There seems to be some small jitter and I can't lock onto the output signal perfectly with my LCD monitor each time I auto adjust and some pixels look a bit wider than others. Are you using fractional ratios of P2:pixel clocks or integer ratios? Fractional ratios could cause this issue.
There's actually a section I missed inserting "resi1" into (4 places)
OOPS - this breaks the code found it - it returns in a rep loop - another trap
.line
xcont m_bv,#0 'do before-visible part of scan line
resi1
rdfast #0,screen_base 'set fifo to read from current screen row
rep @.chr,#cols 'ready to output 8 pixels for each chr
rfbyte pa wc '2 get chr
bitc m_px,#16 '2 select colors using msb of chr
and pa,#$7F '2 trim chr to 7 bits
add pa,font_base '2 add offset for top/middle/bottom lines of chr
rdlut pa,pa '3 look up four lines of chr
shr pa,font_line '2 get line of interest into lower 8 bits
xcont m_px,pa '2 =15 output 8 single-bit pixels to streamer
' resi1 <<<<<<<<<<<<< cannot do this within a rep loop!!!
.chr
xcont m_bs,#0 'do before-sync part of scan line
resi1
xzero m_sn,#1 'do sync part of scan line
resi1
@Cluso99 said: VGA using Interrupts
Attached is VGA_640x480_text_80x40_interrupt which runs the streamer under interrupts
I found a couple of traps
waitx does not work in the non-interrupt code - I presume we cannot interrupt a waitx instruction ?
don't use calls in the interrupt code - I expect/presume we only have one internal stack space
' setint1 #INT_XMT
The INT_xxx definitions (as defined in the spin document 34r/s) appear not to work
You can do CALLs within ISRs. There is as much stack space as was available before the interrupt.
WAITX cannot be interrupted. Better to use loops for timing if interrupts are desired.
The places where calls would be dangerous is calling the blanks because you would return while still inside the call so the stack would likely get confused!!
@rogloh said:
There seems to be some small jitter and I can't lock onto the output signal perfectly with my LCD monitor each time I auto adjust and some pixels look a bit wider than others. Are you using fractional ratios of P2:pixel clocks or integer ratios? Fractional ratios could cause this issue.
Fractional
rdlong pa,#@clkfreq 'calculate streamer frequency
qfrac ##fpix,pa
getqx pa
shr pa,#1
setxfrq pa
I'm just using 216MHz (optimal for 1280x1024) where 50MHz multiples are better for 640x48.
Can you try with _clkfreq = 200_000_000 ?
At 200MHz: I'm not seeing jitter on my LCD but I am seeing some wider pixels (consistent) on some of the characters. It's a little blurry too. I'm using an Acer 24" 1920x1080.
Thanks Roger. I think the wider character pixel rows are because of the scaling in my monitor.
Thanks Brian. I'll try EVENT_XMT. Yes I realised why waitx cannot be interrupted but it's a trap because it's easily forgotten. Initially I just went gungho and changed the lot and of course nothing worked. Then I had to start again, implementing bits at a time and testing.
I've realised the rep section is tight timing and so it's likely missing the interrupt and once that happens the whole process stops. I tried a loop before I realised this. But I think the ultimate may be to do the whole line and display it in one go instead of a character at a time.
There's a few commented out sections.
Tested on P2-EVAL and RetroBlade2
Specify both VGA-Base (4 pins HBGR) and Vsync (VS) pins seperately
Flashes LEDs on both pins 0 and 56
Forgot to test this - it works so apply the change if you wish
Comments
There seems to be some small jitter and I can't lock onto the output signal perfectly with my LCD monitor each time I auto adjust and some pixels look a bit wider than others. Are you using fractional ratios of P2:pixel clocks or integer ratios? Fractional ratios could cause this issue.
You can do CALLs within ISRs. There is as much stack space as was available before the interrupt.
WAITX cannot be interrupted. Better to use loops for timing if interrupts are desired.
There's actually a section I missed inserting "resi1" into (4 places)
OOPS - this breaks the code
found it - it returns in a rep loop - another trap
.line xcont m_bv,#0 'do before-visible part of scan line resi1 rdfast #0,screen_base 'set fifo to read from current screen row rep @.chr,#cols 'ready to output 8 pixels for each chr rfbyte pa wc '2 get chr bitc m_px,#16 '2 select colors using msb of chr and pa,#$7F '2 trim chr to 7 bits add pa,font_base '2 add offset for top/middle/bottom lines of chr rdlut pa,pa '3 look up four lines of chr shr pa,font_line '2 get line of interest into lower 8 bits xcont m_px,pa '2 =15 output 8 single-bit pixels to streamer ' resi1 <<<<<<<<<<<<< cannot do this within a rep loop!!! .chr xcont m_bs,#0 'do before-sync part of scan line resi1 xzero m_sn,#1 'do sync part of scan line resi1I will update the code later
The places where calls would be dangerous is calling the blanks because you would return while still inside the call so the stack would likely get confused!!
Fractional
rdlong pa,#@clkfreq 'calculate streamer frequency qfrac ##fpix,pa getqx pa shr pa,#1 setxfrq paI'm just using 216MHz (optimal for 1280x1024) where 50MHz multiples are better for 640x48.
Can you try with _clkfreq = 200_000_000 ?
At 200MHz: I'm not seeing jitter on my LCD but I am seeing some wider pixels (consistent) on some of the characters. It's a little blurry too. I'm using an Acer 24" 1920x1080.
WAITX blocks interrupts.
Try
setint1 #EVENT_XMT
Yeah that looks cleaner with 200MHz setting Cluso.
Thanks Roger. I think the wider character pixel rows are because of the scaling in my monitor.
Thanks Brian. I'll try EVENT_XMT. Yes I realised why waitx cannot be interrupted but it's a trap because it's easily forgotten. Initially I just went gungho and changed the lot and of course nothing worked. Then I had to start again, implementing bits at a time and testing.
I've realised the rep section is tight timing and so it's likely missing the interrupt and once that happens the whole process stops. I tried a loop before I realised this. But I think the ultimate may be to do the whole line and display it in one go instead of a character at a time.
Here is the code. VGA 640x480 text 80x40
There's a few commented out sections.
Tested on P2-EVAL and RetroBlade2
Specify both VGA-Base (4 pins HBGR) and Vsync (VS) pins seperately
Flashes LEDs on both pins 0 and 56
Forgot to test this - it works so apply the change if you wish
mov ijmp1,#field 'set up streamer-empty interrupt ' setint1 #10 setint1 #EVENT_XMT