[FFmpeg-devel] [libav-devel] [PATCH 1/2] avcodec: Add Cineform HD Decoder
Janne Grunau
janne-libav at jannau.net
Sun Jan 10 23:50:58 CET 2016
On 2016-01-10 00:28:49 +0000, Kieran Kunhya wrote:
> ---
> libavcodec/Makefile | 1 +
> libavcodec/allcodecs.c | 1 +
> libavcodec/avcodec.h | 1 +
> libavcodec/cfhd.c | 565 ++++++++++++++++++++++++++++++++++++++++++++++++
> libavcodec/cfhd.h | 99 +++++++++
> libavcodec/cfhddata.c | 470 ++++++++++++++++++++++++++++++++++++++++
> libavcodec/codec_desc.c | 6 +
> 7 files changed, 1143 insertions(+)
> create mode 100644 libavcodec/cfhd.c
> create mode 100644 libavcodec/cfhd.h
> create mode 100644 libavcodec/cfhddata.c
>
> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> index b9ffdb9..c331a4f 100644
> --- a/libavcodec/Makefile
> +++ b/libavcodec/Makefile
> @@ -210,6 +210,7 @@ OBJS-$(CONFIG_CAVS_DECODER) += cavs.o cavsdec.o cavsdsp.o \
> OBJS-$(CONFIG_CCAPTION_DECODER) += ccaption_dec.o
> OBJS-$(CONFIG_CDGRAPHICS_DECODER) += cdgraphics.o
> OBJS-$(CONFIG_CDXL_DECODER) += cdxl.o
> +OBJS-$(CONFIG_CFHD_DECODER) += cfhd.o cfhddata.o
> OBJS-$(CONFIG_CINEPAK_DECODER) += cinepak.o
> OBJS-$(CONFIG_CINEPAK_ENCODER) += cinepakenc.o elbg.o
> OBJS-$(CONFIG_CLJR_DECODER) += cljrdec.o
> diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
> index 2128546..1d92c8b 100644
> --- a/libavcodec/allcodecs.c
> +++ b/libavcodec/allcodecs.c
> @@ -147,6 +147,7 @@ void avcodec_register_all(void)
> REGISTER_DECODER(CAVS, cavs);
> REGISTER_DECODER(CDGRAPHICS, cdgraphics);
> REGISTER_DECODER(CDXL, cdxl);
> + REGISTER_DECODER(CFHD, cfhd);
> REGISTER_ENCDEC (CINEPAK, cinepak);
> REGISTER_ENCDEC (CLJR, cljr);
> REGISTER_DECODER(CLLC, cllc);
> diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
> index f365775..b958a6c 100644
> --- a/libavcodec/avcodec.h
> +++ b/libavcodec/avcodec.h
> @@ -315,6 +315,7 @@ enum AVCodecID {
> AV_CODEC_ID_SMVJPEG,
> AV_CODEC_ID_APNG,
> AV_CODEC_ID_DAALA,
> + AV_CODEC_ID_CFHD,
>
> /* various PCM "codecs" */
> AV_CODEC_ID_FIRST_AUDIO = 0x10000, ///< A dummy id pointing at the start of audio codecs
> diff --git a/libavcodec/cfhd.c b/libavcodec/cfhd.c
> new file mode 100644
> index 0000000..dc36889
> --- /dev/null
> +++ b/libavcodec/cfhd.c
> @@ -0,0 +1,565 @@
> +/*
> + * Copyright (c) 2015 Kieran Kunhya
> + *
> + * This file is part of FFmpeg.
> + *
> + * FFmpeg is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Lesser General Public
> + * License as published by the Free Software Foundation; either
> + * version 2.1 of the License, or (at your option) any later version.
> + *
> + * FFmpeg is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + * Lesser General Public License for more details.
