cvideo evm test example - Beacon EmbeddedWorks - TDG Forum -

Go to previous topic

Go to next topic

Last Post 12 Oct 2009 07:17 PM by dblockil

cvideo evm test example

1 Replies

Author

Messages

dblockil

New Member

New Member
Posts:

02 Oct 2009 01:33 PM
I have the cvideo example working on my EVM. Now what I am trying to do is convert the YCbCr format to RGB. I think I understand how the data is stored in DDR Ram, at least I am able to produce a good BW image plot in CCS 3.3 image graphic view. My assumption is that Y0,Y2,Y4 etc is store in LUM_TOP buffer, Y1,Y3,Y5 etc is stored in LUM_BOTTOM buffer. CHROM_BOT and CHROM_BOTTOM I think are what have me confused. Is Cb stored in CHROM_TOP and Cr stored in CHROM_BOTTOM? What is the formula to convert from YCrCb to RGB? I have tried the following code below but have not been able to get a good RGB image. The commented code are a number of my different trys. This code is inside a dual r and c for loop, looping through 480X720 rgb values. Thanks Dan Block myrgb = (rgb )(0xC2000000); ytop = (unsigned char )(0xC1000000); ybottom = (unsigned char )(0xC1000000 + 240720); Cb = (unsigned char )(0xC1000000 + 2240720); Cr = (unsigned char )(0xC1000000 + 3240720); /* coeff[] = { 0x2000, 0x2BDD, -0x0AC5, -0x1658, 0x3770 }; / / / / [ 1.0000 0.0000 1.3707 ] [ Y' - 16 ] [ R'] / / [ 1.0000 -0.3365 -0.6982 ] * [ Cb - 128 ] = [ G'] / / [ 1.0000 1.7324 0.0000 ] [ Cr - 128 ] [ B'] / / / myrgb[r720+c].r=(unsigned char)( (ytop[(r/2)720+c]-16) + 1.3707(Cr[(r/2)720+c]-128));myrgb[r720+c].g=(unsigned char)( (ytop[(r/2)720+c]-16) - 0.3365(Cb[(r/2)720+c]-128) - 0.6982(Cr[(r/2)720+c]-128) );myrgb[r720+c].b=(unsigned char)( (ytop[(r/2)720+c]-16) + 1.7324(Cb[(r/2)720+c]-128) ); myrgb[(r+1)720+c].r=(unsigned char)( (ybottom[((r+1)/2)720+c]-16) + 1.3707(Cr[((r+1)/2)720+c]-128)); myrgb[(r+1)720+c].g=(unsigned char)( (ybottom[((r+1)/2)720+c]-16) - 0.3365(Cb[((r+1)/2)720+c]-128) - 0.6982(Cr[((r+1)/2)720+c]-128) ); myrgb[(r+1)720+c].b=(unsigned char)( (ybottom[((r+1)/2)720+c]-16) + 1.7324(Cb[((r+1)/2)720+c]-128)); / 2. Y'CbCr -> RGB conversion with the 219-level range of Y' / / expanded to fill the full RGB dynamic range. (The matrix / / has been scaled by 255/219.) Expected ranges are [16..235] / / for Y' and [16..240] for Cb and Cr. / / / / coeff[] = { 0x2543, 0x3313, -0x0C8A, -0x1A04, 0x408D }; / / / / [ 1.1644 0.0000 1.5960 ] [ Y' - 16 ] [ R'] / / [ 1.1644 -0.3918 -0.8130 ] * [ Cb - 128 ] = [ G'] / / [ 1.1644 2.0172 0.0000 ] [ Cr - 128 ] [ B'] / // myrgb[r720+c].r=(unsigned char)( 1.1644(ytop[(r/2)720+c]-16) + 1.5960(Cr[(r/2)720+c]-128)); // myrgb[r720+c].g=(unsigned char)( 1.1644(ytop[(r/2)720+c]-16) - 0.3918(Cb[(r/2)720+c]-128) - 0.8130(Cr[(r/2)720+c]-128) ); // myrgb[r720+c].b=(unsigned char)( 1.1644(ytop[(r/2)720+c]-16) + 2.0172(Cb[(r/2)720+c]-128) ); // myrgb[(r+1)720+c].r=(unsigned char)( 1.1644(ybottom[((r+1)/2)720+c]-16) + 1.5960(Cr[((r+1)/2)720+c]-128)); // myrgb[(r+1)720+c].g=(unsigned char)( 1.1644(ybottom[((r+1)/2)720+c]-16) - 0.3918(Cb[((r+1)/2)720+c]-128) - 0.8130(Cr[((r+1)/2)720+c]-128) ); // myrgb[(r+1)720+c].b=(unsigned char)( 1.1644(ybottom[((r+1)/2)720+c]-16) + 2.0172(Cb[((r+1)/2)720+c]-128)); // myrgb[r720+c].r=(unsigned char)( ytop[(r/2)720+c] + 1.402Cr[(r/2)720+c]); // myrgb[r720+c].g=(unsigned char)( ytop[(r/2)720+c] - 0.344Cb[(r/2)720+c] - 0.714Cr[(r/2)720+c] ); // myrgb[r720+c].b=(unsigned char)( ytop[(r/2)720+c] + 1.772Cb[(r/2)720+c] ); // myrgb[(r+1)720+c].r=(unsigned char)( ybottom[((r+1)/2)720+c] + 1.402Cr[((r+1)/2)720+c]); // myrgb[(r+1)720+c].g=(unsigned char)( ybottom[((r+1)/2)720+c] - 0.344Cb[((r+1)/2)720+c] - 0.714Cr[((r+1)/2)720+c] ); // myrgb[(r+1)720+c].b=(unsigned char)( ybottom[((r+1)/2)720+c] + 1.772Cb[((r+1)/2)720+c] ); / Cint = (int)ytop[(r/2)720+c] - 16; // Dint = (int)((Cr[(r/2)720+c]1.403) + 0.5) - 128; // Eint = (int)((Cb[(r/2)720+c]1.773) + 0.5) - 128; Dint = (int)((Cr[(r/2)720+c]) + 0.5) - 128; Eint = (int)((Cb[(r/2)720+c]) + 0.5) - 128; redint=( 298Cint + 409Eint + 128)>>8; greenint=( 298Cint - 100Dint - 208Eint + 128 )>>8; blueint=( 298Cint + 516Dint + 128)>>8; myrgb[r720+c].r=(unsigned char)redint; myrgb[r720+c].g=(unsigned char)greenint; myrgb[r720+c].b=(unsigned char)blueint; Cint = (int)ybottom[((r+1)/2)720+c] - 16; redint=( 298Cint + 409Eint + 128)>>8; greenint=( 298Cint - 100Dint - 208Eint + 128 )>>8; blueint=( 298Cint + 516Dint + 128)>>8; myrgb[(r+1)720+c].r=(unsigned char)redint; myrgb[(r+1)720+c].g=(unsigned char)greenint; myrgb[(r+1)720+c].b=(unsigned char)blueint; / // BW // myrgb[r720+c].r=(unsigned char)( ytop[(r/2)720+c]); // myrgb[r720+c].g=(unsigned char)( ytop[(r/2)720+c]); // myrgb[r720+c].b=(unsigned char)( ytop[(r/2)720+c]); // myrgb[(r+1)720+c].r=(unsigned char)( ybottom[((r+1)/2)720+c]); // myrgb[(r+1)720+c].g=(unsigned char)( ybottom[((r+1)/2)720+c]); // myrgb[(r+1)720+c].b=(unsigned char)( ybottom[((r+1)/2)*720+c]);

