Re: [FFmpeg-devel] [PATCH v2 1/3] avcodec/jpegxl_parser: add JPEG XL parser

[Leo Izen <leo.izen@gmail.com>]

On 6/21/23 20:59, James Almer wrote:
> On 6/21/2023 9:43 PM, Leo Izen wrote:
>> diff --git a/libavcodec/jpegxl_parse.c b/libavcodec/jpegxl_parse.c
>> new file mode 100644
>> index 0000000000..be360acb08
>> --- /dev/null
>> +++ b/libavcodec/jpegxl_parse.c
>> @@ -0,0 +1,22 @@
>> +/*
>> + * JPEG XL Header Parser
>> + * Copyright (c) 2023 Leo Izen<leo.izen@gmail.com>
>> + *
>> + * 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
>> + */
>> +
>> +#include "jpegxl_parse.h"
> 
> This is a really weird way to achieve code sharing between libraries.
> What you named jpegxl_parse.h should be jpegxl_parse.c instead, and 
> jpegxl_parse.h should only have the prototypes for 
> ff_jpegxl_collect_codestream_header() and 
> ff_jpegxl_parse_codestream_header(), plus the definition of 
> FFJXLMetadata. The enums can stay in jpegxl.h, and they for that matter 
> don't need the FF_ prefix.


This is based on JEEB's recommendation. He provided these as examples:

https://github.com/jeeb/ffmpeg/commit/b74e47c4ff5bca998936c0d8b9a0892104a7403d
https://github.com/jeeb/ffmpeg/commit/d7a75d21635eab4f4a1efea22945933059c2e36f

How else should I do it? If I do the usual thing where I put prototypes 
and structs in the header file and the implementation in the C file, 
then I won't be able to compile them into avformat as well, as it will 
just grab the stubs, and the ff_funcs won't be exported from avcodec.

