[Unichrome-devel] Fwd: Question about dwQMatrix.

[Kenneth <kea...@on...>]

Sorry, I meant to include you in the receipient of the original message, bu=
t=20
it must have slipped my mind.

Kenneth

=2D---------  Forwarded Message  ----------

Subject: Fwd: Question about dwQMatrix.
Date: Friday 18 June 2004 21:33
=46rom: Kenneth Aafl=F8y <kea...@on...>
To: Fiona Gatt VIA <fio...@vi...>, Joseph Chan=20
<Jos...@vi...>

Since I can't get any information from you to fix my driver, I'm considering
to sign that NDA that you sent me. But then I thought about why I'm even
considering writing this in the first place, and that is that the support y=
ou
give your customers is below par. What if also your datasheets is below par?
You can't even answer a single simple question about using a poorly
documented library api in a weeks notice, I see no reason to help your
ignorant company gain market shares in the linux domain!

Bye,
Kenneth

=2D---------  Forwarded Message  ----------

Subject: Fwd: Question about dwQMatrix.
Date: Friday 18 June 2004 14:04
=46rom: Kenneth Aafl=F8y <kea...@on...>
To: <fio...@vi...>

Hi,
I've yet to receive a reply on this question, maybe you can help?

Was hoping to clear up my and others complaints about Epia with MythTV befo=
re
this weekend.

Kenneth

=2D---------  Forwarded Message  ----------

Subject: Question about dwQMatrix.
Date: Thursday 17 June 2004 01:50
=46rom: Kenneth Aafl=F8y <kea...@on...>
To: <Jos...@vi...>

Hi,

I'm wondering in what order the dwQMatrix array should be loaded, I don't
 know if you know the answer, but if you do not, could you forward this
 message to someone who do?

Right now I have the segment below. But I'm having a hard time making sure
it's correct, because I do not know what order the CLE266 expects to receive
this in. The picture quality looks quite good, but has some artefacts on fa=
st
movements with static objects in the foreground.

What is strange about the segment below is that the high parts of the words
from the matrix goes in the first double-word, then low parts words from go
in the next double-word, is this really right? I can't only get lower quali=
ty
decoding if I read two words into each double-word, regardless of byte orde=
r.

And please don't refer me to the VeXP source, because that is all wrong, and
that is for sure, because if I follow that the picture has vertical lines in
the whole picture (repeat about 16 or 32 pixels).

Regards,
Kenneth

for (i=3D0, j=3D0; j < 16; i +=3D 4, j +=3D 2)
{
VIAMPGSurface.dwQMatrix[0][j+1] =3D
      (s->intra_matrix[ s->intra_scantable.permutated[i+0] ] >> 8) << 24 |
      (s->intra_matrix[ s->intra_scantable.permutated[i+1] ] >> 8) << 16 |
      (s->intra_matrix[ s->intra_scantable.permutated[i+2] ] >> 8) << 8 |
      (s->intra_matrix[ s->intra_scantable.permutated[i+3] ] >> 8) << 0;

VIAMPGSurface.dwQMatrix[0][j+0] =3D
     (s->intra_matrix[ s->intra_scantable.permutated[i+0] ] & 0xff) << 24 |
     (s->intra_matrix[ s->intra_scantable.permutated[i+1] ] & 0xff) << 16 |
     (s->intra_matrix[ s->intra_scantable.permutated[i+2] ] & 0xff) << 8 |
     (s->intra_matrix[ s->intra_scantable.permutated[i+3] ] & 0xff) << 0;
}

for (i =3D 0, j =3D 0; j < 16; i +=3D 4, j +=3D 2)
{
     VIAMPGSurface.dwQMatrix[1][j+1] =3D
      (s->inter_matrix[ s->intra_scantable.permutated[i+0] ] >> 8) << 24 |
      (s->inter_matrix[ s->intra_scantable.permutated[i+1] ] >> 8) << 16 |
      (s->inter_matrix[ s->intra_scantable.permutated[i+2] ] >> 8) << 8 |
      (s->inter_matrix[ s->intra_scantable.permutated[i+3] ] >> 8) << 0;

     VIAMPGSurface.dwQMatrix[1][j+0] =3D
      (s->inter_matrix[ s->intra_scantable.permutated[i+0] ] & 0xff) << 24 |
      (s->inter_matrix[ s->intra_scantable.permutated[i+1] ] & 0xff) << 16 |
      (s->inter_matrix[ s->intra_scantable.permutated[i+2] ] & 0xff) << 8 |
      (s->inter_matrix[ s->intra_scantable.permutated[i+3] ] & 0xff) << 0;
}

=2D------------------------------------------------------

=2D------------------------------------------------------

=2D------------------------------------------------------