[solved]Problem Using OSD Window 0 and Video Window 0

Elric

Hi,

I have a problem and I will be appreciated for any help or advice.

-Enabling the video window 0 is a constraint before enabling the osd window 0.

-Currently I need only osd window 0 because we think rgb565 is suitable for our requirements. Hence while I was experimenting with the VPBE example in spraap3a document, (http://focus.ti.com/lit/an/spraap3a/spraap3a.pdf) I decided to disable the image that video window 0 displays (also disabling video window 1 and cursor window), so only the osd window 0 display shall remain on the display screen.

-While trying this, I realized that if I don't queue a frame buffer of video window 0, creating a channel for osd window 0 returns NULL.

That is:

//test.c

// enqueue the 1st frame buffer
if ( IOM_COMPLETED != FVID_queue( vid0Handle, vid0AllocFB[0] ) ) {
return;
}

If I comment out the code above, then:

//configureOsdWin0_TIDSP_RGB565.c

osd0Handle = FVID_create( "/VPBE0", IOM_INOUT, NULL, &beinitParams, &gioAttrs );

the FVID_create function above returns NULL.

-Do you have any idea why this should happen?

-To solve this, I decided to queue a frame buffer (of video wnd 0) which is set to NULL. (vid0AllocFB[0]->frameBufferPtr = NULL;)

In this case, I can see the image in osd window 0, and of course there is no image in video window 0. But the visible parts of the video window 0 is now set to a distorted green, and the edges of the video window 0 is vibrating.

How can I correct this?

Thank you in advance,

A. Erman Kulunyar

over 17 years ago

0 BrandonAzbell over 17 years ago

TI__Guru 64610 points

Firstly, I would say based on the VPBE User's Guide (SPRU952) in Section 1.2.1, the following restriction exists for the OSD module in the VPBE.

Both the OSD windows and VIDWIN1 should be fully contained inside VIDWIN0. This means that both the Y and X position of either the OSD windows or VIDWIN1 should be greater (not equal) to the Y and X position of VIDWIN0.
This also requires VIDWIN0 to be enabled when using either the OSD windows or VIDWIN1.

Therefore, this substantiates the ddc_vpbe driver in forcing the VIDWIN0 to be enabled, created, etc. Based on your comments and code samples, it appears the FVID_create() for OSDWIN0 is checking if VIDWIN0 is enabled. This is true as I looked at pspdrivers\drivers\vpbe\src\ddc_vpbe.c and the DDC_VPBEOpenHandle() does a check if the PSP_VPBE_VIDEO_0 is active. If not, it returns a NULL handle for PSP_VPBE_OSD_0 or PSP_VPBE_OSD_1. This agrees with the observations you have made.

Elric said:

-To solve this, I decided to queue a frame buffer (of video wnd 0) which is set to NULL. (vid0AllocFB[0]->frameBufferPtr = NULL;)
In this case, I can see the image in osd window 0, and of course there is no image in video window 0. But the visible parts of the video window 0 is now set to a distorted green, and the edges of the video window 0 is vibrating.

How can I correct this?

Did the FVID_queue() complete successfully, without an error status, when providing it a NULL frame buffer as in (vid0AllocFB[0]->frameBufferPtr = NULL;)?
In the pspdrivers\drivers\vpbe\src\ddc_vpbe.c, DDC_VPBEQueueFrameBuffer() performs a check on the frame buffer, which I would have thought returns an error.

0 Elric over 17 years ago in reply to BrandonAzbell

Expert 1250 points

Hi,

I am aware of constraints about video window 0 before activating the osd window 0 or 1. But I thought "activating" process is done by callind FVID_create(). But the problem is, FVID_create() for osdwin0 fails when FVID_create is successfully called for vidwin0 but no frame buffer is enqued of vidwin0. I think this is an interesting problem. Maybe enqueing a frame buffer for vidwin0 is a part of enabling it?

BrandonAzbell said:
Did the FVID_queue() complete successfully, without an error status, when providing it a NULL frame buffer as in (vid0AllocFB[0]->frameBufferPtr = NULL;)?
In the pspdrivers\drivers\vpbe\src\ddc_vpbe.c, DDC_VPBEQueueFrameBuffer() performs a check on the frame buffer, which I would have thought returns an error.

I am not working on my computer now, but I guess the example application was designed to terminate when a FVID function returns NULL, hence it should return without an error.

0 BrandonAzbell over 17 years ago in reply to Elric

TI__Guru 64610 points

The places in the ddc_vpbe.c that I see checks for "Active" are including a test of (TRUE == handleOsdCH->PlaneActive). There are two places where this variable is set true, in DDC_VPBEStartEngine() and DDC_VPBEQueueFrameBuffer().
Therefore, your comment about enqueing a frame buffer for the first time on vidwin0 is valid.

But in the DDC_VPBEStartEngine(), it seems to be implying a valid frame buffer in the queue. So, I would say your comment about needing to queue a frame buffer for vidwin0 is very likely.
I'm basing this on inspection of the driver code as I haven't tried running it.

Processors

Processors forum

[solved]Problem Using OSD Window 0 and Video Window 0