18

I wanna resize a picture by using the ffmpeg's func--->sws_scale().

Is there any one knows how to do it?

Do you have the source code for this function?

Charles
  • 48,924
  • 13
  • 96
  • 136
cyh24
  • 226
  • 1
  • 3
  • 13

2 Answers2

35

First you need to create a SwsContext (you need to do this only once) :

struct SwsContext *resize;
resize = sws_getContext(width1, height1, AV_PIX_FMT_YUV420P, width2, height2, PIX_FMT_RGB24, SWS_BICUBIC, NULL, NULL, NULL);

You need two frames for conversion, frame1 is the original frame, you need to explicitly allocate frame2 :

AVFrame* frame1 = avcodec_alloc_frame(); // this is your original frame

AVFrame* frame2 = avcodec_alloc_frame();
int num_bytes = avpicture_get_size(AV_PIX_FMT_RGB24, width2, height2);
uint8_t* frame2_buffer = (uint8_t *)av_malloc(num_bytes*sizeof(uint8_t));
avpicture_fill((AVPicture*)frame2, frame2_buffer, AV_PIX_FMT_RGB24, width2, height2);

You may use this part inside a loop if you need to resize each frame you receive :

// frame1 should be filled by now (eg using avcodec_decode_video)
sws_scale(resize, frame1->data, frame1->linesize, 0, height1, frame2->data, frame2->linesize);

Note that I also changed pixel format, but you can use the same pixel format for both frames

Timothy Gu
  • 2,977
  • 21
  • 30
arash kordi
  • 2,242
  • 19
  • 22
  • I get a null value from sws_getContext(...). Any idea why? This is my call mpFrameSwsContext = sws_getContext(mAVCodecContext->width, mAVCodecContext->height, mAVCodecContext->pix_fmt, mAVCodecContext->width,mAVCodecContext->height, PIX_FMT_BGR24, SWS_BICUBIC, NULL, NULL, NULL); – Tariq Oct 29 '14 at 16:01
  • It seems like after this code most of the members in `frame2` are not filled out (`width`, `height`, `format`, `pict_type`, etc). How do I get valid values for those? – David Jul 06 '15 at 20:50
  • @Dave: avpicture_fill() should fill the required members – arash kordi Jul 07 '15 at 08:13
  • Can sws_scale produce a "padded" image to fit a display rectangle of a different aspect ratio, instead of always stretching? I ask because I'm trying to do it with avfilter and not having much luck, while I have to use sws_scale for h264->RGB conversion anyway. – KeithS Sep 03 '16 at 01:30
  • @KeithS: just use the same width/height for destination in sws_getContext(), the rest of code is the same. your image will be located on upper left corner of the frame – arash kordi Sep 03 '16 at 08:06
  • Thanks. Turns out the example control I was using was just using GDI halftone scaling, and ffmpeg was never aware of the real destination size. So I stayed in GDI-land, scaled the picture to the maximum allowable dimension, centered it and drew some black rectangles to either side. Works for what we need. – KeithS Sep 15 '16 at 15:56
6

Runnable example in FFmpeg 2.8

Basically using arash's method, but runnable so you can try it out.

Generate one short video procedurally, and then convert it to 3 different sizes.

ffmpeg_encoder_init_frame and ffmpeg_encoder_scale are the key methods.

Source:

#include <libavcodec/avcodec.h>
#include <libavutil/imgutils.h>
#include <libavutil/opt.h>
#include <libswscale/swscale.h>

static AVCodecContext *c = NULL;
static AVFrame *frame;
static AVFrame *frame2;
static AVPacket pkt;
static FILE *file;
static struct SwsContext *sws_context = NULL;

static void ffmpeg_encoder_init_frame(AVFrame **framep, int width, int height) {
    int ret;
    AVFrame *frame;
    frame = av_frame_alloc();
    if (!frame) {
        fprintf(stderr, "Could not allocate video frame\n");
        exit(1);
    }
    frame->format = c->pix_fmt;
    frame->width  = width;
    frame->height = height;
    ret = av_image_alloc(frame->data, frame->linesize, frame->width, frame->height, frame->format, 32);
    if (ret < 0) {
        fprintf(stderr, "Could not allocate raw picture buffer\n");
        exit(1);
    }
    *framep = frame;
}

static void ffmpeg_encoder_scale(uint8_t *rgb) {
    sws_context = sws_getCachedContext(sws_context,
            frame->width, frame->height, AV_PIX_FMT_YUV420P,
            frame2->width, frame2->height, AV_PIX_FMT_YUV420P,
            SWS_BICUBIC, NULL, NULL, NULL);
    sws_scale(sws_context, (const uint8_t * const *)frame->data, frame->linesize, 0,
            frame->height, frame2->data, frame2->linesize);
}

static void ffmpeg_encoder_set_frame_yuv_from_rgb(uint8_t *rgb) {
    const int in_linesize[1] = { 3 * frame->width };
    sws_context = sws_getCachedContext(sws_context,
            frame->width, frame->height, AV_PIX_FMT_RGB24,
            frame->width, frame->height, AV_PIX_FMT_YUV420P,
            0, NULL, NULL, NULL);
    sws_scale(sws_context, (const uint8_t * const *)&rgb, in_linesize, 0,
            frame->height, frame->data, frame->linesize);
}

