[FFmpeg-devel] [PATCH] swresample: misc improvements
Michael Niedermayer
michael at niedermayer.cc
Sat May 13 17:55:15 EEST 2023
On Sat, May 13, 2023 at 08:29:37AM +0200, Paul B Mahol wrote:
> On Sat, May 13, 2023 at 1:37 AM Michael Niedermayer <michael at niedermayer.cc>
> wrote:
>
> > On Thu, May 11, 2023 at 07:13:19PM +0200, Paul B Mahol wrote:
> > > Attached.
> > [...]
> > > @@ -33,64 +33,86 @@
> > >
> > >
> > > #define CONV_FUNC_NAME(dst_fmt, src_fmt) conv_ ## src_fmt ## _to_ ##
> > dst_fmt
> > > +#define CONVP_FUNC_NAME(dst_fmt, src_fmt) convp_ ## src_fmt ## _to_ ##
> > dst_fmt
> > >
> > > //FIXME rounding ?
> > > -#define CONV_FUNC(ofmt, otype, ifmt, expr)\
> > > +#define CONV_FUNC(ofmt, otype, ifmt, itype, expr)\
> > > + \
> > > static void CONV_FUNC_NAME(ofmt, ifmt)(uint8_t *po, const uint8_t *pi,
> > int is, int os, uint8_t *end)\
> > > {\
> > > uint8_t *end2 = end - 3*os;\
> > > while(po < end2){\
> > > + itype x = *(itype*)pi;\
> > > *(otype*)po = expr; pi += is; po += os;\
> > > + x = *(itype*)pi;\
> > > *(otype*)po = expr; pi += is; po += os;\
> > > + x = *(itype*)pi;\
> > > *(otype*)po = expr; pi += is; po += os;\
> > > + x = *(itype*)pi;\
> > > *(otype*)po = expr; pi += is; po += os;\
> > > }\
> > > while(po < end){\
> > > + itype x = *(itype*)pi;\
> > > *(otype*)po = expr; pi += is; po += os;\
> > > }\
> > > +}\
> > > +\
> > > +static void CONVP_FUNC_NAME(ofmt, ifmt)(uint8_t *ddst, const uint8_t
> > *ssrc, int len)\
> > > +{\
> > > + const itype *src = (const itype *)ssrc;\
> > > + otype *dst = (otype *)ddst;\
> > > + for (int n = 0; n < len; n++){\
> > > + itype x = src[n];\
> > > + dst[n] = expr;\
> > > + }\
> > > }
> > >
> > > //FIXME put things below under ifdefs so we do not waste space for
> > cases no codec will need
> > > -CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_U8 , *(const
> > uint8_t*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_U8 , (*(const
> > uint8_t*)pi - 0x80U)<<8)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_U8 , (*(const
> > uint8_t*)pi - 0x80U)<<24)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_U8 ,
> > (uint64_t)((*(const uint8_t*)pi - 0x80U))<<56)
> > > -CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_U8 , (*(const
> > uint8_t*)pi - 0x80)*(1.0f/ (1<<7)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_U8 , (*(const
> > uint8_t*)pi - 0x80)*(1.0 / (1<<7)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S16, (*(const
> > int16_t*)pi>>8) + 0x80)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S16, *(const
> > int16_t*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S16, *(const
> > int16_t*)pi * (1 << 16))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S16,
> > (uint64_t)(*(const int16_t*)pi)<<48)
> > > -CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S16, *(const
> > int16_t*)pi*(1.0f/ (1<<15)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S16, *(const
> > int16_t*)pi*(1.0 / (1<<15)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S32, (*(const
> > int32_t*)pi>>24) + 0x80)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S32, *(const
> > int32_t*)pi>>16)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S32, *(const
> > int32_t*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S32,
> > (uint64_t)(*(const int32_t*)pi)<<32)
> > > -CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S32, *(const
> > int32_t*)pi*(1.0f/ (1U<<31)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S32, *(const
> > int32_t*)pi*(1.0 / (1U<<31)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S64, (*(const
> > int64_t*)pi>>56) + 0x80)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S64, *(const
> > int64_t*)pi>>48)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S64, *(const
> > int64_t*)pi>>32)
> > > -CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S64, *(const
> > int64_t*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S64, *(const
> > int64_t*)pi*(1.0f/ (UINT64_C(1)<<63)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S64, *(const
> > int64_t*)pi*(1.0 / (UINT64_C(1)<<63)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_FLT,
> > av_clip_uint8( lrintf(*(const float*)pi * (1<<7)) + 0x80))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT,
> > av_clip_int16( lrintf(*(const float*)pi * (1<<15))))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT,
> > av_clipl_int32(llrintf(*(const float*)pi * (1U<<31))))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_FLT,
> > llrintf(*(const float*)pi * (UINT64_C(1)<<63)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_FLT, *(const
> > float*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_FLT, *(const
