使用自定义 AVIOContext 查找流信息时出现 FFMPEG 错误

Question

我正在编写将文件作为流接收并对其进行解码的软件。我有以下用于流输入的自定义 AVIO 代码：

/* Allocate a 4kb buffer for copying. */
std::uint32_t bufSize = 4096;
struct vidBuf
{
    std::byte* ptr;
    int size;
};

vidBuf tmpVidBuf = { const_cast<std::byte*>(videoBuffer.data()),
    static_cast<int>(videoBuffer.size()) };
AVIOContext *avioContext =
    avio_alloc_context(reinterpret_cast<std::uint8_t*>(av_malloc(bufSize)),
                       bufSize, 0,
                       reinterpret_cast<void*>(&tmpVidBuf),
                       [](void *opaque, std::uint8_t *buf, int bufSize) -> int
                       {
                           auto &me = *reinterpret_cast<vidBuf*>(opaque);
                           bufSize = std::min(bufSize, me.size);

                           std::copy_n(me.ptr, bufSize, reinterpret_cast<std::byte*>(buf));
                           me.ptr += bufSize;
                           me.size -= bufSize;
                           return bufSize;
                       }, nullptr, nullptr);

auto avFormatPtr = avformat_alloc_context();
avFormatPtr->pb = avioContext;
avFormatPtr->flags |= AVFMT_FLAG_CUSTOM_IO;
//avFormatPtr->probesize = tmpVidBuf.size;
//avFormatPtr->max_analyze_duration = 5000000;

avformat_open_input(&avFormatPtr, nullptr, nullptr, nullptr);

if(auto ret = avformat_find_stream_info(avFormatPtr, nullptr);
   ret < 0)
    logerror << "Could not open the video file: " << makeAVError(ret) << '\n';

但是，当我运行这段代码时，我得到了错误：

[mov,mp4,m4a,3gp,3g2,mj2 @ 0x55d10736d580] stream 0, offset 0x30: partial file
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x55d10736d580] Could not find codec parameters for stream 0 (Video: h264 (avc1 / 0x31637661), none(tv, bt709), 540x360, 649 kb/s): unspecified pixel format
Consider increasing the value for the 'analyzeduration' (0) and 'probesize' (5000000) options
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf58.76.100
  Duration: 00:04:08.41, start: 0.000000, bitrate: N/A
  Stream #0:0(und): Video: h264 (avc1 / 0x31637661), none(tv, bt709), 540x360, 649 kb/s, SAR 1:1 DAR 3:2, 29.97 fps, 29.97 tbr, 30k tbn, 60k tbc (default)
    Metadata:
      handler_name    : ISO Media file produced by Google Inc. Created on: 01/10/2021.
      vendor_id       : [0][0][0][0]
  Stream #0:1(und): Audio: aac (mp4a / 0x6134706D), 22050 Hz, mono, fltp, 69 kb/s (default)
    Metadata:
      handler_name    : ISO Media file produced by Google Inc. Created on: 01/10/2021.
      vendor_id       : [0][0][0][0]
Assertion desc failed at libswscale/swscale_internal.h:677

注意视频流数据中缺少 YUV420p 部分。

这很奇怪，因为如果我运行我的程序使用不同的 mp4 文件，它工作得很好，这个错误只发生在特定的 mp4 文件上。我知道 mp4 文件是有效的，因为 mpv 可以播放它，并且 ffprobe 能够获取它的元数据：

Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'heard.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf58.76.100
  Duration: 00:04:08.41, start: 0.000000, bitrate: 724 kb/s
  Stream #0:0(und): Video: h264 (Main) (avc1 / 0x31637661), yuv420p(tv, bt709), 540x360 [SAR 1:1 DAR 3:2], 649 kb/s, 29.97 fps, 29.97 tbr, 30k tbn, 59.94 tbc (default)
    Metadata:
      handler_name    : ISO Media file produced by Google Inc. Created on: 01/10/2021.
      vendor_id       : [0][0][0][0]
  Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 22050 Hz, mono, fltp, 69 kb/s (default)
    Metadata:
      handler_name    : ISO Media file produced by Google Inc. Created on: 01/10/2021.
      vendor_id       : [0][0][0][0]

正如您在我的代码中所见，我也尝试设置 analyzeduration 和 probesize，但这并没有解决问题。

我也知道这个错误是因为我的自定义io，因为当我avformat_open_input直接打开文件时，它能够很好地解码。我是 ffmpeg 的新手，所以我可能错过了一些简单的东西。

Answer 1

正如 SuRGeoNix 所指出的，我没有为 AVIO 上下文实现搜索功能；我认为这搞砸了 FFMPEG，因为它无法计算出缓冲区的大小。这是我现在的工作代码：

std::uint32_t bufSize = 4096;
struct vidBuf
{
    std::byte* ptr;
    std::byte* origPtr;
    int size;
    int fullSize;
};

vidBuf tmpVidBuf = { const_cast<std::byte*>(videoBuffer.data()),
const_cast<std::byte*>(videoBuffer.data()),
static_cast<int>(videoBuffer.size()),
static_cast<int>(videoBuffer.size()), };

AVIOContext *avioContext =
    avio_alloc_context(reinterpret_cast<std::uint8_t*>(av_malloc(bufSize)),
                       bufSize, 0,
                       reinterpret_cast<void*>(&tmpVidBuf),
                       [](void *opaque, std::uint8_t *buf, int bufSize) -> int
                       {
                           auto &me = *reinterpret_cast<vidBuf*>(opaque);
                           bufSize = std::min(bufSize, me.size);

                           std::copy_n(me.ptr, bufSize, reinterpret_cast<std::byte*>(buf));
                           me.ptr += bufSize;
                           me.size -= bufSize;
                           return bufSize;
                       },
                       nullptr,
                       [](void *opaque, std::int64_t where, int whence) -> std::int64_t
                       {
                           auto me = reinterpret_cast<vidBuf*>(opaque);

                           switch(whence)
                           {
                           case AVSEEK_SIZE:
                               /* Maybe size left? */
                               return me->fullSize;
                               break;
                           case SEEK_SET:
                               if(me->fullSize > where)
                               {
                                   me->ptr = me->origPtr + where;
                                   me->size = me->fullSize - where;
                               }
                               else
                                   return EOF;
                               break;
                           case SEEK_CUR:
                               if(me->size > where)
                               {
                                   me->ptr += where;
                                   me->size -= where;
                               }
                               else
                                   return EOF;
                               break;
                           case SEEK_END:
                               if(me->fullSize > where)
                               {
                                   me->ptr = (me->origPtr + me->fullSize) - where;
                                   int curPos = me->ptr - me->origPtr;
                                   me->size = me->fullSize - curPos;
                               }
                               else
                                   return EOF;
                               break;
                           default:
                           /* On error, do nothing, return current position of file. */
                               logerror << "Could not process buffer seek: "
                                        << whence << ".\n";
                               break;
                           }
                           return me->ptr - me->origPtr;
                       });

使用自定义 AVIOContext 查找流信息时出现 FFMPEG 错误

FFMPEG error when finding stream information with custom AVIOContext

c++

ffmpeg

libav