[FFmpeg-devel] [PATCH 2/2] libavcodec/jpeg2000dec.c: ROI marker support
Gautam Ramakrishnan
gautamramk at gmail.com
Wed Apr 22 06:14:10 EEST 2020
On Wed, Apr 22, 2020 at 3:04 AM Michael Niedermayer
<michael at niedermayer.cc> wrote:
>
> On Tue, Apr 21, 2020 at 01:07:39AM +0530, gautamramk at gmail.com wrote:
> > From: Gautam Ramakrishnan <gautamramk at gmail.com>
> >
> > This patch adds support for decoding images
> > with a Region of Interest. Allows decoding
> > samples such as p0_03.j2k. This patch should
> > fix ticket #4681.
> > ---
> > libavcodec/jpeg2000.h | 1 +
> > libavcodec/jpeg2000dec.c | 57 +++++++++++++++++++++++++++++++++++-----
> > 2 files changed, 51 insertions(+), 7 deletions(-)
> >
> > diff --git a/libavcodec/jpeg2000.h b/libavcodec/jpeg2000.h
> > index 7b78c0193e..0f82716981 100644
> > --- a/libavcodec/jpeg2000.h
> > +++ b/libavcodec/jpeg2000.h
> > @@ -210,6 +210,7 @@ typedef struct Jpeg2000Component {
> > int *i_data;
> > int coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}} -- can be reduced with lowres option
> > int coord_o[2][2]; // border coordinates {{x0, x1}, {y0, y1}} -- original values from jpeg2000 headers
> > + uint8_t roi_shift; // ROI scaling value for the component
> > } Jpeg2000Component;
> >
> > /* misc tools */
> > diff --git a/libavcodec/jpeg2000dec.c b/libavcodec/jpeg2000dec.c
> > index 5a7d9e7882..da19345ee7 100644
> > --- a/libavcodec/jpeg2000dec.c
> > +++ b/libavcodec/jpeg2000dec.c
> > @@ -117,6 +117,7 @@ typedef struct Jpeg2000DecoderContext {
> > Jpeg2000CodingStyle codsty[4];
> > Jpeg2000QuantStyle qntsty[4];
> > Jpeg2000POC poc;
>
> > + uint8_t roi_shift[4];
> >
> > int bit_index;
> >
> > @@ -598,6 +599,29 @@ static int get_coc(Jpeg2000DecoderContext *s, Jpeg2000CodingStyle *c,
> > return 0;
> > }
> >
> > +static int get_rgn(Jpeg2000DecoderContext *s, int n)
> > +{
> > + uint16_t compno;
> > + compno = (s->ncomponents < 257)? bytestream2_get_byte(&s->g):
> > + bytestream2_get_be16u(&s->g);
> > + if (bytestream2_get_byte(&s->g)) {
> > + av_log(s->avctx, AV_LOG_ERROR, "Invalid RGN header.\n");
> > + return AVERROR_INVALIDDATA; // SRgn field value is 0
> > + }
> > + // SPrgn field
> > + if (compno < s->ncomponents) {
> > + if (s->curtileno == -1)
> > + s->roi_shift[compno] = bytestream2_get_byte(&s->g);
>
> theres a check for s->ncomponents < 257 implying that if this check is not a
> dead check ncomponents can be bigger and then compno is just checked by that
> before being used to index into a 4 entry array (roi_shift)
>
> something in here is not entirely correct
For our decoder, it is a dead check as we are supposed to return an error
before this marker reaches if the number of components is greater than 4.
However, I have still kept this check keeping in mind support for more than
4 components in the future.
