{"id":8078,"date":"2026-03-25T13:51:55","date_gmt":"2026-03-25T05:51:55","guid":{"rendered":"https:\/\/kyle.ai\/blog\/?p=8078"},"modified":"2026-04-03T15:25:48","modified_gmt":"2026-04-03T07:25:48","slug":"%e4%bd%bf%e7%94%a8-golang-%e6%8f%90%e5%8f%96%e8%a7%86%e9%a2%91%e6%96%87%e4%bb%b6%e4%b8%ad%e7%9a%84%e5%ad%97%e5%b9%95%e6%96%87%e4%bb%b6","status":"publish","type":"post","link":"https:\/\/kyle.ai\/blog\/8078.html","title":{"rendered":"\u4f7f\u7528 Golang \u63d0\u53d6\u89c6\u9891\u6587\u4ef6\u4e2d\u7684\u5b57\u5e55\u6587\u4ef6"},"content":{"rendered":"\n

\u672c\u6765\u662f\u60f3\u7814\u7a76\u4e00\u4e0b\uff0c\u80fd\u4e0d\u80fd\u53ea\u8bfb\u53d6\u89c6\u9891\u6587\u4ef6\u7684\u524d\u9762\u5f88\u5c0f\u4e00\u90e8\u5206\u6570\u636e\uff0c\u4f8b\u5982\u53ea\u8bfb\u5f00\u5934\u7684 2MB\uff0c\u5c31\u53ef\u4ee5\u628a\u5185\u7f6e\u5b57\u5e55\u6587\u4ef6\u5bfc\u51fa\u6765\u3002<\/p>\n\n\n\n

\u7ed3\u679c\u53d1\u73b0\u884c\u4e0d\u901a\uff0cMP4 \u548c MKV \u6587\u4ef6\uff0c\u5b83\u4eec\u7684\u5b57\u5e55\u90fd\u662f\u8ddf\u968f\u89c6\u9891\u6570\u636e\u4e00\u8d77\u7684\uff0c\u5206\u6563\u5230\u6587\u4ef6\u7684\u4e0d\u540c\u4f4d\u7f6e\uff0c\u6240\u4ee5\u8981\u5bfc\u51fa\u5b8c\u6574\u5b57\u5e55\u6587\u4ef6\uff0c\u4f60\u9700\u8981\u8bfb\u53d6\u6574\u4e2a\u89c6\u9891\u6587\u4ef6\u3002<\/p>\n\n\n\n

mp4 \u6587\u4ef6<\/h2>\n\n\n\n

MP4 \u6587\u4ef6\u662f BOX \u6a21\u578b\uff0c\u8fd9\u91cc\u6709\u4e2a\u5f88\u597d\u7684\u8bf4\u660e\u6587\u6863 https:\/\/www.pedestrian.com.cn\/user\/video\/mp4_muxer.html<\/a><\/p>\n\n\n\n

\u6211\u4eec\u9700\u8981\u8bfb\u53d6\u7684 BOX \u5d4c\u5957\u89c4\u5219\u5982\u4e0b<\/p>\n\n\n\n

moov -> trak -> mdia -> mdhd\n                     -> hdlr\n                     -> minf -> stbl -> stts\n                                     -> stsz\n                                     -> stco<\/code><\/pre>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u793a\u4f8b\u7a0b\u5e8f\uff0c\u8bfb\u53d6 mp4 \u6587\u4ef6\uff0c\u6253\u5370\u51fa\u5b57\u5e55\u5217\u8868\uff0c\u5e76\u5c06\u5176\u5bfc\u51fa\u4e3a srt \u5b57\u5e55\u6587\u4ef6<\/p>\n\n\n
