Re: OMAP3 ISP and UYVY422

2011-09-08 Thread Gary Thomas

On 2011-09-08 07:40, Gary Thomas wrote:

On 2011-09-08 07:22, Enrico wrote:

On Wed, Sep 7, 2011 at 5:55 PM, Gary Thomas wrote:

My UYVY422 data looks like this (raw-3.0):
000: 0080 0080 007f 0080 007f 0080 007f 0080 
010: 0080 0080 0080 0080 0080 0080 0080 0080 
020: 0080 0080 0080 0080 007f 0080 0080 0080 
030: 0080 007f 0080 007f 0080 007f 0080 007f 

It should look more like this (raw-2.6.32):
000: 8034 8033 8034 8034 8034 8034 8034 8034 .4.3.4.4.4.4.4.4
010: 8034 8033 8034 8034 8034 8034 8034 8033 .4.3.4.4.4.4.4.3
020: 8034 8034 8034 8034 8034 8034 8033 8032 .4.4.4.4.4.4.3.2
030: 8034 8035 8033 8034 8033 8034 8033 8034 .4.5.3.4.3.4.3.4

n.b. these are grabbed from the same image on the camera, on the same
board - either running the new media controller code (3.0+) or old TI
PSP code (2.6.32)

I've compared the CCDC registers between the two systems and they look
pretty good to me (none of the differences explain the behaviour above)

It looks to me like the 8 bit data coming into the CCDC is not being
packed properly, as well as the second byte of each pair is being
dropped.

Any hints on where to look, what might be mis-configured, etc?


Apart from that (i have the same issue) do you get the full 720
horizontal pixels?

Because this is what i get:

http://imageshack.us/f/215/newkernel0.png/

It's not simply "stretched", the right part is missing. Did you change
some ccdc parameters in the files i sent you?


This is precisely what I see. If you look at the raw UYVY data as
I did, you'll see that 1/2 of the data is being lost. It looks like
there is some setup wrong in how the data is being moved from the CCDC
to memory but I don't know enough about the code to know where that
might be configured.



Note: I just saw a set of patches for exactly this CCDC with BT656
support on the Linux-OMAP mailing list.  I'll look at them to see
if they are anything different from what Laurent provided.

--

Gary Thomas |  Consulting for the
MLB Associates  |Embedded world

--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: OMAP3 ISP and UYVY422

2011-09-08 Thread Gary Thomas

On 2011-09-08 07:22, Enrico wrote:

On Wed, Sep 7, 2011 at 5:55 PM, Gary Thomas  wrote:

My UYVY422 data looks like this (raw-3.0):
  000: 0080 0080 007f 0080 007f 0080 007f 0080  
  010: 0080 0080 0080 0080 0080 0080 0080 0080  
  020: 0080 0080 0080 0080 007f 0080 0080 0080  
  030: 0080 007f 0080 007f 0080 007f 0080 007f  

It should look more like this (raw-2.6.32):
  000: 8034 8033 8034 8034 8034 8034 8034 8034  .4.3.4.4.4.4.4.4
  010: 8034 8033 8034 8034 8034 8034 8034 8033  .4.3.4.4.4.4.4.3
  020: 8034 8034 8034 8034 8034 8034 8033 8032  .4.4.4.4.4.4.3.2
  030: 8034 8035 8033 8034 8033 8034 8033 8034  .4.5.3.4.3.4.3.4

n.b. these are grabbed from the same image on the camera, on the same
board - either running the new media controller code (3.0+) or old TI
PSP code (2.6.32)

I've compared the CCDC registers between the two systems and they look
pretty good to me (none of the differences explain the behaviour above)

It looks to me like the 8 bit data coming into the CCDC is not being
packed properly, as well as the second byte of each pair is being
dropped.

Any hints on where to look, what might be mis-configured, etc?


Apart from that (i have the same issue) do you get the full 720
horizontal pixels?

Because this is what i get:

http://imageshack.us/f/215/newkernel0.png/

It's not simply "stretched", the right part is missing. Did you change
some ccdc parameters in the files i sent you?


This is precisely what I see.  If you look at the raw UYVY data as
I did, you'll see that 1/2 of the data is being lost.  It looks like
there is some setup wrong in how the data is being moved from the CCDC
to memory but I don't know enough about the code to know where that
might be configured.

