[FFmpeg-devel] [PATCH] avcodec/exr: preserve half-float NaN bits and add fate test

[Paul B Mahol]

On Sun, Dec 6, 2020 at 4:20 AM Mark Reid <mindmark at gmail.com> wrote:

> On Sun, Nov 22, 2020 at 8:32 PM <mindmark at gmail.com> wrote:
>
> > From: Mark Reid <mindmark at gmail.com>
> >
> > Hi,
> > This patch handles NaNs more like the offical implentation handles them,
> > preserving
> > the original bits.
> >
> >
> >
> https://github.com/AcademySoftwareFoundation/openexr/blob/RB-2.5/IlmBase/Half/toFloat.cpp#L111
> >
> > It also adds a fate test that is a 256x256 exr containing all possible
> > 16bit half-float values.
> >
> > Here is a link to download the fate test file, if someone could add it to
> > fate me
> >
> >
> https://www.dropbox.com/s/2q4jg8w489aunsf/rgb_scanline_zip_half_float_0x0_to_0xFFFF.exr
> >
> > ---
> >  libavcodec/exr.c                                    | 3 +--
> >  tests/fate/image.mak                                | 2 ++
> >  tests/ref/fate/exr-rgb-scanline-zip-half-0x0-0xFFFF | 6 ++++++
> >  3 files changed, 9 insertions(+), 2 deletions(-)
> >  create mode 100644 tests/ref/fate/exr-rgb-scanline-zip-half-0x0-0xFFFF
> >
> > diff --git a/libavcodec/exr.c b/libavcodec/exr.c
> > index d233dd43fb..6e6ce4275c 100644
> > --- a/libavcodec/exr.c
> > +++ b/libavcodec/exr.c
> > @@ -193,8 +193,7 @@ static union av_intfloat32 exr_half2float(uint16_t
> hf)
> >          // half-float NaNs will be converted to a single precision NaN
> >          // half-float Infs will be converted to a single precision Inf
> >          exp = FLOAT_MAX_BIASED_EXP;
> > -        if (mantissa)
> > -            mantissa = (1 << 23) - 1;    // set all bits to indicate a
> NaN
> > +        mantissa <<= 13; // preserve half-float NaN bits if set
> >      } else if (exp == 0x0) {
> >          // convert half-float zero/denorm to single precision value
> >          if (mantissa) {
> > diff --git a/tests/fate/image.mak b/tests/fate/image.mak
> > index 22072a62f1..c453f0f79c 100644
> > --- a/tests/fate/image.mak
> > +++ b/tests/fate/image.mak
> > @@ -317,6 +317,8 @@ fate-exr-rgb-scanline-half-zip-dw-outside: CMD =
> > framecrc -i $(TARGET_SAMPLES)/e
> >  FATE_EXR += fate-exr-rgb-tile-half-zip-dw-outside
> >  fate-exr-rgb-tile-half-zip-dw-outside: CMD = framecrc -i
> > $(TARGET_SAMPLES)/exr/rgb_tile_half_zip_dw_outside.exr -pix_fmt gbrpf32le
> >
> > +FATE_EXR += fate-exr-rgb-scanline-zip-half-0x0-0xFFFF
> > +fate-exr-rgb-scanline-zip-half-0x0-0xFFFF: CMD = framecrc -i
> > $(TARGET_SAMPLES)/exr/rgb_scanline_zip_half_float_0x0_to_0xFFFF.exr
> > -pix_fmt gbrpf32le
> >
> >  FATE_EXR-$(call DEMDEC, IMAGE2, EXR) += $(FATE_EXR)
> >
> > diff --git a/tests/ref/fate/exr-rgb-scanline-zip-half-0x0-0xFFFF
> > b/tests/ref/fate/exr-rgb-scanline-zip-half-0x0-0xFFFF
> > new file mode 100644
> > index 0000000000..b6201116fe
> > --- /dev/null
> > +++ b/tests/ref/fate/exr-rgb-scanline-zip-half-0x0-0xFFFF
> > @@ -0,0 +1,6 @@
> > +#tb 0: 1/25
> > +#media_type 0: video
> > +#codec_id 0: rawvideo
> > +#dimensions 0: 256x256
> > +#sar 0: 1/1
> > +0,          0,          0,        1,   786432, 0x1445e411
> > --
> > 2.29.2
> >
> >
> ping
>

Uploaded sample to server.
Will apply in the next 48h.


> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".