\npackage main\nimport (\n\t"encoding\/binary"\n\t"fmt"\n\t"io"\n\t"os"\n\t"time"\n\t"github.com\/abema\/go-mp4"\n)\nfunc locateMoov(rs io.ReadSeeker) {\n\tfor {\n\t\t\/\/ \u83b7\u53d6\u5f53\u524d\u4f4d\u7f6e\n\t\tcurrentPos, _ := rs.Seek(0, io.SeekCurrent)\n\t\t\/\/ \u8bfb\u53d6 Box Header (8 \u5b57\u8282)\n\t\tvar header [8]byte\n\t\tif _, err := rs.Read(header[:]); err != nil {\n\t\t\tbreak \/\/ \u5230\u8fbe\u6587\u4ef6\u672b\u5c3e\n\t\t}\n\t\tsize := binary.BigEndian.Uint32(header[0:4])\n\t\tboxType := string(header[4:8])\n\t\tif boxType == "moov" {\n\t\t\tfmt.Printf("\u627e\u5230 moov! \u8d77\u59cb\u504f\u79fb\u91cf: %d\uff0c\u5927\u5c0f\uff1a%d\\n", currentPos, size)\n\t\t\trs.Seek(currentPos, io.SeekStart) \/\/ \u56de\u5230\u8d77\u70b9\u5f00\u59cb\u8be6\u7ec6\u89e3\u6790\n\t\t\treturn\n\t\t}\n\t\tif boxType == "mdat" {\n\t\t\tvar actualSize int64\n\t\t\tif size == 1 {\n\t\t\t\t\/\/ \u5904\u7406 64 \u4f4d\u5927\u6587\u4ef6\u957f\u5ea6\n\t\t\t\tvar largeSize [8]byte\n\t\t\t\trs.Read(largeSize[:])\n\t\t\t\tactualSize = int64(binary.BigEndian.Uint64(largeSize[:]))\n\t\t\t} else {\n\t\t\t\tactualSize = int64(size)\n\t\t\t}\n\t\t\tfmt.Printf("\u8df3\u8fc7 mdat, \u5927\u5c0f: %d \u5b57\u8282\uff0c\u5f53\u524d offset %d\\n", actualSize, currentPos)\n\t\t\t\/\/ \u6838\u5fc3\u64cd\u4f5c\uff1a\u7cbe\u51c6\u8df3\u8fc7 mdat \u7684\u4e3b\u4f53\n\t\t\trs.Seek(currentPos+actualSize, io.SeekStart)\n\t\t\tcontinue\n\t\t}\n\t\t\/\/ \u5176\u4ed6 Box (\u5982 ftyp, free)\uff0c\u6839\u636e\u5176 size \u6b63\u5e38\u8df3\u8fc7\n\t\trs.Seek(currentPos+int64(size), io.SeekStart)\n\t}\n}\n\/\/ \u8f85\u52a9\u51fd\u6570\uff1a\u5c06 MP4 \u5185\u90e8\u7684 uint16 \u8bed\u8a00\u4ee3\u7801\u8f6c\u4e3a\u5b57\u7b26\u4e32 (\u5982 "chi", "eng")\nfunc decodeLanguage(lang uint16) string {\n\t\/\/ \u7b97\u6cd5\uff1a\u6bcf\u4e2a\u5b57\u7b26\u5360 5 \u4f4d\uff0c\u504f\u79fb\u91cf\u4e3a 0x60\n\tc1 := rune((lang>>10)&0x1F) + 0x60\n\tc2 := rune((lang>>5)&0x1F) + 0x60\n\tc3 := rune(lang&0x1F) + 0x60\n\treturn string([]rune{c1, c2, c3})\n}\ntype SampleInfo struct {\n\tOffset int64\n\tSize   uint32\n\tStart  time.Duration\n\tEnd    time.Duration\n}\ntype SubtitleTrack struct {\n\tID         uint32\n\tLanguage   string\n\tTimescale  uint32\n\tSamples    []SampleInfo\n\tIsSubtitle bool\n}\nfunc main() {\n\tif len(os.Args) < 2 {\n\t\tpanic("\u8bf7\u63d0\u4f9b\u6587\u4ef6\u8def\u5f84")\n\t}\n\tfilePath := os.Args[1]\n\tf, err := os.Open(filePath) \/\/ \u66ff\u6362\u4e3a\u4f60\u7684\u6587\u4ef6\n\tif err != nil {\n\t\tpanic(err)\n\t}\n\tdefer f.Close()\n\t\/\/ \u8df3\u8fc7\u5176\u5b83 box\uff0c\u76f4\u63a5\u8bfb\u53d6 moov \u7ed3\u6784\n\tlocateMoov(f)\n\tvar tracks []*SubtitleTrack\n\tvar cur *SubtitleTrack\n\t\/\/ 1. \u89e3\u6790\u7ed3\u6784\n\t_, err = mp4.ReadBoxStructure(f, func(h *mp4.ReadHandle) (any, error) {\n\t\tswitch h.BoxInfo.Type.String() {\n\t\tcase "moov", "mdia", "minf", "stbl":\n\t\t\treturn h.Expand()\n\t\tcase "trak":\n\t\t\t\/\/ \u8fdb\u5165\u65b0\u8f68\u9053\uff0c\u521d\u59cb\u5316\u4e34\u65f6\u5bf9\u8c61\n\t\t\tcur = &SubtitleTrack{ID: uint32(len(tracks) + 1)}\n\t\t\t_, err := h.Expand()\n\t\t\tif cur.IsSubtitle {\n\t\t\t\ttracks = append(tracks, cur)\n\t\t\t}\n\t\t\treturn nil, err\n\t\tcase "hdlr":\n\t\t\tpayload, _, _ := h.ReadPayload()\n\t\t\thdlr := payload.(*mp4.Hdlr)\n\t\t\tif hdlr.HandlerType == [4]byte{'s', 'b', 't', 'l'} {\n\t\t\t\tcur.IsSubtitle = true\n\t\t\t}\n\t\tcase "mdhd":\n\t\t\tpayload, _, _ := h.ReadPayload()\n\t\t\tmdhd := payload.(*mp4.Mdhd)\n\t\t\tcur.Timescale = mdhd.Timescale\n\t\t\tlangCode := uint16(mdhd.Language[0])<<10 | uint16(mdhd.Language[1])<<5 | uint16(mdhd.Language[2])\n\t\t\tcur.Language = decodeLang(langCode)\n\t\tcase "stts": \/\/ \u65f6\u95f4\u8868\n\t\t\tpayload, _, _ := h.ReadPayload()\n\t\t\tstts := payload.(*mp4.Stts)\n\t\t\tvar runningTime uint64\n\t\t\tfor _, entry := range stts.Entries {\n\t\t\t\tfor i := uint32(0); i < entry.SampleCount; i++ {\n\t\t\t\t\tstart := time.Duration(runningTime) * time.Second \/ time.Duration(cur.Timescale)\n\t\t\t\t\trunningTime += uint64(entry.SampleDelta)\n\t\t\t\t\tend := time.Duration(runningTime) * time.Second \/ time.Duration(cur.Timescale)\n\t\t\t\t\tcur.Samples = append(cur.Samples, SampleInfo{Start: start, End: end})\n\t\t\t\t}\n\t\t\t}\n\t\tcase "stsz": \/\/ \u5927\u5c0f\u8868\n\t\t\tpayload, _, _ := h.ReadPayload()\n\t\t\tstsz := payload.(*mp4.Stsz)\n\t\t\tfor i := 0; i < len(cur.Samples) && i < len(stsz.EntrySize); i++ {\n\t\t\t\tcur.Samples[i].Size = stsz.EntrySize[i]\n\t\t\t}\n\t\tcase "stco": \/\/ \u504f\u79fb\u91cf\u8868\n\t\t\tpayload, _, _ := h.ReadPayload()\n\t\t\tstco := payload.(*mp4.Stco)\n\t\t\tfor i := 0; i < len(cur.Samples) && i < len(stco.ChunkOffset); i++ {\n\t\t\t\tcur.Samples[i].Offset = int64(stco.ChunkOffset[i])\n\t\t\t}\n\t\tcase "co64": \/\/ 64\u4f4d\u504f\u79fb\u91cf\n\t\t\tpayload, _, _ := h.ReadPayload()\n\t\t\tco64 := payload.