--

Gary Thomas |  Consulting for the
MLB Associates  |Embedded world

--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: OMAP3 ISP and UYVY422

2011-09-08 Thread Enrico
On Wed, Sep 7, 2011 at 5:55 PM, Gary Thomas  wrote:
> My UYVY422 data looks like this (raw-3.0):
>  000: 0080 0080 007f 0080 007f 0080 007f 0080  
>  010: 0080 0080 0080 0080 0080 0080 0080 0080  
>  020: 0080 0080 0080 0080 007f 0080 0080 0080  
>  030: 0080 007f 0080 007f 0080 007f 0080 007f  
>
> It should look more like this (raw-2.6.32):
>  000: 8034 8033 8034 8034 8034 8034 8034 8034  .4.3.4.4.4.4.4.4
>  010: 8034 8033 8034 8034 8034 8034 8034 8033  .4.3.4.4.4.4.4.3
>  020: 8034 8034 8034 8034 8034 8034 8033 8032  .4.4.4.4.4.4.3.2
>  030: 8034 8035 8033 8034 8033 8034 8033 8034  .4.5.3.4.3.4.3.4
>
> n.b. these are grabbed from the same image on the camera, on the same
> board - either running the new media controller code (3.0+) or old TI
> PSP code (2.6.32)
>
> I've compared the CCDC registers between the two systems and they look
> pretty good to me (none of the differences explain the behaviour above)
>
> It looks to me like the 8 bit data coming into the CCDC is not being
> packed properly, as well as the second byte of each pair is being
> dropped.
>
> Any hints on where to look, what might be mis-configured, etc?

Apart from that (i have the same issue) do you get the full 720
horizontal pixels?

Because this is what i get:

http://imageshack.us/f/215/newkernel0.png/

It's not simply "stretched", the right part is missing. Did you change
some ccdc parameters in the files i sent you?

Enrico
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


OMAP3 ISP and UYVY422

2011-09-07 Thread Gary Thomas

I now have my camera, attached to TVP5150 running in BT656 mode,
grabbing data.  Sadly, the output is not yet correct.

My pipeline is set up as:
  media-ctl -r -l '"tvp5150 2-005c":0->"OMAP3 ISP CCDC":0[1], "OMAP3 ISP 
CCDC":1->"OMAP3 ISP CCDC output":0[1]'
  media-ctl -f '"tvp5150 2-005c":0[UYVY2X8 720x480], "OMAP3 ISP CCDC":0[UYVY2X8 720x480], 
"OMAP3 ISP CCDC":1[UYVY2X8 720x480]'

Running this command to grab the data:
  ffmpeg -r 14/1 -pix_fmt uyvy422 -s 720x480 -f video4linux2 -i /dev/video2 -f 
rawvideo raw-3.0

My UYVY422 data looks like this (raw-3.0):
  000: 0080 0080 007f 0080 007f 0080 007f 0080  
  010: 0080 0080 0080 0080 0080 0080 0080 0080  
  020: 0080 0080 0080 0080 007f 0080 0080 0080  
  030: 0080 007f 0080 007f 0080 007f 0080 007f  

It should look more like this (raw-2.6.32):
  000: 8034 8033 8034 8034 8034 8034 8034 8034  .4.3.4.4.4.4.4.4
  010: 8034 8033 8034 8034 8034 8034 8034 8033  .4.3.4.4.4.4.4.3
  020: 8034 8034 8034 8034 8034 8034 8033 8032  .4.4.4.4.4.4.3.2
  030: 8034 8035 8033 8034 8033 8034 8033 8034  .4.5.3.4.3.4.3.4

n.b. these are grabbed from the same image on the camera, on the same
board - either running the new media controller code (3.0+) or old TI
PSP code (2.6.32)

I've compared the CCDC registers between the two systems and they look
pretty good to me (none of the differences explain the behaviour above)

It looks to me like the 8 bit data coming into the CCDC is not being
packed properly, as well as the second byte of each pair is being
dropped.

Any hints on where to look, what might be mis-configured, etc?

Thanks

--

Gary Thomas |  Consulting for the
MLB Associates  |Embedded world

--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html