>
>
>
> > + else {
> > + if (s->tile[s->curtileno].tp_idx != 0)
> > + return AVERROR_INVALIDDATA; // marker occurs only in first tile part of tile
> > + s->tile[s->curtileno].comp[compno].roi_shift = bytestream2_get_byte(&s->g);
> > + }
> > + return 0;
> > + }
> > + return AVERROR_INVALIDDATA;
> > +}
> > +
> > /* Get common part for QCD and QCC segments. */
> > static int get_qcx(Jpeg2000DecoderContext *s, int n, Jpeg2000QuantStyle *q)
> > {
> > @@ -947,6 +971,9 @@ static int init_tile(Jpeg2000DecoderContext *s, int tileno)
> > comp->coord[1][0] = ff_jpeg2000_ceildivpow2(comp->coord_o[1][0], s->reduction_factor);
> > comp->coord[1][1] = ff_jpeg2000_ceildivpow2(comp->coord_o[1][1], s->reduction_factor);
> >
> > + if (!comp->roi_shift)
> > + comp->roi_shift = s->roi_shift[compno];
> > +
> > if (ret = ff_jpeg2000_init_component(comp, codsty, qntsty,
> > s->cbps[compno], s->cdx[compno],
> > s->cdy[compno], s->avctx))
> > @@ -1615,9 +1642,9 @@ static void decode_clnpass(Jpeg2000DecoderContext *s, Jpeg2000T1Context *t1,
> >
> > static int decode_cblk(Jpeg2000DecoderContext *s, Jpeg2000CodingStyle *codsty,
> > Jpeg2000T1Context *t1, Jpeg2000Cblk *cblk,
> > - int width, int height, int bandpos)
> > + int width, int height, int bandpos, uint8_t roi_shift)
> > {
> > - int passno = cblk->npasses, pass_t = 2, bpno = cblk->nonzerobits - 1;
> > + int passno = cblk->npasses, pass_t = 2, bpno = cblk->nonzerobits - 1 + roi_shift;
> > int pass_cnt = 0;
> > int vert_causal_ctx_csty_symbol = codsty->cblk_style & JPEG2000_CBLK_VSC;
> > int term_cnt = 0;
>
> > @@ -1691,6 +1718,19 @@ static int decode_cblk(Jpeg2000DecoderContext *s, Jpeg2000CodingStyle *codsty,
> > return 1;
> > }
> >
> > +static inline int roi_shift_param(Jpeg2000Component *comp,
> > + int quan_parameter)
> > +{
> > + uint8_t roi_shift;
> > + int val;
> > + roi_shift = comp->roi_shift;
> > + val = (quan_parameter < 0)?-quan_parameter:quan_parameter;
> > +
> > + if (val > (1 << roi_shift))
> > + return (quan_parameter < 0)?-(val >> roi_shift):(val >> roi_shift);
> > + return quan_parameter;
> > +}
> > +
> > /* TODO: Verify dequantization for lossless case
> > * comp->data can be float or int
> > * band->stepsize can be float or int
> > @@ -1708,7 +1748,7 @@ static void dequantization_float(int x, int y, Jpeg2000Cblk *cblk,
> > float *datap = &comp->f_data[(comp->coord[0][1] - comp->coord[0][0]) * (y + j) + x];
> > int *src = t1->data + j*t1->stride;
> > for (i = 0; i < w; ++i)
> > - datap[i] = src[i] * band->f_stepsize;
> > + datap[i] = roi_shift_param(comp, src[i]) * band->f_stepsize;
> > }
> > }
> >
> > @@ -1724,11 +1764,11 @@ static void dequantization_int(int x, int y, Jpeg2000Cblk *cblk,
> > int *src = t1->data + j*t1->stride;
> > if (band->i_stepsize == 32768) {
> > for (i = 0; i < w; ++i)
> > - datap[i] = src[i] / 2;
> > + datap[i] = roi_shift_param(comp, src[i]) / 2;
> > } else {
> > // This should be VERY uncommon
> > for (i = 0; i < w; ++i)
> > - datap[i] = (src[i] * (int64_t)band->i_stepsize) / 65536;
> > + datap[i] = (roi_shift_param(comp, src[i]) * (int64_t)band->i_stepsize) / 65536;
> > }
> > }
> > }
> > @@ -1743,7 +1783,7 @@ static void dequantization_int_97(int x, int y, Jpeg2000Cblk *cblk,
> > int32_t *datap = &comp->i_data[(comp->coord[0][1] - comp->coord[0][0]) * (y + j) + x];
> > int *src = t1->data + j*t1->stride;
> > for (i = 0; i < w; ++i)
> > - datap[i] = (src[i] * (int64_t)band->i_stepsize + (1<<15)) >> 16;
> > + datap[i] = (roi_shift_param(comp, src[i]) * (int64_t)band->i_stepsize + (1<<15)) >> 16;
> > }
> > }
>
> the dequantization functions matter speedwise adding the whole
> roi_shift_param() stuff repeatly to each sample would slow the code down
> more than neccessary
>
I shall make the shifting conditional now, so the shift is performed beforehand.
> thx
>
> [...]
>
> --
> Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>
> I am the wisest man alive, for I know one thing, and that is that I know
> nothing. -- Socrates
> _______________________________________________
> 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".
--
-------------
Gautam |
More information about the ffmpeg-devel
mailing list