Extracting/reading PNG iTXt chunks with Exiv2

Added by Arnold Wiegert about 5 years ago

Being a newb to Exiv2, I have been trying to find a way to check for and extract or read the various chunks, particularly the iTXt chunks, in a PNG, file using the current Exiv2 library.
Any comments or pointer will much appreciated.
TIA

Replies (4)

RE: Extracting/reading PNG iTXt chunks with Exiv2 - Added by Robin Mills about 5 years ago

Arnold

Yes, you can do this with the -pS (printStructure in v0.25) and -pR (printRecursive in trunk/v0.26) options.

951 rmills@rmillsmbp:~/gnu/exiv2/trunk $ exiv2 -pS test/data/imagemagick.png 
STRUCTURE OF PNG FILE: test/data/imagemagick.png
 address | index | chunk_type |  length | data
       8 |     0 |       IHDR |      13 | ...@........
      33 |     1 |       iCCP |    1404 | icc..x...i8.......af\...w_3...+
    1449 |     2 |       sBIT |       3 | ...
    1464 |     3 |       zTXt |      87 | Software..x...A.. ......B....}.N
    1563 |     4 |       tEXt |   24482 | Raw profile type exif..exif.   1
   26057 |     5 |       tEXt |     471 | Raw profile type iptc..iptc.    
   26540 |     6 |       IDAT |    8192 | x...Y.$Wv&v.{.{l.T.......[w.=m$G
   34744 |     7 |       IDAT |    8192 | .4X.y.AR...4....:Ue..U.|1..:..D.
   42948 |     8 |       IDAT |    8192 | 'g.!... ...n...s..Jdz.........6.
   51152 |     9 |       IDAT |    8192 | ........k....CY/75I..u;.. .z...}
   59356 |    10 |       IDAT |    8192 | .f>..]....UKqD2s.(.q....=x.l.\V.
   67560 |    11 |       IDAT |    8192 | .i.{!!B0...C!4.p..`D g`.........
   75764 |    12 |       IDAT |    8192 | .*.].4..Q..}(9...S0&.......T.9.+
   83968 |    13 |       IDAT |    8192 | ..k...6....g.1..}.].&.H........
   92172 |    14 |       IDAT |    8192 | .j..S.........z..!U.G0*.m%..09<,
  100376 |    15 |       IDAT |    8192 | .....t.>!.....6^.<..;..?$I..M.9
  108580 |    16 |       IDAT |    8192 | W.&5.5J........FW`....3.N.9Pk;.s
  116784 |    17 |       IDAT |    8192 | .....d.z".`...v=g-..-.c8...Z.5.x
  124988 |    18 |       IDAT |    8192 | .."...o<&."....1M....1&. ..5....
  133192 |    19 |       IDAT |    8192 | k........!..B*.....\*.(!..0.s..B
  141396 |    20 |       IDAT |    3346 | .Y.L@I$M.Z[.0A ...K#.t.0+.G(.jX.
  144754 |    21 |       IEND |       0 | 
952 rmills@rmillsmbp:~/gnu/exiv2/trunk $

I don't know for which purpose you are interested in the iTXt chunks. Some can be extracted by exiv2. For example, the following command will extract an ICC profile to stdout.

$ exiv2 -pC foo.png

The following command will extract XMP (xml) to stdout:

$ exiv2 -pX foo.png

If you have other chunks of interest, you can write a script to parse the output of $ exiv2 -pS foo.png to obtain the offset and length of a chunk which can then be extracted with dd. This is discussed in relation to TIFF files in this article: http://dev.exiv2.org/projects/exiv2/wiki/The_Metadata_in_TIFF_files

Print Recursive -pR on the trunk and heading for v0.26 shows the same information and descends into data structures it recognises.

