mawen苣菊,我也遇到了加入HQDering后FPS速度大增的情况,由1.8XX暴增至2.3XXFPS,求解。
下面POST脚本,MPP是刚开始用,望批评指教。
机器:E3-1230v2 8G RAM GTS250
AVS(读入滤镜的命令省略):
MP_Pipeline("""
SetMemoryMax(500)
DGDecode_mpeg2source("D:\Movies\动漫\VIDEO_TS\VTS_01_1.d2v")
tfm(order=1).tdecimate(mode=1)
mt_merge(Deblock(quant=30, aOffset=0, bOffset=0).Deblock(quant=40, aOffset=0, bOffset=0), last, tcanny.mt_expand, luma=false)
### prefetch: 16, 8
### lock threads to cores
### ###
SetMemoryMax(800)
nnedi3_resize16(1280, 720, lsb_in=false, lsb=true, curve="601")
### prefetch: 16, 8
### lock threads to cores
### ###
SetMemoryMax(1000)
src16=last
nr=src16.SMDegrain(tr=2, thSAD=400, RefineMotion=True,contrasharp=False, interlaced=False, plane=4, prefilter=2, chroma=True, lsb_in=true, lsb=true, lsb_out=True, Show=False)
### export clip: nr,src16
### prefetch: 16, 8
### lock threads to cores
### ###
SetMemoryMax(800)
db=src16.GradFun3(smode=1, thr=0.61, elast=1.50, ampn=0, radius=12, ampo=0.9, dthr=0.011, lsb=true, lsb_in=true, mask=0).F3kdb(range=15,Y=64, Cb=64, Cr=64, grainY=0, grainC=0, input_mode=1, output_depth=16, output_mode=1)
### export clip: nr,db,src16
### prefetch: 16, 8
### lock threads to cores
### ###
SetMemoryMax(800)
nrs=nr.CSmod16(src16, lsb_in=true, lsb=true, chroma=true, preblur=2, edgemode=2) #这个锐化还原一流,但不太懂用。AA后用锯齿也回去了。。。。
### export clip: src16,nrs,db
### prefetch: 16, 8
### lock threads to cores
### ###
SetMemoryMax(1000)
ALMerge(db, nrs, src16, U=3, V=3)
HQDeringmod(mrad=2, mthr=60, minp=1, sharp=True, Y=3, U=2, V=2, lsb_in=true, lsb=true)
Down10(10, stack=false)
### prefetch: 16, 8
### lock threads to cores
""")