There is a special case I missed, its simpler to just store
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
We need to allocate space for 64 coefficients per block;
24dbc4c2e8 wanted to
perform the calculation 64*sizeof(MJpegHuffmanCode)
at compile time, yet in the end did it in a way that
made it allocate 64 times as much memory as needed.
Reported-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The kVTVideoDecoderReferenceMissingErr constant was only added
in the macOS 12 and iOS 15 SDKs. Use a hardcoded value instead
of the named constant, to fix building with older SDKs
after c6214b0d69.
Signed-off-by: Martin Storsjö <martin@martin.st>
For all encoders and all decoders except MPEG-4 the unquantize
functions to use don't change at all and therefore needn't be
kept in the context. So discard them after setting them;
for MPEG-4, the functions get assigned on a per-frame basis.
Decoders not using any unquantize functions (H.261, MPEG-1/2)
as well as decoders that only call ff_mpv_reconstruct_mb()
through error resilience (RV30/40, the VC-1 family) don't have
the remaining pointers set at all.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
block_last_index and nCoeffs is an optimization designed
to avoid processing unnecessarily many coefficients; yet
it would be legal to always process all coefficients
(all coefficients beyond nCoeffs are zero anyway and
zeros are always unquantized to zeros). Therefore
one does not need a scalar tail.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Don't do it depending upon dct_algo, as this is not supposed
to influence the unquantize function selection.
(This check originated in 05c4072b45
where it was used for the dct_quantize function only.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This affected the WMV1/2 encoders (but not when running FATE because
the encoding part uses the fastint dct, so this code isn't used
then anyway).
It did not affect anything else, because a) only WMV1/2 use different
scantables, b) ff_msmpeg4_decode_block() (and therefore the WMV1
decoder) already unquantize inter macroblocks as they are parsed
c) the WMV2 decoder does not use the unquantize functions for inter
macroblocks at all.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is in preparation for only keeping the actually used
unquantize functions in MpegEncContext.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Use spell out what me_cmp_func means.
Avoids inclusions in the H.264 decoder as well as all
mpegvideo decoders.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Only set them after mb_x and mb_y are known which happens
only after the call to ff_h261_reorder_mb_index().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Reset them in ff_me_init_pic(). It is the appropriate place for it
and allows to remove resetting code from multiple places
(e.g. mpegvideo_enc.c resetted it in two different places
(one for the main slice context, one for all the others)).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Also avoid casts and parentheses.
(This is only possible now because ff_update_duplicate_context()
no longer touches the PutBitContext.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
These pointers point to static tables which must not be modified
by anyone after they have been initialized. So constify the pointees.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This happens currently for the non-main slice contexts.
But these variables get reset at the start of encode_thread()
anyway for all slices, so this is unnecessary.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Many of the fields of MpegEncContext (which is also used by decoders)
are actually only used by encoders. Therefore this commit adds
a new encoder-only structure and moves all of the encoder-only
fields to it except for those which require more explicit
synchronisation between the main slice context and the other
slice contexts. This synchronisation is currently mainly provided
by ff_update_thread_context() which simply copies most of
the main slice context over the other slice contexts. Fields
which are moved to the new MPVEncContext no longer participate
in this (which is desired, because it is horrible and for the
fields b) below wasteful) which means that some fields can only
be moved when explicit synchronisation code is added in later commits.
More explicitly, this commit moves the following fields:
a) Fields not copied by ff_update_duplicate_context():
dct_error_sum and dct_count; the former does not need synchronisation,
the latter is synchronised in merge_context_after_encode().
b) Fields which do not change after initialisation (these fields
could also be put into MPVMainEncContext at the cost of
an indirection to access them): lambda_table, adaptive_quant,
{luma,chroma}_elim_threshold, new_pic, fdsp, mpvencdsp, pdsp,
{p,b_forw,b_back,b_bidir_forw,b_bidir_back,b_direct,b_field}_mv_table,
[pb]_field_select_table, mb_{type,var,mean}, mc_mb_var, {min,max}_qcoeff,
{inter,intra}_quant_bias, ac_esc_length, the *_vlc_length fields,
the q_{intra,inter,chroma_intra}_matrix{,16}, dct_offset, mb_info,
mjpeg_ctx, rtp_mode, rtp_payload_size, encode_mb, all function
pointers, mpv_flags, quantizer_noise_shaping,
frame_reconstruction_bitfield, error_rate and intra_penalty.
c) Fields which are already (re)set explicitly: The PutBitContexts
pb, tex_pb, pb2; dquant, skipdct, encoding_error, the statistics
fields {mv,i_tex,p_tex,misc,last}_bits and i_count; last_mv_dir,
esc_pos (reset when writing the header).
d) Fields which are only used by encoders not supporting slice
threading for which synchronisation doesn't matter: esc3_level_length
and the remaining mb_info fields.
e) coded_score: This field is only really used when FF_MPV_FLAG_CBP_RD
is set (which implies trellis) and even then it is only used for
non-intra blocks. For these blocks dct_quantize_trellis_c() either
sets coded_score[n] or returns a last_non_zero value of -1
in which case coded_score will be reset in encode_mb_internal().
Therefore no old values are ever used.
The MotionEstContext has not been moved yet.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This avoids relying on ff_update_duplicate_context() to copy
these fields to the slice contexts.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