(*mp4.Co64)\n\t\t\tfor i := 0; i < len(cur.Samples) && i < len(co64.ChunkOffset); i++ {\n\t\t\t\tcur.Samples[i].Offset = int64(co64.ChunkOffset[i])\n\t\t\t}\n\t\t}\n\t\treturn nil, nil\n\t})\n\t\/\/ 2. \u5bfc\u51fa\n\tfor _, t := range tracks {\n\t\tfmt.Printf("%s %d\\n", t.Language, len(t.Samples))\n\t\tsaveToSrt(f, t)\n\t}\n}\n\/\/ \u89e3\u7801 MP4 \u5185\u90e8\u8bed\u8a00\u4ee3\u7801 (ISO-639-2\/T)\nfunc decodeLang(l uint16) string {\n\t\/\/ 5\u4f4d\u4e00\u4e2a\u5b57\u6bcd\uff0c\u504f\u79fb\u91cf 0x60\n\tc1 := byte((l>>10)&0x1F) + 0x60\n\tc2 := byte((l>>5)&0x1F) + 0x60\n\tc3 := byte(l&0x1F) + 0x60\n\treturn string([]byte{c1, c2, c3})\n}\nfunc saveToSrt(f io.ReadSeeker, t *SubtitleTrack) {\n\toutName := fmt.Sprintf("track_%d_%s.srt", t.ID, t.Language)\n\tout, _ := os.Create(outName)\n\tdefer out.Close()\n\tfmt.Printf("\u5bfc\u51fa\u8f68\u9053 %d [%s], \u5171 %d \u6761\u5b57\u5e55...\\n", t.ID, t.Language, len(t.Samples))\n\tfor i, s := range t.Samples {\n\t\tf.Seek(s.Offset, io.SeekStart)\n\t\tdata := make([]byte, s.Size)\n\t\tf.Read(data)\n\t\tif len(data) < 2 {\n\t\t\tcontinue\n\t\t}\n\t\ttxtLen := binary.BigEndian.Uint16(data[:2])\n\t\tif int(txtLen)+2 > len(data) {\n\t\t\ttxtLen = uint16(len(data) - 2)\n\t\t}\n\t\tcontent := string(data[2 : 2+txtLen])\n\t\tfmt.Fprintf(out, "%d\\n%s --> %s\\n%s\\n\\n",\n\t\t\ti+1, formatSrtTime(s.Start), formatSrtTime(s.End), content)\n\t}\n}\nfunc formatSrtTime(d time.Duration) string {\n\tms := d.Milliseconds() % 1000\n\tsec := int64(d.Seconds()) % 60\n\tmin := int64(d.Minutes()) % 60\n\thr := int64(d.Hours())\n\treturn fmt.Sprintf("%02d:%02d:%02d,%03d", hr, min, sec, ms)\n}\n\n<\/pre><\/div>\n\n\n
mkv \u6587\u4ef6<\/h2>\n\n\n\n
\u793a\u4f8b\u7a0b\u5e8f\uff0c\u8bfb\u53d6 mkv \u6587\u4ef6\uff0c\u6253\u5370\u51fa\u5b57\u5e55\u5217\u8868\u3002\u8fd9\u91cc\u5c31\u6ca1\u7ee7\u7eed\u7814\u7a76\u5bfc\u51fa\u5b57\u5e55\u6587\u4ef6\u4e86\uff0c\u8981\u8bfb\u53d6\u6574\u4e2a mkv \u6587\u4ef6\uff0c\u6bd4\u8f83\u9ebb\u70e6\uff0c\u800c\u4e14\u901f\u5ea6\u4e5f\u8f83\u6162\u3002<\/p>\n\n\n
