nesdev.com

Hi tokumaru
As the topic states eject from tile position can you go into detail i'm aware when logical (lda _ObjectY,x..and %00000111) my object the low bits are how any pixels i'm into the block.How do I add these back is my question.Sorry if this is a newbie question but i'm also new to game coding.

I'm assuming a no on my question then,Tepples wanna help?i'm assuming add the low bits+1 back to the object postion.I apologise for my english, i'm Canadian and english is my third language .

Also sorry for the edits i had to look up French,German to English a few times

Could I see a bit of context, such as how you are computing this value whose low 3 bits are the penetration distance?

(Also run-ons.)

Yes of course this is checking the right side of my aabb

Didn't have time to check this out yet. I'm not home now, will try to take a look later.

No worries tokumaru i've Tepples helping me.

Also Tepples can you go into deep detail or link me as i understand most of you're explanations and find it difficult to understand other members.(I can understand tokumaru&Tepples can the rest of you guys not add into the conversation until they both reply?)

I'm not yet in a position to sit down and understand the entire code snippet. But I did find a few things that if you clarified them I could understand better.


What are the 21 and the 7?


Why not just AND #$F0?

Very unusual way to start a topic, since not all members read everything... luckily I read almost everything that isn't SNES, music, repros or mods!

When in doubt... DRAW! Here's a character (blue rectangle) running right into a wall, inside a tiny 3x3 metatile area:


Since the object moved left (something we know because the horizontal displacement was positive), we have to check all the blocks from the top right corner ($24, $04) to the bottom right corner ($24, $1F), to know whether they're solid or not. Dividing by 16, we have to look at all blocks between (2, 0) and (2, 1). If any of those blocks are solid, you'll know the object has hit a wall, which is the case here.

Now we have to figure out how many pixels to push the object back. Regardless of where the object's hospot is, and how you calculated the rightmost horizontal coordinate, you can easily tell just from looking that this coordinate is 5 pixels into the wall. Math confirms this: $24 AND $0f = $04, the fifth pixel of the metatile. You can now simply subtract 4 + 1 (since we count pixels from 0) from the object's coordinate and you can be sure that its right side will not be inside the wall anymore.

The math is a bit different for the left side, since objects enter the wall from the opposite side (column $0f):


Now we're looking at the corners ($0B, $04) and ($0B, $1F). $0B AND $0F = $0B, which is the 12th column of the metatile, but since the object is entering from the right, that doesn't mean it's 12 pixels in. The actual value you're looking for is 16 - 12 = 4, which is the same we'd get if the object was moving right, like in the previous example. In assembly, you can use a shortcut to find this number: just EOR it with $0F. $0B EOR $0F = $04, which is the number you're looking for. Now you can add that (plus 1) to eject the object from the wall.

Does this help?

I noticed now that you mentioned the mask %00000111, which makes me believe you're using 8x8-pixel blocks, not metatiles as in my example, right? If so, that's not a problem, the logic is always the same, you just have to keep the block size in mind when dividing and masking/negating bits.

[quote="tepples"]I'm not yet in a position to sit down and understand the entire code snippet. But I did find a few things that if you clarified them I could understand better.

[/code]
What are the 21 and the 7?
21 is my yposition AABB [i add 7 pixels for my xposition]
Sorry Tepples i want help but i wasen't clear






Why not just AND #$F0
I didn't think of this
I'm new to 6502

Thanks tokumaru & Tepples i need to read what tokumaru posted.
Now i get it thanks guys...I was having a problem with my object gravity he was to many pixels into the block.

I'm coding in z80 brother but the same rules apply

Tepples I'm used to this


ld b,(ix+constantYInt)
ld c,(ix+ContantYFraction)
ld l,(ix+objectYFraction)
ld h(ix+ObjectYInt)
add hl,bc
ld (ix+objectYFraction),l
ld h(ix+ObjectYInt),h
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
Is this the equivalent to as 6502 has direct memory access
charlievelocitycont:
lda _blobgravityf
sec
sbc _blobconstgravityf
sta _blobgravityf
lda _blobvelocity
sbc #0
sta _blobvelocity

Yes, that's how you typically subtract 16-bit values on 6502.