[mythtv-users] Code to convert Xfree86 modeline into an 18 byte DTD array? (Jean-Yves Avenard)

[Stephen P. Villano]

On 2/8/14, 2:12 AM, Stephen Hocking wrote:
> Hi all,
>
> After some hacking I finally managed to get the desired Detailed
> Timing Descriptors into the EDID file I was editing. Using this EDID
> file with the nvidia driver gave HDMI audio (after some fiddling), but
> the overscan remained. Firing up xvidtune, it became clear that the
> desired modeline was indeed being fed to the receiver, but the
> receiver was then munging the actual signal to force overscan. Very
> frustrating. It means that I can have HDMI sound over the same cable
> that video is delivered through, but only with overscan. A suggested
> method setting static_hdmi_pcm=1 for the snd_hda_codec_hdmi did not
> work. The sound devices were as silent as the grave.
>
> As part of finding all this out, I found a utility that parsed edid
> files (you can find it in most Linux distros), but it had the
> limitation that it would only parse the first block, which was all
> that existed back when it was originally written. The extension blocks
> are where the fun part resides, containing the following:
>
>   1) More info about the monitor (color depths etc)
>   2) More standard video modes that were not around when version 1
> came out,
>   3) Data blocks that describe the various sound bitstreams that the
> receiver can handle (LPCM DTS, AC3, DD+ DTS-HD, etc), along with the
> bit depths, sampling rates and maximum bitrate where application, plus
> the maximum channels for each.
>   4) Speaker configurations that are acceptable
>   5) Vendor specific data. This was the interesting bit.
>
> I wrote a program that would selectively copy various portions of the
> extension block over to the new EDID file. The only one whose presence
> was necessary for the Nvidia driver to instantiate the HDMI sound, was
> the vendor specific data. All the others did not matter.
>  This was also the block that forced the overscan. I'm trying to pick
> this one apart to see what else it's responsible for, and what will
> happen -  the receiver vendor data block has more bytes than the
> recommended length, which the spec says should be zero.
>
> If anyone's interested, I could probably post the code somewhere. It's
> all just command line stuff.
>
>
> _______________________________________________
>
While at current, I have no use for your tool, I cannot say that would
be true in the future.
As I do believe that knowledge is power and I insist on being nuclear
armed in terms of knowledge, please do post the code or create a
repository for the code.

Personally, along such journeys of learning the capabilities,
limitations and coding of devices, I tend to relax by the consideration
of percussion programming.
The use of a 10 pound sledgehammer to reprogram a device.
8 pounds tends to bounce off of things that are *really* hard, 23 pounds
is too damnably heavy to consider (I have a range of such hammers
around, courtesy of my father, who worked in construction).
But, the thought relieves stress and "zeroes the bits" for a bit.

*True* stress release is destruction of equipment to prevent data
recovery by ill intending persons.
Thermite is cook on hard drives.   ;)
The 8 pound reprogramming tool is excellent for the rest.
Or just revenge for decades of dealing with recalcitrant hardware.  :P

Good night all, I have to crash. My shoulder is giving me hell, courtesy
of a trial separation, it's not going well. Blew out one knee, courtesy
of my not considering my knees when getting up earlier. Various and
sundry areas are giving me aggravation, due to rather ancient injuries
and the weather is changing and I'm a *really* good biological barometer.
Tomorrow morning will either suck or be cool. Either way, it's cool, as
I get to wake up again.
Because, I've known plenty of good men who died, none came back to tell
me how cool dead was. I've only had living people suggest that and to be
honest, I really don't think that they had my best interests at heart.
So, good night, may the bedbugs not bite, or if they do, may they die of
food poisoning, like any that might bite me would.  ;)

--No, someone did not leave a Hershey bar in the bottom of my gene pool.
Someone both urinated *and* defecated into my gene pool and here I am
before you. So, now what?

~~ Me.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mythtv.org/pipermail/mythtv-users/attachments/20140208/0ec7e6cd/attachment.html>