• State
  • Locked Locked
  • Replies 15 replies
  • Subscribers 102 subscribers
  • Views 1483 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Video outputs on DM6446

Mindong
Prodigy 30 points

Other Parts Discussed in Thread: OMAP3530

Hi all,

In our design, using DM6446, the system needs two different video outputs to different display terminals. You know there are 4 DAC video outputs on DM6446 device. Does It can support to output two video channels on these two DAC line, which binding to two different video display devices?

As I know, there is only one OSD and VENC modules in VPBE block. If displaying two different video on two devices, how does it work?

Thanks,

Mindong

 

   

  • 0
    TI__Mastermind 37340 points

    Mindong,

    As you mentioned, there is only one channel that can go thru OSD and VENC at any one time.   Hence you can certainly send this single channel source to two different displays, but if you need two different, this will not be possible with this part.  Optionally, you can display one channel while recording a separate channel into DDR2 or HDD for display at a later time.

  • 0 in reply to Juan Gonzales
    Prodigy 30 points

    Hi Juan,

    Thank you for clear clarification. As your understanding, the only one channel video can be outputed to all the linked display terminals.  So, the digital display interfce and analog display interface of VPBE is also drived by same video source. Is it right?  Does it mean that DM6446 VPBE module only supports one video channel to all digital and analog interface?

    I noticed OMAP3530 or 3525 has a dual-display interface. The 2 buffer video data can feed to two DAC.  In thie view, is it possible to disable DM6446 OSD, and feed 2 video datas to 2 of 4 DACs of DM6446 for supporting our requirements?

    Regards,

    Mindong

     

     

     

  • 0 in reply to Mindong
    TI__Mastermind 37340 points

    OMAP3530 hardware architecture is different from DM6446.  In DM6446, video data goes from DDR2 to OSD to VENC to DACs (and/or digital output).  There is no way to bypass OSD or VENC and these can only support one channel at a time.  Therefore, all analog and digital outputs are fed from the same video source on DM6446.

    If you do not need to OSD, DM6467 may be a good option (supports two input channels and two output channels) and is very similar to DM6446 in most other aspects.  DM648 is a multiple video port part, but it is much more different from DM6446.  And of course, OMAP3530 can also accomplish this.  however, DM6467 would be the closest part to DM6446.

  • 0 in reply to Juan Gonzales
    Prodigy 30 points

    Hi Juan, Thank you very much.

  • 0 in reply to Mindong
    Prodigy 90 points

    I have the same task as topic starter had but I should use DM6437 chip for this.  I decided not to create separate topic and just ask my questions here.

    1. Am I right that DM6437 has the same FPBE as DM6446. And I can not get out two different output video streams?
    2. Can I show the same input video on two different DACs but add bitmap only to one of the DAC?

     

  • 0 in reply to Vasily9089
    TI__Mastermind 37340 points

    1. Yes, VPBE should be mostly the same for both of these platforms

    2. Unfortunately, this will not be possible with this platform, all video and OSD windows are merged prior to selecting which output channels will display the data.  You can select which output channels to enable or disable, but all will pretty much show the same video data.

  • 0 in reply to Juan Gonzales
    Intellectual 615 points

    Hi Juan,

    I find a mysterious problem, I testing the video out signal from Davinci DM6441 with the assistance of V4Monitor (Magin Monitor compact monitoring system)

    I find, that video out signal that come out from the Davinci DM6441 having Horizontal jitter signal.( Horizontal signal frame are shake)

    Actuality the video out look on the TV screen stable and see excellent.

    Furthermore I have DVEVM and the video out signal result is same.

    I download the VPBE PDF file( sprue37c.pdf)  but I didn`t know if this is the point of problem.

    Any idea?

     

  • 0 in reply to EranS
    TI__Expert 6635 points

    I am not familiar with the monitoring / testing device you mentioned, but how does your system look like? Is it your own board you are testing?

    Which silicon revision and Linux LSP are you using? Can it be that you are having the issue mentioned in advisory 2.1.42 of the silicon errata?

