V9958 documentation unclear

ページ 2/3
1 | | 3

By Accumulator

Champion (351)

Accumulator さんの画像

28-03-2023, 22:24

Just ff quick reply.. You are totally correct. The dot was placed incorrectly. 19.268 should be.
Maybe the incorrectness was because I was also thinking about 192 display lines in V9918, V9938 and V9958. And thinking if there was relation.
If it was a movie, Mr. blue was a person, he was always f#cked.. Some modes always less than Mr. Red or Mr. Green.

By gdx

Enlighted (6427)

gdx さんの画像

31-03-2023, 13:47

TomH wrote:

bits 0–7 of register 3 are A6 upwards of the colour table address. Bits 0–1 of register 4 are address bits 11 and 12 for the pattern generator table. This answer feels so simple that I must have misunderstood the question, or failed to perceive a relevant complexity? Apologies.

Not so simple! See here. TMS VDP are buggy (C6-C7 bits). That's why these bits are not documented. I don't know if Toshiba and Yamaha clones have this bug. V99x8 don't have it.

By Accumulator

Champion (351)

Accumulator さんの画像

31-03-2023, 15:45

Well, TMS VDP are in some cases absolutely better than V9938+.
For example polling bit 7 from status register 0 , TMS never misses the F flag!
V9938+ misses from time to time the F flag. (when setting and polling goes simultaneously)

By TomH

Champion (375)

TomH さんの画像

31-03-2023, 16:41

gdx wrote:
TomH wrote:

bits 0–7 of register 3 are A6 upwards of the colour table address. Bits 0–1 of register 4 are address bits 11 and 12 for the pattern generator table. This answer feels so simple that I must have misunderstood the question, or failed to perceive a relevant complexity? Apologies.

Not so simple! See here. TMS VDP are buggy (C6-C7 bits). That's why these bits are not documented. I don't know if Toshiba and Yamaha clones have this bug. V99x8 don't have it.

I'm failing to see the complexity; that looks like an incredibly lengthy way of saying: take the logical AND of the ephemeral VDP-generated address and the value written by the programmer. As above, I assume this is a failure of comprehension on my part.

By gdx

Enlighted (6427)

gdx さんの画像

31-03-2023, 16:50

I explained it this way because the logical AND doesn't seem right. I don't see how an AND could cause the effects I observed.

By PingPong

Enlighted (4155)

PingPong さんの画像

31-03-2023, 21:22

the TMS does miss the interrupt flag as the v99x8, absolutely.
It is well documented. The V99x8 unfortunately inherits this bug (kinda of race condition)

By PingPong

Enlighted (4155)

PingPong さんの画像

31-03-2023, 21:25

Absolutely better you say?
Like the sprite ghosting 'feature' that has to do with pattern or color table ? (never remind what is ), and that start to show when the TMS does warm?
Are you kidding? ;-)

By Accumulator

Champion (351)

Accumulator さんの画像

01-04-2023, 09:09

@PingPong,
If what you are saying is true, about F flag, I am really stunned...... VG8020 seems not have this error... or .. GGGRRRRRRR....

By gdx

Enlighted (6427)

gdx さんの画像

01-04-2023, 12:26

TomH wrote:

TMS VDP are in some cases absolutely better than V9938+.
For example polling bit 7 from status register 0 , TMS never misses the F flag!
V9938+ misses from time to time the F flag. (when setting and polling goes simultaneously)

This is not really a problem because in general it is not necessary to read this register outside the ISR. TMS VDP is not always better. It still requires waiting to access VRAM.

By DarkSchneider

Paragon (1030)

DarkSchneider さんの画像

01-04-2023, 15:02

What I miss on 99x8 is a 2KB pattern mode for G3, it cannot do it at least officially.
It is clear that Yamaha didn’t care about the tile mode at all and focused in bitmap mode.

ページ 2/3
1 | | 3