> + *
> + * You should have received a copy of the GNU Lesser General Public
> + * License along with FFmpeg; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
> + */
> +
> +/**
> + * @file
> + * CFHD Video Decoder
> + */
> +
> +#include "libavutil/buffer.h"
> +#include "libavutil/common.h"
> +#include "libavutil/intreadwrite.h"
> +#include "libavutil/imgutils.h"
> +#include "libavutil/opt.h"
> +
> +#include "avcodec.h"
> +#include "bswapdsp.h"
> +#include "internal.h"
> +#include "bytestream.h"
> +#include "cfhd.h"
> +
> +static av_cold int cfhd_decode_init(AVCodecContext *avctx)
> +{
> + CFHDContext *s = avctx->priv_data;
> +
> + avctx->pix_fmt = AV_PIX_FMT_YUV422P10;
> + avctx->bits_per_raw_sample = 10;
> + s->avctx = avctx;
> +
> + return ff_cfhd_init_vlcs(s);
> +}
> +
> +static void init_plane_defaults(CFHDContext *s)
> +{
> + s->subband_num = 0;
> + s->level = 0;
> + s->subband_num_actual = 0;
> +}
> +
> +static void init_frame_defaults(CFHDContext *s)
> +{
> + s->bpc = 10;
> + s->channel_cnt = 4;
> + s->subband_cnt = 10;
> + s->channel_num = 0;
> + s->lowpass_precision = 16;
> + s->quantisation = 1;
> + s->wavelet_depth = 3;
> + s->pshift = 1;
> + s->codebook = 0;
> + init_plane_defaults(s);
> +}
> +
> +static inline int dequant_and_decompand(int level, int quantisation)
> +{
> + int abslevel = abs(level);
> + return (abslevel + ((768 * abslevel * abslevel * abslevel) / (255 * 255 * 255))) * FFSIGN(level) * quantisation;
> +}
> +
> +static inline void filter(int16_t *output, ptrdiff_t out_stride, int16_t *low, ptrdiff_t low_stride,
> + int16_t *high, ptrdiff_t high_stride, int len)
> +{
> + int32_t tmp;
> +
> + int i;
> + for (i = 0; i < len; i++) {
> + if( i == 0 )
> + {
> + tmp = (11*low[0*low_stride] - 4*low[1*low_stride] + low[2*low_stride] + 4) >> 3;
> + output[(2*i+0)*out_stride] = (tmp + high[0*high_stride]) >> 1;
> + tmp = ( 5*low[0*low_stride] + 4*low[1*low_stride] - low[2*low_stride] + 4) >> 3;
> + output[(2*i+1)*out_stride] = (tmp - high[0*high_stride]) >> 1;
> + }
> + else if( i == len-1 )
> + {
> + tmp = ( 5*low[i*low_stride] + 4*low[(i-1)*low_stride] - low[(i-2)*low_stride] + 4) >> 3;
> + output[(2*i+0)*out_stride] = (tmp + high[i*high_stride]) >> 1;
> + tmp = (11*low[i*low_stride] - 4*low[(i-1)*low_stride] + low[(i-2)*low_stride] + 4) >> 3;
> + output[(2*i+1)*out_stride] = (tmp - high[i*high_stride]) >> 1;
> + }
> + else
> + {
> + tmp = (low[(i-1)*low_stride] - low[(i+1)*low_stride] + 4) >> 3;
> + output[(2*i+0)*out_stride] = (tmp + low[i*low_stride] + high[i*high_stride]) >> 1;
> + tmp = (low[(i+1)*low_stride] - low[(i-1)*low_stride] + 4) >> 3;
> + output[(2*i+1)*out_stride] = (tmp + low[i*low_stride] - high[i*high_stride]) >> 1;
> + }
> + }
> +}
> +
> +static void horiz_filter(int16_t *output, int16_t *low, int16_t *high, int width)
> +{
> + filter(output, 1, low, 1, high, 1, width);
> +}
> +
> +static void vert_filter(int16_t *output, int out_stride, int16_t *low, int low_stride,
> + int16_t *high, int high_stride, int len)
> +{
> + filter(output, out_stride, low, low_stride, high, high_stride, len);
> +}
> +
> +static int cfhd_decode(AVCodecContext *avctx, void *data, int *got_frame,
> + AVPacket *avpkt)
> +{
> + CFHDContext *s = avctx->priv_data;
> + GetByteContext gb;
> + AVFrame *pic = data;
> + int ret = 0, i, j;
> + int16_t *plane[3] = {NULL};