> > float*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_DBL,
> > av_clip_uint8( lrint(*(const double*)pi * (1<<7)) + 0x80))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL,
> > av_clip_int16( lrint(*(const double*)pi * (1<<15))))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL,
> > av_clipl_int32(llrint(*(const double*)pi * (1U<<31))))
> > > -CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_DBL, llrint(*(const
> > double*)pi * (UINT64_C(1)<<63)))
> > > -CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_DBL, *(const
> > double*)pi)
> > > -CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_DBL, *(const
> > double*)pi)
> > > -
> > > -#define FMT_PAIR_FUNC(out, in) [(out) + AV_SAMPLE_FMT_NB*(in)] =
> > CONV_FUNC_NAME(out, in)
> > > -
> > > -static conv_func_type * const
> > fmt_pair_to_conv_functions[AV_SAMPLE_FMT_NB*AV_SAMPLE_FMT_NB] = {
> > > +CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_U8 , uint8_t, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_U8 , uint8_t, (x -
> > 0x80U)<<8)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_U8 , uint8_t, (x -
> > 0x80U)<<24)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_U8 , uint8_t,
> > (uint64_t)(x - 0x80U)<<56)
> > > +CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_U8 , uint8_t, (x -
> > 0x80)*(1.0f/ (1<<7)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_U8 , uint8_t, (x -
> > 0x80)*(1.0 / (1<<7)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S16, int16_t,
> > (x>>8) + 0x80)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S16, int16_t, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S16, int16_t, x *
> > (1 << 16))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S16, int16_t,
> > (uint64_t)(x)<<48)
> > > +CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S16, int16_t,
> > x*(1.0f/ (1<<15)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S16, int16_t,
> > x*(1.0 / (1<<15)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S32, int32_t,
> > (x>>24) + 0x80)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S32, int32_t, x>>16)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S32, int32_t, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S32, int32_t,
> > (uint64_t)(x)<<32)
> > > +CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S32, int32_t,
> > x*(1.0f/ (1U<<31)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S32, int32_t,
> > x*(1.0 / (1U<<31)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S64, int64_t,
> > (x>>56) + 0x80)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S64, int64_t, x>>48)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S64, int64_t, x>>32)
> > > +CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S64, int64_t, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S64, int64_t,
> > x*(1.0f/ (UINT64_C(1)<<63)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S64, int64_t,
> > x*(1.0 / (UINT64_C(1)<<63)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_FLT, float,
> > av_clip_uint8( lrintf(x * (1<<7)) + 0x80))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float,
> > av_clip_int16( lrintf(x * (1<<15))))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float,
> > av_clipl_int32(llrintf(x * (1U<<31))))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_FLT, float,
> > llrintf(x * (UINT64_C(1)<<63)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_FLT, float, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_FLT, float, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_DBL, double,
> > av_clip_uint8( lrint(x * (1<<7)) + 0x80))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double,
> > av_clip_int16( lrint(x * (1<<15))))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double,
> > av_clipl_int32(llrint(x * (1U<<31))))
> > > +CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_DBL, double,
> > llrint(x * (UINT64_C(1)<<63)))
> > > +CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_DBL, double, x)
> > > +CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_DBL, double, x)
> >
> > i think the new cases are longer const, is that intended ?
> > (it would cast const to non const)
> >
>
> You mean I removed const from old macro?
yes
> Can fix that if that is the case.
thx
[...]
--
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
Complexity theory is the science of finding the exact solution to an
approximation. Benchmarking OTOH is finding an approximation of the exact
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20230513/51174fea/attachment.sig>
More information about the ffmpeg-devel
mailing list