Scale bug ?

[KanedaFr]

Hi, 

I read on some topics on AA what there is a "known" bug on scaling.

Since emulators are able to render this bug, could someone explain me this bug ?

 

I'm trying to stack 2x scaled objects on screen but I have +1, +2 or +3 to ypos to avoid empty line.

So I'd like to know if I could use scaling or just forget about it and scale bitmap data outside of the Jaguar...

 

Thanks

[42bs]

Scaling works perfectly. The Jaguar boot is a scaled object. JagNiccc2000 uses a few of them.

 

Scaled and mirrored.

 

Edited 4 hours ago by 42bs

[KanedaFr]

I have exactly the same issue that in this thread -> 

 

 

 

Is it JagStudio related ?

 

Edited 4 hours ago by KanedaFr

[42bs]

16 minutes ago, KanedaFr said:

Is it JagStudio related ?

 

I'd say only with respect to you using JS and did something wrong

 

Maybe show your code and/or a picture of the result.

 

[KanedaFr]

It's hard to show code w/o filling a full page since I'll need to share rapinit.s 

 

I'll try to isolate the problem on one sample project and share the zip + binary

[42bs]

Just now, KanedaFr said:

It's hard to show code w/o filling a full page since I'll need to share rapinit.s 

 

I'll try to isolate the problem on one sample project and share the zip + binary

The part of the object list should be enough. But JS experts might need only a look into your rapinit.S to pin-point the problem.

[+CyranoJ]

When using a scaled object set the bitmap pointer address to the address of the bitmap +1 line worth of data. This is fully documented in the... documentation.

 

[edit] or it could be -1 line

 

Too tired to go check, but its one of those.

[KanedaFr]

I use a dynamic list : 20 instances of the same object, scaled x2 (64), edited by code

so it's not that obvious to debug just with a rapinit.s

 

Something like:

object2.y_ =  240-(scale*object2.height)

object1.y_ =  240-(scale*object2.height)-(scale*object1.height)

[KanedaFr]

3 minutes ago, CyranoJ said:

When using a scaled object set the bitmap pointer address to the address of the bitmap +1 line worth of data. This is fully documented in the... documentation.

Oh no!!!

not again...:(((

What the hell did I miss...?!

 

I only saw

The object must be on a 32 byte boundary in 64 bit RAM.

which I was trying to debug

 

[KanedaFr]

ARG !!!!!!

 

Please note that for scaled bitmap objects, HEIGHT
should actually be the bitmap height -1.

 

 

So, for exemple, 16 if not scaled, but 15 if scaled ?

and what about framez ? and vbox ?

Edited 3 hours ago by KanedaFr

[Seedy1812]

If you are passing the same values in for both scaled sprites ( different x and y - on screen ) then they should be drawn the same.
The pictures you shown from another thread has a shift in the x whilst  you talk about the y.

Chances are its the SBMO Remainder