> + int16_t *tmp[3] = {NULL};
> + int16_t *subband[3][10] = {{0}};
> + int16_t *l_h[3][8];
> + int16_t *coeff_data;
> +
> + avcodec_get_chroma_sub_sample(avctx->pix_fmt, &s->chroma_x_shift, &s->chroma_y_shift);
> +
> + for (i = 0; i < 3; i++) {
> + int width = i ? avctx->width >> s->chroma_x_shift : avctx->width;
> + int height = i ? avctx->height >> s->chroma_y_shift : avctx->height;
> + int stride = FFALIGN(width / 8, 8) * 8;
FFALIGN(width, 64) looks clearer to me
> + int w8, h8, w4, h4, w2, h2;
> + height = FFALIGN(height / 8, 2) * 8;
same here
> + s->plane[i].width = width;
> + s->plane[i].height = height;
> + s->plane[i].stride = stride;
> +
> + w8 = FFALIGN(s->plane[i].width / 8, 8);
> + h8 = FFALIGN(s->plane[i].height / 8, 2);
s->plane[i].height is already aligned to 16 so the FFALIGN is not
needed, the same argument could be made for w8 with stride but that's
less compelling. Probably ok if both stay as is as defensive measure
against code changes.
> + w4 = w8 * 2;
> + h4 = h8 * 2;
> + w2 = w4 * 2;
> + h2 = h4 * 2;
> +
> + plane[i] = av_malloc(height * stride * sizeof(*plane[i]));
> + tmp[i] = av_malloc(height * stride * sizeof(*tmp[i]));
> + if (!plane[i] || !tmp[i]) {
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> +
> + subband[i][0] = plane[i];
> + subband[i][1] = plane[i] + 2 * w8 * h8;
> + subband[i][2] = plane[i] + 1 * w8 * h8;
> + subband[i][3] = plane[i] + 3 * w8 * h8;
> + subband[i][4] = plane[i] + 2 * w4 * h4;
> + subband[i][5] = plane[i] + 1 * w4 * h4;
> + subband[i][6] = plane[i] + 3 * w4 * h4;
> + subband[i][7] = plane[i] + 2 * w2 * h2;
> + subband[i][8] = plane[i] + 1 * w2 * h2;
> + subband[i][9] = plane[i] + 3 * w2 * h2;
> +
> + l_h[i][0] = tmp[i];
> + l_h[i][1] = tmp[i] + 2 * w8 * h8;
> + //l_h[i][2] = ll2;
> + l_h[i][3] = tmp[i];
> + l_h[i][4] = tmp[i] + 2 * w4 * h4;
> + //l_h[i][5] = ll1;
> + l_h[i][6] = tmp[i];
> + l_h[i][7] = tmp[i] + 2 * w2 * h2;
> + }
> +
> + init_frame_defaults(s);
> +
> + if ((ret = ff_get_buffer(avctx, pic, 0)) < 0)
> + return ret;
> +
> + bytestream2_init(&gb, avpkt->data, avpkt->size);
> +
> + while (bytestream2_tell(&gb) < avpkt->size) {
> + /* Bit weird but implement the tag parsing as the spec says */
> + uint16_t tagu = bytestream2_get_be16(&gb);
> + int16_t tag = (int16_t)tagu;
> + int8_t tag8 = (int8_t)(tagu >> 8);
> + uint16_t abstag = abs(tag);
> + int8_t abs_tag8 = abs(tag8);
> + uint16_t data = bytestream2_get_be16(&gb);
> + if (abs_tag8 >= 0x60 && abs_tag8 <= 0x6f) {
> + av_log(avctx, AV_LOG_DEBUG, "large len %x \n", ((tagu & 0xff) << 16) | data);
> + } else if (tag == 20) {
> + av_log(avctx, AV_LOG_DEBUG, "Width %"PRIu16" \n", data);
> + avctx->width = data;
> + } else if (tag == 21) {
> + av_log(avctx, AV_LOG_DEBUG, "Height %"PRIu16" \n", data);
> + avctx->height = data;
> + } else if (tag == 101) {
> + av_log(avctx, AV_LOG_DEBUG, "Bits per component: %"PRIu16" \n", data);
> + s->bpc = data;
> + } else if (tag == 12) {
> + av_log(avctx, AV_LOG_DEBUG, "Channel Count: %"PRIu16" \n", data);
> + s->channel_cnt = data;
> + if (data != 3) {
> + av_log(avctx, AV_LOG_ERROR, "Channel Count of %"PRIu16" is unsupported\n", data);
> + ret = AVERROR_PATCHWELCOME;
> + break;
goto end; would be clearer since makes clear that nothing else in this
function will happen. I missed the '!ret' in the for condition first.
