I am testing this on Sony hb-f700d, real hardware,
Including settings like reg8 VR select, using bits 7 and 2 reg 1, bit 7 reg 0, max amount of sprites,
Currently only tested in basic, not assembly yet. In basic the settings official returns only part of screen when loading image.
Furthermore with some settings for get a similar results as screen '9' and sometimes a screen like 7 with a border around it.
If the same result was achieved with value 6 and 7 set to m5-m3 , and it would be 6 for screen8 not 7.
There must be a good (technic/politic) reason. Maybe this mode is also available in another VDP (for example YM7101/2/3) and was especially made for some other company,with exclusive modes.
Any assistance to unravel this mistery is welcome
Now it is also clear why V9958 has additional screen 10-12 and not a 9 . Nine already existed...
I always had the impression that the system was designed to be able to cohabit that of the SG-1000 and Master system. At least at the start.
Maybe the famous "screen 4" used on Sega Master System?
https://www.smspower.org/Development/Modes
"The Master System VDP is an extended version of the TMS9918a VDP, with an enhanced mode for "SMS mode" graphics, featuring:
16-colour tiles using one of two palettes
64 sprites
X and Y scrolling
This is known as "mode 4", with "mode 5" appearing on the Genesis. Modes 0-3 are the standard TMS9918a "modes", seen in SG-1000 and SC-3000 software.
Mode 4 is enabled by setting bit 2 of VDP Register 0. This is the mode used by almost all Master System and Game Gear games.
When in mode 4, the TMS9918a mode bits enable extended height modes. One of these is used by some Codemasters games, such as Micro Machines, to enable a 256x224 resolution."
Sega and MSX definitely have some compatibility.
One result is for example the Sakhr AX-990. MSX1 with V9938, and able to play sms/Genesis games.
I cannot recall the documentation I have read about Sega, but it mentioned, like cheat mode with game controller, move joystick or joypads counter clockwise around and pressing fire button a and than b, to select M3 than M2, than M5 and M4, to enable the specific screen mode.
The Japanese are sometimes smarter than they look
In my opinion and logic I think it should be a mode with 64 colors. Why?
When selected M5 of VDP #0 starts with screen 6 (screen mode in basic)
Screen 6 | 4 colors
Screen 7 | 4 * 4 = 16 colors
Screen 8 | 16 * 4 = 64 colors ( undocumented and expected)
Screen 9 | 64 * 4 = 256 colors (current screen 8 in Basic)
If this is the case, then comes another question: are the 64 colors a free selection within 512 colors or just 64 fixed colors?
In other words: closer to official SCREEN 5 or closer to official SCREEN 8 or maybe a subtle mix of free colors + fixed colors (what could anticipate the mix operated later with SCREEN 10/11) ?
Thanx accumulator. May i suggest to to another test? (In basic)
1) set screen 8 from basic
2) draw a filled circle on screen with some color (maybe some border color and fill color)
3) Show some sprites, maybe a row 8 16x16 square sprites.
4) by using Vdp(n) set the unknown mode
Report us what's happening
something like this:
10 Screen 8,2 20 circle (127,95), 90, 100 30 paint step (0,0), 19,100 40 S$=STRING$(32,255) 45 FORT%=0TO15:D$=D$+CHR$(T%):NEXT: SPRITE$(0)=S$ 50 FOR T%=0 TO 7 52 PUT SPRITE T%, (T%*18, 100),,0 55 COLORSPRITE$(T%)=D$ 57 NEXT T% 58 REM HERE SET UNKNOWN MODE VIA VDP(n) 60 GOTO 60
Fascinating. Has nobody played around with this before? Almost can't be. Also very interesting to see what will happen on a 9958.
Fascinating. Has nobody played around with this before? Almost can't be
Yeah, I'm thinking the same thing.
It's almost impossible to imagine that no one has ever played with these bits.
Haha, I was thinking exactly the same.
If you take a look at how the modes are calculated, the screen mode for screen 3 is also bit M3 and M2 set together.
M1 enabled is text mode, who claims a multicolor screen '0', font size is not possible.
@PingPong: result is a white screen. I have had better results.
I am testing on a Sony HB-F700D , converted to 2+, so with V9958, and possibility to switch on 7Mhz.
Loading a screen '8' picture is possible with undocumented mode, same/similar result.
So, officially, I can only confirm results on V9958, not V9938
Counting on BIOS calls is a no-no, it blocks some bits/registers like 5FH, so correct results is obtained by addresses the VDP directly with assembly code.
And all this after almost 35 years, almost impossible!!
I was checking all VDP registers and logic and in order to find a way to get parallax scroling possible the hardware way, and not using 2 VDP's by using video in pin, like laser disc does.
To be honest, I almost would expect Arjen Z/Ultrasoft would have the solution/answer.
I have not understood if anyone had figured what this unknown bit configuration do.
can we havev a screen shot?
It this already emulated on openMSX?