    Perhaps the evm is based on a previous silicon so that you do not see this effect on the evm?

  • 0 in reply to EranS
    TI__Mastermind 37340 points

    I am not too familiar with V4Monitor either, can you provide a more detailed description of what you expect to see vs what you are seeing (horizontal jitter).  Our video port is very floxible and programmable and I suspect that is there is some mis-aligment with the video timing, it can be resolved by properly adjusting the VPBE registers.

  • 0 in reply to Juan Gonzales
    Intellectual 615 points

    Of course, as you can see the first picture is from Davinci DM644x and you can see the signal frame as V4Monitor transmit it to TV.

    On the other hand, the second picture is Color Bar from Bar generator ( stable video signal).
     
    This is the Horizontal jitter who I mean to.
     
    This video signal took form DVEVM board, but is look same from my DM6441 board.
     
     
    I"m using with the DVEVM and Davinci DM6441 chip:
     
    silicon revision :  $NE - 7AA7HIW TMS320 DM6441ZWT

    Linux LSP: 1_30

    This is the DVEVM Boot setup:

     
    bootargs=video=davincifb:vid0=720x576x16,2500K:vid1=720x576x16,2500K:osd0=720x576x16,2025Kdavinci_enc_mngr.ch0_output=COMPOSITE davinci_enc_mngr.ch0_mode=pal console=ttyS0,115200n8 noinitrd rw ip=dhcp root=/dev/nfs nfsroot=console=ttyS0,115200n8 noinitrd rw ip=dhcp root=/dev/nfs nfsroot=192.168.10.93:/home/user/workdir/filesys,nolock mem=120M:
     
     

     

     

     

  • 0 in reply to GiPa
    Intellectual 615 points

    Of course, as you can see the first picture is from Davinci DM644x and you can see the signal frame as V4Monitor transmit it to TV.

    On the other hand, the second picture is Color Bar from Bar generator ( stable video signal).
     
    This is the Horizontal jitter who I mean to.
     
    This video signal took form DVEVM board, but is look same from my DM6441 board.
     
     
    I"m using with the DVEVM and Davinci DM6441 chip:
     
    silicon revision :  $NE - 7AA7HIW TMS320 DM6441ZWT

    Linux LSP: 1_30

    This is the DVEVM Boot setup:

     
    bootargs=video=davincifb:vid0=720x576x16,2500K:vid1=720x576x16,2500K:osd0=720x576x16,2025Kdavinci_enc_mngr.ch0_output=COMPOSITE davinci_enc_mngr.ch0_mode=pal console=ttyS0,115200n8 noinitrd rw ip=dhcp root=/dev/nfs nfsroot=console=ttyS0,115200n8 noinitrd rw ip=dhcp root=/dev/nfs
     nfsroot=192.168.10.93:/home/user/workdir/filesys,nolock mem=120M:
  • 0 in reply to GiPa
    Intellectual 615 points

    DVEVM is behave the same and put out error horizontal signal!!

     

  • 0 in reply to EranS
    TI__Mastermind 37340 points

    from looking at the images, it appears like an interlaced effect where one of the fields is not properly offset.

    Our latest DVSDK demos use resizer to de-interlace video which may yield better results for you, though the offset seems quite large in your screen captures.

  • 0 in reply to Juan Gonzales
    Intellectual 615 points

    Juan

    Can you be more specific , which last version of DVSDK have the de-interlace?

    As you say, if the offset is look large, what more than resizer can be the problem?

  • 0 in reply to EranS
    TI__Mastermind 37340 points

    Since we are talking about DM644x, the latest DVSDK would be available via the corresponding product folder, which should eventually take you to http://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/dvsdk/DVSDK_2_00/latest/index_FDS.html

    What I ment about the offset being large is that aside from the interlaced problem, you may also be looking at correcting the line offset for one of the two fields.  It appears one of the fields (even or odd lines) has a much bigger offset than the other (or than you expect from interlacing).