The same applies to the other errors
> + }
> + } else if (tag == 14) {
> + av_log(avctx, AV_LOG_DEBUG, "Subband Count: %"PRIu16" \n", data);
> + if (data != 10) {
> + av_log(avctx, AV_LOG_ERROR, "Subband Count of %"PRIu16" is unsupported\n", data);
> + ret = AVERROR_PATCHWELCOME;
> + break;
> + }
I'd expect these tags to be only seen once per frame at the beginning of
the frame. If that assumption is true for all non-broken streams it
would make sense to parse this before the memory allocation and
ff_get_buffer(). It would allow raw decoding without parsing of single
frames. I'd also expect it to ioncrease rebustness against fuzzing.
> + } else if (tag == 62) {
> + s->channel_num = data;
> + av_log(avctx, AV_LOG_DEBUG, "Channel number %"PRIu16" \n", data);
> + init_plane_defaults(s);
> + } else if (tag == 48) {
> + if (s->subband_num != 0 && data == 1) // hack
> + s->level++;
> + av_log(avctx, AV_LOG_DEBUG, "Subband number %"PRIu16" \n", data);
> + s->subband_num = data;
> + } else if (tag == 51) {
> + av_log(avctx, AV_LOG_DEBUG, "Subband number actual %"PRIu16" \n", data);
> + s->subband_num_actual = data;
> + } else if (tag == 35)
> + av_log(avctx, AV_LOG_DEBUG, "Lowpass precision bits: %"PRIu16" \n", data);
> + else if (tag == 53) {
> + s->quantisation = data;
> + av_log(avctx, AV_LOG_DEBUG, "Quantisation: %"PRIu16" \n", data);
> + } else if (tag == 109) {
> + s->prescale_shift[0] = (data >> 0) & 0x7;
> + s->prescale_shift[1] = (data >> 3) & 0x7;
> + s->prescale_shift[2] = (data >> 6) & 0x7;
> + av_log(avctx, AV_LOG_DEBUG, "Prescale shift (VC-5): %x \n", data);
> + } else if (tag == 27) {
> + s->plane[s->channel_num].band[0][0].width = data;
> + s->plane[s->channel_num].band[0][0].stride = data;
> + av_log(avctx, AV_LOG_DEBUG, "Lowpass width %"PRIu16" \n", data);
> + } else if (tag == 28) {
> + s->plane[s->channel_num].band[0][0].height = data;
> + av_log(avctx, AV_LOG_DEBUG, "Lowpass height %"PRIu16" \n", data);
> + } else if (tag == 1)
> + av_log(avctx, AV_LOG_DEBUG, "Sample type? %"PRIu16" \n", data);
> + else if (tag == 10) {
> + if (data != 0) {
> + av_log(avctx, AV_LOG_ERROR, "Transform type of %"PRIu16" is unsupported\n", data);
> + ret = AVERROR_PATCHWELCOME;
> + break;
> + }
> + av_log(avctx, AV_LOG_DEBUG, "Transform-type? %"PRIu16" \n", data);
> + } else if (abstag >= 0x4000 && abstag <= 0x40ff) {
> + av_log(avctx, AV_LOG_DEBUG, "Small chunk length %"PRIu16" %s \n", data * 4, tag < 0 ? "optional" : "required");
> + bytestream2_skipu(&gb, data * 4);
> + } else if (tag == 23) {
> + av_log(avctx, AV_LOG_DEBUG, "Skip frame \n");
> + av_log(avctx, AV_LOG_ERROR, "Skip frame not supported \n");
> + ret = AVERROR_PATCHWELCOME;
> + break;
> + } else if (tag == 2) {