dblockil

New Member

New Member
Posts:

12 Oct 2009 07:17 PM
I think I have answered my own question after a bunch of trial and error. Here is the way I think the YCbCr data is stored in the cvideo example. ytop = (unsigned char )(0xC1000000); ybottom = (unsigned char )(0xC1000000 + 240720); Ctop = (unsigned char )(0xC1000000 + 2240720); Cbottom = (unsigned char )(0xC1000000 + 3240*720); Naming convention Pixel Row 0 Column 0 Y0Cb0Cr0 Pixel Row 1 Column 0 Y1Cb0Cr0 Pixel Row 0 Column 1 Y2Cb2Cr2 Pixel Row 1 Column 1 Y3Cb2Cr2 Pixel Row 0 Column 2 Y4Cb4Cr4 Pixel Row 1 Column 2 Y5Cb4Cr4 Pixel Row 0 Column 3 Y6Cb6Cr6 Pixel Row 1 Column 3 Y7Cb6Cr6 Then you pull the YCbCr data out in this fashion Y0 = ytop(0) ; Y1 = ybottom(0); Cb0 = Ctop(0); Cr0 = Ctop(1); Y2 = ytop(1); Y3 = ybottom(1); Cb2 = Cbottom(0); Cr2 = Cbottom(1); Then repeating again Y4 = ytop(2) ; Y5 = ybottom(2); Cb4 = Ctop(2); Cr4 = Ctop(3); Y6 = ytop(3); Y7 = ybottom(3); Cb6 = Cbottom(2); Cr6 = Cbottom(3); Does this look correct?

Forums Texas Instruments SOMs and Development Kits OMAP-L138 EVM Development Kit