Curt Vendel Posted July 11, 2005 Share Posted July 11, 2005 I thought everyone would enjoy this little tidbit I dug up tonight while scanning through some internal Atari media: December 22, 1983 TO: Alan deSchweinitz FROM: Akio Tanaka To accomodate the video disc for the MARIA chip the color system clock must adhere to the broadcast standard. The complete dot interlace occur over 4 field sequence in NTSC. The complete dot interlace occur over 8 field sequence in PAL. For NTSC game the horizontal line has 227 color clocks. With external video horizental has 227.5 color clocks so a half clock is stuffed. Also normally the frame has 524 vertical lines. With external video frame has 525 vertical lines so a line is stuffed. For PAL game the horizontal line has 283.75 color clocks. With external video horizontal has 283.75 color clocks so they naturally match up. Also normally the frame has 624 vertical lines. With external video frame has 625 vertical lines so a line is stuffed. There is no need to do pulse correction for PAL beacause by using the 4/5 color clock as system clock , 227 system clock come to exactly 283.75 color clocks which is the PAL standard. The color clock is at 4433618.75 Hz. System clock is at Fc x 4/5 = 3546895.00 Hz. Now 228 system clocks equals 285 color clocks,but 227 system clocks equals 283.75 color clocks which is exactly what we want. So for PAL if we just synchronize the incoming burst and have the system clock PLL to 4/5 of color clock then chroma phase rotation is taken care of automatically. So for PAL there is no requirement for external pulse corection just line correction. COLOR TABLE FOR NTSC-PAL COLOR CODE | NTSC | PAL PAL ------------------------------------------------------------ | | EVEN ODD EVEN ODD ------------------------------------------------------------ 0000 | 0000 - | 1111 0000 - - 0001 | 0001 1D* | 0000 1111 1D 1D 0010 | 0010 2D | 0001 1110 2D 15D 0011 | 0011 3D | 0010 1101 3D* 14D* 0100 | 0100 4D | 0011 1100 4D 13D 0101 | 0101 5D | 0100 1011 5D 12D 0110 | 0110 6D | 0101 1010 6D 11D 0111 | 0111 7D | 0110 1001 7D 10D 1000 | 1000 8D | 0111 1000 8D 9D 1001 | 1001 9D | 1000 0111 9D 8D 1010 | 1010 10D | 1001 0110 10D 7D 1011 | 1011 11D | 1010 0101 11D 6D 1100 | 1100 12D | 1011 0100 12D 5D 1101 | 1101 13D | 1100 0011 13D 4D 1110 | 1110 14D | 1101 0010 14D 3D 1111 | 1111 15D | 1110 0001 15D 2D ------------------------------------------------------------ 1. For NTSC color code is directly decoded and the delay matches the color code. The color burst is coincident with 1D. 2. For PAL on EVEN field the color code is subtracted by one, however, the delay selected is same as the color code. On the ODD field decoder input is the complement of color code subtracted by one. The color burst is coincident with 1D. The EVEN field burst occurs at 3D. The ODD field burst occurs at 14D. COLOR TABLE FOR NTSC-PAL (GCC) COLOR CODE | NTSC | PAL PAL ------------------------------------------------------------ | | EVEN ODD EVEN ODD ------------------------------------------------------------ 0000 | 0000 - | 1111 0000 - - 0001 | 0001 1D* | 0000 1111 15D 15D 0010 | 0010 2D | 0001 1110 1D 14D 0011 | 0011 3D | 0010 1101 2D* 13D* 0100 | 0100 4D | 0011 1100 3D 12D 0101 | 0101 5D | 0100 1011 4D 11D 0110 | 0110 6D | 0101 1010 5D 10D 0111 | 0111 7D | 0110 1001 6D 9D 1000 | 1000 8D | 0111 1000 7D 8D 1001 | 1001 9D | 1000 0111 8D 7D 1010 | 1010 10D | 1001 0110 9D 6D 1011 | 1011 11D | 1010 0101 10D 5D 1100 | 1100 12D | 1011 0100 11D 4D 1101 | 1101 13D | 1100 0011 12D 3D 1110 | 1110 14D | 1101 0010 13D 2D 1111 | 1111 15D | 1110 0001 14D 1D ------------------------------------------------------------ 1. For NTSC color code is directly decoded and the delay matches the color code. The color burst is coincident with 1D. 2. For PAL on EVEN field the color code is subtracted by one, so the delay is color code subtracted by one. On the ODD field decoder input is the complement of color code subtracted by one. The color burst is coincident with 15D. The EVEN field burst occurs at 2D. The ODD field burst occurs at 13D. NTSC-PAL COLOR TABLE (PRESENT GCC) TRUTH TABLE (STAGE 1) DIN | DOUT COUT -------------------- 0 | 0 0 1 | 1 1 TRUTH TABLE (STAGE 2-3) DIN CIN | DOUT COUT ---------------------------- 0 0 | 0 0 0 1 | 1 1 1 0 | 1 1 1 1 | 0 1 COLOR CODE DECODER INPUT DELAY COLOR BURST ---------------------------------------------------------------- EVEN ODD EVEN ODD EVEN ODD ---------------------------------------------------------------- 0000 0000 1111 - - 0001 1111 0000 15D 15D 0010 1110 0001 0011 1101** 0010** 0100 1100 0011 0101 1011 0100 0110 1010 0101 0111 1001 0110 1000 1000 0111 1001 0111 1000 1010 0110 1001 1011 0101 1010 1100 0100 1011 1101 0011 1100 1110 0010 1101 1101 0010 1111 0001 1110 SINCE THE FF IS TOGGLED ON THE NEGATIVE EDGE OF THE CBURST SIGNAL, ON THE SAME LINE BURST AND THE REST OF THE SIGANL ARE ON COMPLEMENTARY LINE. Curt Quote Link to comment Share on other sites More sharing options...
supercat Posted July 11, 2005 Share Posted July 11, 2005 I thought everyone would enjoy this little tidbit I dug up tonight while scanning through some internal Atari media: My recollection of playing the 7800 is that the color phase is the same every line. Cutting a half cycle would reduce artifacting (breaking games that relied upon it, but making other games look better). Is this piece suggesting that the Maria chip should have been modified to do this? Quote Link to comment Share on other sites More sharing options...
Ferris Posted July 11, 2005 Share Posted July 11, 2005 Bout time this surfaced. Quote Link to comment Share on other sites More sharing options...
Curt Vendel Posted July 11, 2005 Author Share Posted July 11, 2005 The Maria still had refinements including some fairly major "fixes" up to about April 84 and then final product in May, so it is possible. Curt I thought everyone would enjoy this little tidbit I dug up tonight while scanning through some internal Atari media: My recollection of playing the 7800 is that the color phase is the same every line. Cutting a half cycle would reduce artifacting (breaking games that relied upon it, but making other games look better). Is this piece suggesting that the Maria chip should have been modified to do this? 889519[/snapback] Quote Link to comment Share on other sites More sharing options...
EricBall Posted July 11, 2005 Share Posted July 11, 2005 Although the stock MARIA probably doesn't bother with the half cycle, it would be necessary if it was going to sync properly with an external video signal, i.e. a LaserDisc player via the expansion port. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.