> + av_log(avctx, AV_LOG_DEBUG, "tag=2 header - skipping %i tag/value pairs \n", data);
> + for (i = 0; i < data; i++) {
> + av_log(avctx, AV_LOG_DEBUG, "Tag/Value = %x %x \n", bytestream2_get_be16(&gb), bytestream2_get_be16(&gb));
> + }
> + } else if (tag == 41) {
> + s->plane[s->channel_num].band[s->level][s->subband_num].width = data;
> + s->plane[s->channel_num].band[s->level][s->subband_num].stride = FFALIGN(data, 8);
> + av_log(avctx, AV_LOG_DEBUG, "Highpass width %i channel %i level %i subband %i \n", data, s->channel_num, s->level, s->subband_num);
> + } else if (tag == 42) {
> + s->plane[s->channel_num].band[s->level][s->subband_num].height = data;
> + av_log(avctx, AV_LOG_DEBUG, "Highpass height %i \n",
> data);
> + } else if (tag == 49) {
> + s->plane[s->channel_num].band[s->level][s->subband_num].width = data;
> + s->plane[s->channel_num].band[s->level][s->subband_num].stride = FFALIGN(data, 8);
> + av_log(avctx, AV_LOG_DEBUG, "Highpass width2 %i \n", data);
> + } else if (tag == 50) {
> + s->plane[s->channel_num].band[s->level][s->subband_num].height = data;
> + av_log(avctx, AV_LOG_DEBUG, "Highpass height2 %i \n", data);
> + } else if (tag == 71) {
> + s->codebook = data;
> + av_log(avctx, AV_LOG_DEBUG, "Codebook %i \n", s->codebook);
> + } else if (tag == 72) {
> + s->codebook = data;
> + av_log(avctx, AV_LOG_DEBUG, "Other codebook? %i \n", s->codebook);
> + } else
> + av_log(avctx, AV_LOG_DEBUG, "Unknown tag %i data %x \n", tag, data);
> +
> + coeff_data = subband[s->channel_num][s->subband_num_actual];
> +
> + /* Lowpass coefficients */
> + if (tag == 4 && data == 0xf0f) {
> + int lowpass_height = s->plane[s->channel_num].band[0][0].height;
> + int lowpass_width = s->plane[s->channel_num].band[0][0].width;
> +
> + if (lowpass_width * lowpass_height * sizeof(int16_t) > bytestream2_get_bytes_left(&gb)) {
> + av_log(avctx, AV_LOG_ERROR, "Too many lowpass coefficients \n");
> + ret = AVERROR(EINVAL);
> + goto end;
> + }
> +
> + av_log(avctx, AV_LOG_DEBUG, "Start of lowpass coeffs component %"PRIu16" \n", s->channel_num);
> + for (i = 0; i < lowpass_height; i++) {
> + for (j = 0; j < lowpass_width; j++) {
> + coeff_data[j] = AV_RB16(gb.buffer);
> +
> + gb.buffer += 2;
bytestream2_get_be16u
> + }
> + coeff_data += lowpass_width;
> + }
> +
> + /* Copy last line of coefficients if odd height */
> + if (lowpass_height & 1) {
> + memcpy(&coeff_data[lowpass_height * lowpass_width],
> + &coeff_data[(lowpass_height - 1) * lowpass_width],
> + lowpass_width * sizeof(*coeff_data));
> + }
> +
> + av_log(avctx, AV_LOG_DEBUG, "Lowpass coefficients %"PRIu16" \n", lowpass_width * lowpass_height);
> + }
> +
> + if (tag == 55 && s->subband_num_actual != 255) {
> + int highpass_height = s->plane[s->channel_num].band[s->level][s->subband_num].height;
