@randombler I know DPB is before your time but maybe you can remember any details of the DPB booting up? We think these dots are part of the painting menu being drawn. But this could be first output from the Paintbox Emulator!
-
-
When the old DPB from the Beeb was fired up with no HDD, it just used to output what looked like maybe an uninitialized framebuffer (especially so soon after power-up). Also locked like Hsync wasn't locked, though... https://youtu.be/xGEXQm7ytrA?t=231 …pic.twitter.com/1Py6apUK7I
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjille @AshEvans81, @DextersTechLab ja
@TheMogMiner Could you possibly upload a printf log of the writes to the framebuffer? Could be interesting to see what it's trying to do.1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Sure, but here's an overview: It's issuing a bunch of "Fast Wipe Framestore" and "Draw" commands to Framestore II. Both Luma and Chroma enabled, /Brush Invert and Brush Zero asserted. The Draw commands have Fixed Color Select enabled, the Fast Wipe commands don't.
1 vastaus 0 uudelleentwiittausta 1 tykkäys -
Vastauksena käyttäjille @TheMogMiner, @AshEvans81 ja
It's issuing all of the commands with CXPOS and CYPOS in the first 128 lines of Frame Store II, which according to the service manual means it's trying to draw a menu. Currently I just plot a single pixel at CXPOS and CYPOS. The structure looks rather like the menu system.
1 vastaus 0 uudelleentwiittausta 1 tykkäys -
Interesting. I expected it would try to clear the FB, but not actually draw the menu. I thought the bulk of the code for that would get loaded from HDD on boot. I recall the SM saying that the menu is always drawn, though, just hidden from view by twiddling address bits.
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
I'm surprised it doesn't clear the FB, but it makes sense given the photo of the screen you posted. It's also doing writes with CSR=8, but I don't log the values. Per the Brush Store Card's desc, that's how data goes from System Bus to Brush Data Bus, so perhaps I should.
1 vastaus 0 uudelleentwiittausta 1 tykkäys -
Vastauksena käyttäjille @TheMogMiner, @AshEvans81 ja
8000, 8032, 8000, 80ad, 8000, 8032... according to the Brush Store Card schematic, the upper 8 bits toggle between U and V, and the lower 8 bits are always Y. So yeah. Hex 0x80 sounds about right for some grayscale pixel data. Let's try storing it and visualizing it.
1 vastaus 0 uudelleentwiittausta 1 tykkäys -
Yep, that sounds fairly standard for bt.606 / bt.656 style vid. So U and V will have a value of 128 at the "zero" point. I think some stuff just assumes that say the U sample is output first, directly after Hblank?
3 vastausta 0 uudelleentwiittausta 1 tykkäys -
Vastauksena käyttäjille @AshEvans81, @TheMogMiner ja
I can't remember if it would be full range 0-255 or not for each sample, though? https://en.wikipedia.org/wiki/YUV#Studio_swing_for_BT.601 …
2 vastausta 0 uudelleentwiittausta 1 tykkäys
Not that much of an issue quite yet, for the time being I'm just using a YUV -> RGB formula that I snaffled off of some site somewhere. I'll worry about whether it's full-swing or not a bit later down the line. ;)
-
-
True, it usually looks passable with the wrong range for Y anyway. Don't need that stuff to see some output. ;)
0 vastausta 0 uudelleentwiittausta 0 tykkäystäKiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
U and V range was 16-240, with offset of 128, so white/black/grey was u,V=128. Y was 16-235. By using luma only or chroma only painting you could generate illegal colours on the NTSC output of early machines that looked good on your local monitor but caused transmitter to choke.
0 vastausta 0 uudelleentwiittausta 3 tykkäystäKiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.