scanning support for Kaffeine
  • Fix the ATSC video rendering problem for Kaffeine
  • Make ATSC scanning work for Me TV
  • Finish the ATSC closed captioning support for Xine/Kaffeine
  • Come up with a good solution for audio in TvTime
  • Finish debugging the URB problem with Paul's LinuxStamp
  • Wrap up support for a couple of the analog-only em28xx devices I have kicking around

Thursday, February 19, 2009

HVR-950q analog support

So, my blog has been pretty quiet for the past four weeks in terms of the work I've been doing. I've been working on analog support for the HVR-950q every night and weekend since January 23rd.

For those who might be interested in trying it out and providing feedback, check out the following repository:

http://linuxtv.org/hg/~dheitmueller/hvr950q-analog/

If all goes well, the plan is to solicit feedback over the next week or two, and then get it merged into the mainline.

There are still some performance issues to be worked out in the xc5000 tuning, and the VBI support isn't implemented yet, but the code should be good enough for most users.

Wednesday, February 18, 2009

Poor domain naming

I was looking for some sample code to figure out how to get AIX CPU statistics, and came across the following site for the Nagios monitoring tool:

http://www.nagiosexchange.org

Now, for those of you who might be thinking about getting a domain name, learn from their mistake:

The above website could be read as "Nagios Exchange" or it could be read as "Nagio Sex Change".

Monday, February 16, 2009

Still alive...

I haven't posted anything in a while. I've been really busy with a new project.

On Valentine's Day Victoria and I went to the Trapeze School of New York. Here are a couple of videos for entertainment value.