> + int highpass_stride = s->plane[s->channel_num].band[s->level][s->subband_num].stride;
> + int expected = highpass_height * highpass_stride;
> + int level, run, coeff;
> + int count = 0, bytes;
> +
> + av_log(avctx, AV_LOG_DEBUG, "Start subband coeffs plane %i level %i codebook %i expected %i \n", s->channel_num, s->level, s->codebook, expected);
> +
> + init_get_bits(&s->gb, gb.buffer, bytestream2_get_bytes_left(&gb) * 8);
> + OPEN_READER(re, &s->gb);
> + if (!s->codebook) {
> + for (;;) {
> + UPDATE_CACHE(re, &s->gb);
> + GET_RL_VLC(level, run, re, &s->gb, s->table_9_rl_vlc,
> + VLC_BITS, 3, 1);
> +
> + /* escape */
> + if (level == 64)
> + break;
> +
> + count += run;
> +
> + if (count > expected)
> + break;
> +
> + coeff = dequant_and_decompand(level, s->quantisation);
> + for (i = 0; i < run; i++)
> + *coeff_data++ = coeff;
> + }
> + } else {
> + for (;;) {
> + UPDATE_CACHE(re, &s->gb);
> + GET_RL_VLC(level, run, re, &s->gb, s->table_18_rl_vlc,
> + VLC_BITS, 3, 1);
> +
> + /* escape */
> + if (level == 255 && run == 2)
> + break;
> +
> + count += run;
> +
> + if (count > expected)
> + break;
> +
> + coeff = dequant_and_decompand(level, s->quantisation);
> + for (i = 0; i < run; i++)
> + *coeff_data++ = coeff;
> + }
> + }
> + CLOSE_READER(re, &s->gb);
> +
> + if (count > expected) {
> + av_log(avctx, AV_LOG_ERROR, "Escape codeword not found, probably corrupt data \n");
> + ret = AVERROR(EINVAL);
> + goto end;
> + }
> +
> + bytes = FFALIGN(FF_CEIL_RSHIFT(get_bits_count(&s->gb), 3), 4);
note to the person who will push this: FF_CEIL_RSHIFT is not yet in
libav and will be added as AV_CEIL_RSHIFT
> + if (bytes > bytestream2_get_bytes_left(&gb)) {
> + av_log(avctx, AV_LOG_ERROR, "Bitstream overread error \n");
> + ret = AVERROR(EINVAL);
> + goto end;
> + } else
> + gb.buffer += bytes;
bytestream2_seek(gb, bytes, SEEK_CUR);
> +
> + av_log(avctx, AV_LOG_DEBUG, "End subband coeffs %i extra %i %p \n", count, count - expected, subband[s->subband_num_actual]);
> + s->codebook = 0;
> +
> + /* Copy last line of coefficients if odd height */
> + if (highpass_height & 1) {
> + memcpy(&coeff_data[highpass_height * highpass_stride],
> + &coeff_data[(highpass_height - 1) * highpass_stride],
> + highpass_stride * sizeof(*coeff_data));
> + }
> + }
The coefficient bitstream parsing blocks are good candidates to factor
out to their own functions to make cfhd_decode smaller.
> + }
> +
> + for (int plane = 0; plane < 3 && !ret; plane++) {
same for the filtering, the 'for' would move to the function and there
is a single 'if (!ret)' before a function call and no 'goto end' in the
while loop.
the rest looks ok to me
Janne
More information about the ffmpeg-devel
mailing list