> 
>> diff --git a/libavcodec/jpegxl_parser.c b/libavcodec/jpegxl_parser.c
>> new file mode 100644
>> index 0000000000..096b22be0f
>> --- /dev/null
>> +++ b/libavcodec/jpegxl_parser.c
>> @@ -0,0 +1,165 @@
>> +/*
>> + * JPEG XL Codec Parser
>> + * Copyright (c) 2023 Leo Izen <leo.izen@gmail.com>
>> + *
>> + * 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
>> + */
>> +
>> +#include "libavutil/intreadwrite.h"
>> +
>> +#include "codec_id.h"
>> +#include "jpegxl.h"
>> +#include "parser.h"
>> +
>> +typedef struct JXLParseContext {
>> +    ParseContext pc;
> 
> You don't need this. You're not assembling a packet.
> 
>> +    FFJXLMetadata meta;
>> +    int parsed_header;
>> +} JXLParseContext;
>> +
>> +static int jpegxl_parse(AVCodecParserContext *s, AVCodecContext *avctx,
>> +                        const uint8_t **poutbuf, int *poutbuf_size,
>> +                        const uint8_t *buf, int buf_size)
>> +{
>> +    JXLParseContext *ctx = s->priv_data;
>> +    int ret;
>> +
>> +    *poutbuf_size = 0;
>> +    *poutbuf = NULL;
>> +
>> +    if (!ctx->parsed_header) {
>> +        if (AV_RL64(buf) == FF_JPEGXL_CONTAINER_SIGNATURE_LE) {
>> +            int copied;
>> +            uint8_t codestream_header[4096];
>> +            ret = ff_jpegxl_collect_codestream_header(buf, buf_size, 
>> codestream_header,
>> +                                                      
>> sizeof(codestream_header), &copied);
>> +            if (ret < 0)
>> +                return ret;
>> +            /* copied may be larger than the bufsize if stuff was 
>> skipped */
>> +            copied = FFMIN(copied, sizeof(codestream_header));
>> +            ret = 
>> ff_jpegxl_parse_codestream_header(codestream_header, copied, 0, 
>> &ctx->meta);
>> +            if (ret < 0)
>> +                return ret;
>> +        } else {
>> +            ret = ff_jpegxl_parse_codestream_header(buf, buf_size, 1, 
>> &ctx->meta);
>> +            if (ret < 0)
>> +                return ret;
>> +        }
>> +        avctx->width = ctx->meta.width;
>> +        avctx->height = ctx->meta.height;
>> +
>> +        switch (ctx->meta.csp) {
>> +        case FF_JPEGXL_CS_RGB:
>> +        case FF_JPEGXL_CS_XYB:
>> +            avctx->colorspace = AVCOL_SPC_RGB;
>> +            break;
>> +        default:
>> +            avctx->colorspace = AVCOL_SPC_UNSPECIFIED;
>> +        }
>> +
>> +        if (ctx->meta.wp == FF_JPEGXL_WP_D65) {
>> +            switch (ctx->meta.primaries) {
>> +            case FF_JPEGXL_PR_SRGB:
>> +                avctx->color_primaries = AVCOL_PRI_BT709;
>> +                break;
>> +            case FF_JPEGXL_PR_P3:
>> +                avctx->color_primaries = AVCOL_PRI_SMPTE432;
>> +                break;
>> +            case FF_JPEGXL_PR_2100:
>> +                avctx->color_primaries = AVCOL_PRI_BT2020;
>> +                break;
>> +            default:
>> +                avctx->color_primaries = AVCOL_PRI_UNSPECIFIED;
>> +            }
>> +        } else if (ctx->meta.wp == FF_JPEGXL_WP_DCI && 
>> ctx->meta.primaries == FF_JPEGXL_PR_P3) {
>> +            avctx->color_primaries = AVCOL_PRI_SMPTE431;
>> +        } else {
>> +            avctx->color_primaries = AVCOL_PRI_UNSPECIFIED;
>> +        }
>> +
>> +        if (ctx->meta.trc > FF_JPEGXL_TR_GAMMA) {
>> +            FFJXLTransferCharacteristic trc = ctx->meta.trc - 
>> FF_JPEGXL_TR_GAMMA;
>> +            switch (trc) {
>> +            case FF_JPEGXL_TR_BT709:
>> +                avctx->color_trc = AVCOL_TRC_BT709;
>> +                break;
>> +            case FF_JPEGXL_TR_LINEAR:
>> +                avctx->color_trc = AVCOL_TRC_LINEAR;
>> +                break;
>> +            case FF_JPEGXL_TR_SRGB:
>> +                avctx->color_trc = AVCOL_TRC_IEC61966_2_1;
>> +                break;
>> +            case FF_JPEGXL_TR_PQ:
>> +                avctx->color_trc = AVCOL_TRC_SMPTEST2084;
>> +                break;
>> +            case FF_JPEGXL_TR_DCI:
>> +                avctx->color_trc = AVCOL_TRC_SMPTE428;
>> +                break;
>> +            case FF_JPEGXL_TR_HLG:
>> +                avctx->color_trc = AVCOL_TRC_ARIB_STD_B67;
>> +                break;
>> +            default:
>> +                avctx->color_trc = AVCOL_TRC_UNSPECIFIED;
>> +            }
>> +        } else if (ctx->meta.trc > 0) {
>> +            if (ctx->meta.trc > 45355 && ctx->meta.trc < 45555)
>> +                avctx->color_trc = AVCOL_TRC_GAMMA22;
>> +            else if (ctx->meta.trc > 35614 && ctx->meta.trc < 35814)
>> +                avctx->color_trc = AVCOL_TRC_GAMMA28;
>> +            else
>> +                avctx->color_trc = AVCOL_TRC_UNSPECIFIED;
>> +        } else {
>> +            avctx->color_trc = AVCOL_TRC_UNSPECIFIED;
>> +        }
>> +
>> +        if (ctx->meta.csp == FF_JPEGXL_CS_GRAY) {
>> +            if (ctx->meta.bit_depth <= 8)
>> +                avctx->pix_fmt = ctx->meta.have_alpha ? 
>> AV_PIX_FMT_YA8 : AV_PIX_FMT_GRAY8;.
> 
> Set s->format, not avctx->pix_fmt. The format in an AVCodecContext is 
> set by a decoder, which may not match what the parser sets.
> 
>> +            else if (ctx->meta.bit_depth <= 16)
>> +                avctx->pix_fmt = ctx->meta.have_alpha ? 
>> AV_PIX_FMT_YA16 : AV_PIX_FMT_GRAY16;
>> +            else
>> +                avctx->pix_fmt = ctx->meta.have_alpha ? 
>> AV_PIX_FMT_NONE : AV_PIX_FMT_GRAYF32;
>> +        } else {
>> +            if (ctx->meta.bit_depth <= 8)
>> +                avctx->pix_fmt = ctx->meta.have_alpha ? 
>> AV_PIX_FMT_RGBA : AV_PIX_FMT_RGB24;
>> +            else if (ctx->meta.bit_depth <= 16)
>> +                avctx->pix_fmt = ctx->meta.have_alpha ? 
>> AV_PIX_FMT_RGBA64 : AV_PIX_FMT_RGB48;
>> +            else
>> +                avctx->pix_fmt = ctx->meta.have_alpha ? 
>> AV_PIX_FMT_RGBAF32 : AV_PIX_FMT_RGBF32;
>> +        }
>> +
>> +        ctx->parsed_header = 1;
> 
> So this is done once for the entire stream?

That's the intention.

> 
>> +    }
>> +
>> +    ctx->pc.frame_start_found = 1;
> 
> Unnecessary.
> 
>> +
>> +    ret = ff_combine_frame(&ctx->pc, END_NOT_FOUND, &buf, &buf_size);
> 
> Also unnecessary.
> 
>> +    if (ret < 0)
>> +        return buf_size;
>> +
>> +    *poutbuf = buf;
>> +    *poutbuf_size = buf_size;
>> +
>> +    return END_NOT_FOUND;
> 
> ???
> 
> The process succeeded. Return 0.

If I'm not packetizing, I wasn't sure if I could return any other value, 
but I can change it to return 0.

> 
>> +}
>> +
>> +const AVCodecParser ff_jpegxl_parser = {
>> +    .codec_ids      = { AV_CODEC_ID_JPEGXL },
>> +    .priv_data_size = sizeof(JXLParseContext),
>> +    .parser_parse   = jpegxl_parse,
>> +    .parser_close   = ff_parse_close,
>> +};
> 
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".