@Cluso, I linked Eric's drivers above. When he and I were tinkering with P1 video, a few different ideas fell out of that, plus he made a very nice all software driver that generates very good color signals given the coarse P1 DAC. That one should port over very nicely, IMHO. Some of that came from looking at how older computers did graphics, and other ideas came from various ways NTSC can be exploited to get lots of colors from just a few bits of DAC capability.
PAL is less forgiving in some ways, but... it's also exploitable in some other ways. That never got done much on P1 just because it never really output a great PAL signal. Maybe this time...
I kind of like the idea of adding the two signals together electrically. Right now, I'm setting up on P1 to attempt that, and running some code to get timings and test a few things. It's been a while.
I kind of like the idea of adding the two signals together electrically. Right now, I'm setting up on P1 to attempt that, and running some code to get timings and test a few things. It's been a while.
Yes I like that idea too.
I think Bill Henning used to output s-video at one point, if I remember correctly.
FWIW I've decided to make a small Si504 oscillator board, about same size as HC49 crystal, so we can retrofit a flexible oscillator to xtal pluggable boards, and control its frequency from a servo connection
Yes, his "Propcade" board had an S-video connector on it. P1 does have a VCFG option to output the chroma on start_pin+3, which is what was used on Bill's board.
On P1, this two DAC thing can be roughly simulated with two COGS, each using the coarse video DAC, synchronized. One nice thing about doing it that way is the ability to run the monochrome channel at pretty high resolution, optionally toggling the colorburst, and then running the color channel at a more modest resolution. Let the TV circuits sort it all out, and given smooth signals, it's going to look pretty great.
1cog idea
vga mode on 4 pins. for luminance/sync
Use counterB as a phase-shiftable 3.579545 MHz square wave to pin 5, use 3 lowpass filters in series to create something close to a sine wave.
Instead of a 5mhz crystal, use a 3.58 or 7.15909 Mhz ? as powerof2 will be jitter free.
4.43361875 or 8.86724Mhz for PAL
While 4 pixels are displaying (low resolution) using %%4321
You have time to make 2 changes to PHSB to change its phase.
So 4:2 resolution that is pretty much the norm in NTSC broadcast anyway, but harder to bitmap
While PAL was better (more scan lines) than NTSC, monitors and tvs here (in OZ where we are PAL) have all been capable of doing NTSC for more than a decade. So perhaps PAL is no longer a requirement, since it is harder to implement.
The P1 sure made it easy to do NTSC color. It would be nice to have at least something simple to generate at least one composite color signal.
potatohead: interested to hear how you go with the tests.
I had one of those... PAL sets are hard to get where I live. Next best thing is a capture card, and those tend to make way too many compensations for realistic testing.
Comments
PAL is less forgiving in some ways, but... it's also exploitable in some other ways. That never got done much on P1 just because it never really output a great PAL signal. Maybe this time...
I kind of like the idea of adding the two signals together electrically. Right now, I'm setting up on P1 to attempt that, and running some code to get timings and test a few things. It's been a while.
Yes I like that idea too.
I think Bill Henning used to output s-video at one point, if I remember correctly.
FWIW I've decided to make a small Si504 oscillator board, about same size as HC49 crystal, so we can retrofit a flexible oscillator to xtal pluggable boards, and control its frequency from a servo connection
On P1, this two DAC thing can be roughly simulated with two COGS, each using the coarse video DAC, synchronized. One nice thing about doing it that way is the ability to run the monochrome channel at pretty high resolution, optionally toggling the colorburst, and then running the color channel at a more modest resolution. Let the TV circuits sort it all out, and given smooth signals, it's going to look pretty great.
vga mode on 4 pins. for luminance/sync
Use counterB as a phase-shiftable 3.579545 MHz square wave to pin 5, use 3 lowpass filters in series to create something close to a sine wave.
Instead of a 5mhz crystal, use a 3.58 or 7.15909 Mhz ? as powerof2 will be jitter free.
4.43361875 or 8.86724Mhz for PAL
While 4 pixels are displaying (low resolution) using %%4321
You have time to make 2 changes to PHSB to change its phase.
So 4:2 resolution that is pretty much the norm in NTSC broadcast anyway, but harder to bitmap
While PAL was better (more scan lines) than NTSC, monitors and tvs here (in OZ where we are PAL) have all been capable of doing NTSC for more than a decade. So perhaps PAL is no longer a requirement, since it is harder to implement.
The P1 sure made it easy to do NTSC color. It would be nice to have at least something simple to generate at least one composite color signal.
potatohead: interested to hear how you go with the tests.
But I think a dedicated summer + driver is best, this one is 3.3v and is under $1 (0.65mm msop8).
It got SAG so you can use 100uF+22uF caps as that is cheaper/smaller than a single 470uF
http://www.mouser.com/Search/ProductDetail.aspx?R=NJM2274R-TE1virtualkey51310000virtualkey513-NJM2274R-TE1
datasheet: http://www.njr.com/semicon/PDF/NJM2274_E.pdf
This IC don't need any output caps at all as it generate negative 0.3v, but it does NOT come in a YC mixer version unfortunately.
http://www.mouser.com/ProductDetail/Analog-Devices/ADA4431-1YCPZ-R7/?qs=sGAEpiMZZMutXGli8Ay4kOSwB%252bAxEb%252baTuPzCYAIkKM%3d