dif4.dsp

载dsp上实现16dct和8dct的程序

.module/ram do_DIF4;        {*** 4 4-way DIFs ***}
.external tmp;
.entry DIF4;

DIF4:
        i0=^tmp;
        i1=^tmp+2;
        i2=^tmp;
        i3=^tmp+2;
        m2=3;

        ax1=dm(i3,m1);
        af=pass ax1, ax0=dm(i2,m1);
        ar=ax0+af, ax1=dm(i3,m2), my0=pm(i6,m6);
        ar=ax0-af, dm(i0,m1)=ar;
        mr=ar*my0(ss);
        af=pass ax1, ax0=dm(i2,m2);
        ar=ax0+af,                my1=pm(i6,m6);
        ar=ax0-af, dm(i0,m2)=ar;
        mr=ar*my1(ss), dm(i1,m1)=mr1;
        dm(i1,m2)=mr1;

        ax1=dm(i3,m1);
        af=pass ax1, ax0=dm(i2,m1);
        ar=ax0+af, ax1=dm(i3,m2);
        ar=ax0-af, dm(i0,m1)=ar;
        mr=ar*my0(ss);
        af=pass ax1, ax0=dm(i2,m2);
        ar=ax0+af;
        ar=ax0-af, dm(i0,m2)=ar;
        mr=ar*my1(ss), dm(i1,m1)=mr1;
        dm(i1,m2)=mr1;

        ax1=dm(i3,m1);
        af=pass ax1, ax0=dm(i2,m1);
        ar=ax0+af, ax1=dm(i3,m1);
        ar=ax0-af, dm(i0,m1)=ar;
        mr=ar*my0(ss);
        af=pass ax1, ax0=dm(i2,m0);
        ar=ax0+af;
        ar=ax0-af, dm(i0,m0)=ar;
        mr=ar*my1(ss), dm(i1,m1)=mr1;
        dm(i1,m1)=mr1;

        ax1=dm(i2,m1);
        af=pass ax1, ax0=dm(i3,m1);
        ar=ax0+af, ax1=dm(i2,m1);
        ar=ax0-af, dm(i1,m1)=ar;
        mr=ar*my0(ss);
        af=pass ax1, ax0=dm(i3,m1);
        ar=ax0+af;
        ar=ax0-af, dm(i1,m1)=ar;
        mr=ar*my1(ss), dm(i0,m1)=mr1;
        dm(i0,m1)=mr1;
        rts;                               {*** end 4 4-way DIFs ***}
.endmod;