I modified the DM6446 EVM video inputs to accept component video inputs. We've modified these files in order to configure the TVP5146 for component video:
/lsp/ti-davinci/include/media/davinci/tvp5146.h
/lsp/ti-davinci/drivers/media/video/davinci/tvp5146.c
/lsp/ti-davinci/drivers/media/video/davinci/davinci_vpfe.c
When I try to capture video using the demo app ">/opt/dvsdk/dm6446/encode -v video_enc.264 -s audio_enc.g711 -t 20" I get a message that no video was detected. The videocfg boot environment variable is currently set to:
sendln 'setenv videocfg video=davincifb:osd0=720x480x16,1350K:osd1=720x480,1350K:vid0=720x480,2025K:vid1=720x480,2025K davinci_enc_mngr.ch0_output=COMPOSITE
davinci_enc_mngr.ch0_mode=NTSC'
I'm using the video output RCA jacks for the component video input which prevents us from having a composite video output, but we do have S-Video output. I changed the the videocfg bott variable to:
sendln 'setenv videocfg video=davincifb:osd0=720x480x16,1350K:osd1=720x480,1350K:vid0=720x480,2025K:vid1=720x480,2025K davinci_enc_mngr.ch0_output=SVIDEO davinci_enc_mngr.ch0_mode=NTSC'
We added two #define statements to tvp5146.h
/* analog muxing mode */
#define TVP5146_AMUX_COMPOSITE 0
#define TVP5146_AMUX_SVIDEO 1
#define TVP5146_AMUX_RGB 2
#define TVP5146_AMUX_YPRPB 3
My question is there a davinci_enc_mngr.ch0_input to change the input similar to the davinci_enc_mngr.ch0_output? Do I have to make changes to the kernel configuration to build the image? I'm concerned that the kernel config is not passing the right variable for the configuration we need.
Have we made all the changes we need to make this work?
Mike