IRC log of #cubox of Wed 28 Aug 2013. All times are in CEST < Back to index

01:32 vitelmussi hola quien por aqui
13:55 wumpus _rmk_: yes, the stretch blit is pretty useless, it doesn't support any filtering, the filter blit does but it doesn't support clipping
13:57 wumpus I was first assuming stretch would do some kind of interpolation, so I tried to draw gradients using it, but just got huge pixels :-)
13:59 wumpus the filter blit is pretty neat though, apart from that it needs two passes
15:31 jnettlet I assume you guys are using either gco2D_SetStretchFactors, or gco2D_SetStretchRectFactors first.
15:41 wumpus yes
15:42 wumpus I compute the stretch factors according to ((src_width - 1) << 16) / (dst_width - 1) and vice versa for height, as it says in the docs
15:43 jnettlet and the destination surface has a format that supports Alpha blending.
15:44 wumpus yes, simply A8R8G8B8
15:44 wumpus is it supposed to do a smooth scaling then?
15:50 jnettlet nope. If you have a yuv format or need smoothscale then you have to do a filterblit
15:51 jnettlet yuv src format that is
15:51 wumpus ok, yes that's what I thought
15:59 jnettlet and you can also set the scaling quality to use.
16:01 _rmk_ err, why does the destination need to support alpha blending, that sounds soo obtuse.
16:01 _rmk_ and where does the alpha come from with a yuv source...
16:02 jnettlet if you have a yuv source then you can only use a filterblit, in which case you can have a non-alpha destination
16:03 jnettlet the alpha blending destination is a limit on the stretch blit
16:06 jnettlet I believe it has to do with the limitation of only being able to have either the ROP or alpha blending engine on for an operation
16:26 dbsx I found the spi flash bug in 3.11!
16:28 wumpus woohoo
16:28 dbsx It took a while
16:28 dbsx sleep now
16:31 _rmk_ congrats
16:36 dbsx Not sure how to fix it
16:37 dbsx ttfn
20:25 dv_ jnettlet_: you ran the vmeta tests on the XO-4, and saw the flickering, do I remember this correctly?