void generate_rgb(int width, int height, int pts, uint8_t **rgbp) {
    int x, y, cur;
    uint8_t *rgb = *rgbp;
    rgb = realloc(rgb, 3 * sizeof(uint8_t) * height * width);
    for (y = 0; y < height; y++) {
        for (x = 0; x < width; x++) {
            cur = 3 * (y * width + x);
            rgb[cur + 0] = 0;
            rgb[cur + 1] = 0;
            rgb[cur + 2] = 0;
            if ((frame->pts / 25) % 2 == 0) {
                if (y < height / 2) {
                    if (x < width / 2) {
                        /* Black. */
                    } else {
                        rgb[cur + 0] = 255;
                    }
                } else {
                    if (x < width / 2) {
                        rgb[cur + 1] = 255;
                    } else {
                        rgb[cur + 2] = 255;
                    }
                }
            } else {
                if (y < height / 2) {
                    rgb[cur + 0] = 255;
                    if (x < width / 2) {
                        rgb[cur + 1] = 255;
                    } else {
                        rgb[cur + 2] = 255;
                    }
                } else {
                    if (x < width / 2) {
                        rgb[cur + 1] = 255;
                        rgb[cur + 2] = 255;
                    } else {
                        rgb[cur + 0] = 255;
                        rgb[cur + 1] = 255;
                        rgb[cur + 2] = 255;
                    }
                }
            }
        }
    }
    *rgbp = rgb;
}

void ffmpeg_encoder_start(const char *filename, int codec_id, int fps, int width, int height, float factor) {
    AVCodec *codec;
    int ret;
    int width2 = width * factor;
    int height2 = height * factor;
    avcodec_register_all();
    codec = avcodec_find_encoder(codec_id);
    if (!codec) {
        fprintf(stderr, "Codec not found\n");
        exit(1);
    }
    c = avcodec_alloc_context3(codec);
    if (!c) {
        fprintf(stderr, "Could not allocate video codec context\n");
        exit(1);
    }
    c->bit_rate = 400000;
    c->width = width2;
    c->height = height2;
    c->time_base.num = 1;
    c->time_base.den = fps;
    c->gop_size = 10;
    c->max_b_frames = 1;
    c->pix_fmt = AV_PIX_FMT_YUV420P;
    if (codec_id == AV_CODEC_ID_H264)
        av_opt_set(c->priv_data, "preset", "slow", 0);
    if (avcodec_open2(c, codec, NULL) < 0) {
        fprintf(stderr, "Could not open codec\n");
        exit(1);
    }
    file = fopen(filename, "wb");
    if (!file) {
        fprintf(stderr, "Could not open %s\n", filename);
        exit(1);
    }
    ffmpeg_encoder_init_frame(&frame, width, height);
    ffmpeg_encoder_init_frame(&frame2, width2, height2);
}

void ffmpeg_encoder_finish(void) {
    uint8_t endcode[] = { 0, 0, 1, 0xb7 };
    int got_output, ret;
    do {
        fflush(stdout);
        ret = avcodec_encode_video2(c, &pkt, NULL, &got_output);
        if (ret < 0) {
            fprintf(stderr, "Error encoding frame\n");
            exit(1);
        }
        if (got_output) {
            fwrite(pkt.data, 1, pkt.size, file);
            av_packet_unref(&pkt);
        }
    } while (got_output);
    fwrite(endcode, 1, sizeof(endcode), file);
    fclose(file);
    avcodec_close(c);
    av_free(c);
    av_freep(&frame->data[0]);
    av_frame_free(&frame);
    av_freep(&frame2->data[0]);
    av_frame_free(&frame2);
}

void ffmpeg_encoder_encode_frame(uint8_t *rgb) {
    int ret, got_output;
    ffmpeg_encoder_set_frame_yuv_from_rgb(rgb);
    ffmpeg_encoder_scale(rgb);
    frame2->pts = frame->pts;
    av_init_packet(&pkt);
    pkt.data = NULL;
    pkt.size = 0;
    ret = avcodec_encode_video2(c, &pkt, frame2, &got_output);
    if (ret < 0) {
        fprintf(stderr, "Error encoding frame\n");
        exit(1);
    }
    if (got_output) {
        fwrite(pkt.data, 1, pkt.size, file);
        av_packet_unref(&pkt);
    }
}

static void encode_example(float factor) {
    char filename[255];
    int pts;
    int width = 320;
    int height = 240;
    uint8_t *rgb = NULL;
    snprintf(filename, 255, "tmp." __FILE__ ".%.2f.h264", factor);
    ffmpeg_encoder_start(filename, AV_CODEC_ID_H264, 25, width, height, factor);
    for (pts = 0; pts < 100; pts++) {
        frame->pts = pts;
        generate_rgb(width, height, pts, &rgb);
        ffmpeg_encoder_encode_frame(rgb);
    }
    ffmpeg_encoder_finish();
    free(rgb);
}

int main(void) {
    encode_example(0.5);
    encode_example(1.0);
    encode_example(2.0);
    return EXIT_SUCCESS;
}

Run with:

gcc main.c -lavformat -lavcodec -lswresample -lswscale -lavutil -lx264
./a.out
ffplay tmp.main.c.0.50.h264
ffplay tmp.main.c.1.00.h264
ffplay tmp.main.c.2.00.h264

Tested on Ubuntu 16.04. Source on GitHub.

  • do you have to do color conversion and scaling in different steps? When decoding I thought the point was you do both at the same time, shouldn't encoding allow you to do the same? – Matt Wolfe May 05 '16 at 02:05