hello,
Am Samstag, 27. Mai 2006 21:03 schrieb Miller Puckette:
I used to think the same thing, that floating-point computation would never be feasible on video... but Mathieu has demonstrated that, if you schedule the computations carefully, you can do blindingly fast video crunching in floating-point. Trouble with it is simply that it takes much more storage space, so you end up using a lot of memory bandwidth (a read or write to external memory costs the same as tens or hundreds of CPU cycles these days).
uhm.... blindingly fast? i dont believe that .... unless you use sse/sse2 for the floating point ops. and even then you would be slower. the alu for integer math is much quicker and has a lower latency as the fpu. also, on a p4 for example, you have actually 2 alu's for integer stuff ....
to put it in other words: how comes that demuxing/converting a lot of audio channels (for example from a hammerfall) takes a considerable amount of cpu, if using floats would be blindingly fast? and even on 16 channels of audio you dont have that much data as with a simple full pal-sized video ....
But anyway, I'll probably make it conditionally compilable in other formats so we can just benchmark it and find out for sure.
why conditional? why not runtime-selectable, per branch/subpatch? i mean, there are quite some formats actively in use, like rgb(a), yuv, hsb ...
cheers Miller
greets,
chris