\npackage main\nimport (\n\t"fmt"\n\t"io"\n\t"os"\n\t"github.com\/at-wat\/ebml-go"\n)\n\/\/ SubtitleTrack \u5b58\u50a8 MKV \u5b57\u5e55\u8f68\u9053\u4fe1\u606f\ntype SubtitleTrack struct {\n\tNumber   uint64\n\tLanguage string\n\tName     string\n\tCodecID  string\n}\nfunc main() {\n\tif len(os.Args) < 2 {\n\t\tpanic("\u8bf7\u63d0\u4f9b\u6587\u4ef6\u8def\u5f84")\n\t}\n\tfilePath := os.Args[1]\n\tf, err := os.Open(filePath)\n\tif err != nil {\n\t\tpanic(err)\n\t}\n\tdefer f.Close()\n\theadReader := io.NewSectionReader(f, 0, 100*1024) \/\/ \u8bfb\u53d6\u524d 100k\uff0c\u901a\u5e38\u8db3\u591f\u5305\u542b\u5934\u90e8\u4fe1\u606f\n\ttracks, err := listMKVSubtitles(headReader)\n\tif err != nil {\n\t\tfmt.Printf("\u89e3\u6790\u5217\u8868\u65f6\u51fa\u9519 (\u53ef\u80fd\u662f\u6570\u636e\u4e0d\u8db3): %v\\n", err)\n\t}\n\tif len(tracks) == 0 {\n\t\tfmt.Println("\u672a\u53d1\u73b0\u5185\u7f6e\u5b57\u5e55\u8f68\u9053")\n\t\treturn\n\t}\n\tfmt.Println("--- \u53d1\u73b0\u5b57\u5e55\u5217\u8868 ---")\n\tfor _, t := range tracks {\n\t\tfmt.Printf("[%d] \u8bed\u8a00: %s, \u683c\u5f0f: %s, \u540d\u79f0: %s\\n", t.Number, t.Language, t.CodecID, t.Name)\n\t}\n}\nfunc listMKVSubtitles(r io.ReadSeeker) ([]SubtitleTrack, error) {\n\tr.Seek(0, io.SeekStart)\n\tvar header struct {\n\t\tSegment struct {\n\t\t\tTracks struct {\n\t\t\t\tTrackEntry []struct {\n\t\t\t\t\tTrackNumber uint64 `ebml:"TrackNumber"`\n\t\t\t\t\tTrackType   uint64 `ebml:"TrackType"`\n\t\t\t\t\tLanguage    string `ebml:"Language"`\n\t\t\t\t\tName        string `ebml:"Name"`\n\t\t\t\t\tCodecID     string `ebml:"CodecID"`\n\t\t\t\t} `ebml:"TrackEntry"`\n\t\t\t} `ebml:"Tracks"`\n\t\t} `ebml:"Segment"`\n\t}\n\t\/\/ Unmarshal \u4f1a\u5c1d\u8bd5\u89e3\u6790\u6574\u4e2a\u7ed3\u6784\uff0c\u4f46\u5728\u8bfb\u53d6\u5b8c Tracks \u540e\u6211\u4eec\u5c31\u53ef\u4ee5\u505c\u6b62\n\t\/\/ \u5373\u4f7f\u56e0\u4e3a\u6587\u4ef6\u4e0d\u5b8c\u6574\u62a5\u9519\uff0cheader \u91cc\u7684\u4fe1\u606f\u901a\u5e38\u4e5f\u5df2\u7ecf\u586b\u5145\u4e86\n\terr := ebml.Unmarshal(r, &header)\n\tvar result []SubtitleTrack\n\tfor _, t := range header.Segment.Tracks.TrackEntry {\n\t\tif t.TrackType == 17 { \/\/ 17 \u662f MKV \u6807\u51c6\u4e2d\u7684 Subtitle \u7c7b\u578b\n\t\t\tresult = append(result, SubtitleTrack{\n\t\t\t\tNumber:   t.TrackNumber,\n\t\t\t\tLanguage: t.Language,\n\t\t\t\tName:     t.Name,\n\t\t\t\tCodecID:  t.CodecID,\n\t\t\t})\n\t\t}\n\t}\n\treturn result, err\n}\n<\/pre><\/div>\n\n\n