953 rmills@rmillsmbp:~/gnu/exiv2/trunk $ exiv2 -pR test/data/imagemagick.pngSTRUCTURE OF PNG FILE: test/data/imagemagick.png length | data IHDR |      13 | ...@........ iCCP |    1404 | icc..x...i8.......af\...w_3...+ 3 | ... zTXt |      87 | Software..x...A.. ......B....}.N ( libpng version 1.2.8 - December 3, 2004 (header) ) tEXt |   24482 | Raw profile type exif..exif.   1 |      type |    count |    offset | value |     SLONG |        1 |       320 | 320 |     SLONG |        1 |       211 | 211 |     ASCII |       18 |       146 | NIKON CORPORATION |     ASCII |       10 |       164 | NIKON D70 |     SHORT |        1 |     65536 | 1 |  RATIONAL |        1 |       174 | 174/0 |  RATIONAL |        1 |       182 | 182/0 |     SHORT |        1 |    131072 | 2 |     ASCII |       18 |       190 | digiKam-0.9.0-svn |     ASCII |       20 |       208 | 2006:02:04 16:09:30 |      LONG |        1 |       228 | 228 |      type |    count |    offset | value |  RATIONAL |        1 |       546 | 546/0 |  RATIONAL |        1 |       554 | 554/0 |     SHORT |        1 |    262144 | 4 |     SHORT |        1 |  13107200 | 200 |     ASCII |       20 |       562 | 2006:02:04 16:09:30 |     ASCII |       20 |       582 | 2006:02:04 16:09:30 |  RATIONAL |        1 |       602 | 602/0 |  RATIONAL |        1 |       610 | 610/0 |  RATIONAL |        1 |       618 | 618/0 |  RATIONAL |        1 |       626 | 626/0 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 |  RATIONAL |        1 |       634 | 634/0 | UNDEFINED |     6989 |       642 | Nikon.....MM.*.....+........0210 ... FILE (MM): MemIo |      type |    count |    offset | value | UNDEFINED |        4 | 808595760 | 0210 |     SHORT |        2 |       200 | 0 200 |     ASCII |        8 |       530 | RAW |     ASCII |       13 |       538 | AUTO |     ASCII |        7 |       551 | NORMAL |     ASCII |        7 |       558 | AF-S |     ASCII |       13 |       565 | NORMAL |     ASCII |       13 |       578 | ProcessingSoftware        |    SSHORT |        1 |         0 | 0 | UNDEFINED |        4 |     67072 | ... | UNDEFINED |        4 |     68608 | ... |      LONG |        1 |      7006 | 7006 | UNDEFINED |        4 |     67072 | ... GPSDestLatitudeRef        |     SHORT |        2 |       200 | 0 200 | UNDEFINED |        4 |     67072 | ... | UNDEFINED |        4 |     67072 | ... GPSDestDistanceRef        | SRATIONAL |        1 |       591 | 591/0 |     ASCII |        9 |       599 | NORMAL |      BYTE |        1 | 100663296 | . |  RATIONAL |        4 |       608 | 180/0 10/0 500/0 10/0 |      BYTE |        1 |         0 | | UNDEFINED |        4 |  16777217 | .... |     SHORT |        1 |         0 | 0 | UNDEFINED |        4 |1208028160 | H.. | UNDEFINED |     4160 |       640 | I.............................. ... |     ASCII |        9 |      4800 | MODE2 |     ASCII |       12 |      4809 | NATURAL | UNDEFINED |      465 |      4821 | 0103........................... ... |    SSHORT |        1 |         0 | 0 |     SHORT |        1 |     65536 | 1 |     ASCII |        5 |      5286 | OFF | UNDEFINED |     1412 |      5291 | D..H.H.H.H...................... ... | UNDEFINED |      140 |      6703 | 0103.................:...z...... ... | UNDEFINED |       31 |      6843 | ............................... |     SHORT |        2 |  99615728 | 1520 1008 |  RATIONAL |        2 |      6874 | 3/296 1/0 |     ASCII |       21 |      6890 | .........2........... | UNDEFINED |        4 | 808595504 | 0200 |      LONG |        1 |      5670 | 5670 | UNDEFINED |       20 |      6911 | ..........a~NIKON CO |     ASCII |       16 |      6931 | RPORATION.NIKON |     ASCII |       16 |      6947 | D70....,......., |     ASCII |       16 |      6963 | ....Bibble 4.5. |     SLONG |        1 |       320 | 320 |     SLONG |        1 |       211 | 211 |      BYTE |        1 |  33554432 | . |      BYTE |        1 |  16777216 | . |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 FocalLengthIn35mmFilm     |     SHORT |        1 |   4915200 | 75 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 SubjectDistanceRange      |     SHORT |        1 |         0 | 0 |     SHORT |        1 |    393216 | 6 |      LONG |        1 |      7673 | 7673 JPEGInterchangeFormatLeng |      LONG |        1 |      4376 | 4376 tEXt |     471 | Raw profile type iptc..iptc. | Length | Data |      4 | ovni |      1 | 0 |      3 | I - |     13 | SPCL Specials |      4 | E.T. |      9 | Spielberg |     21 | L'ovni de la Bastille |      5 | Paris |     13 | ..le de Franc |     12 | FRA - France |      3 | 750 |     10 | Y'en a pas TransmissionReference    |      2 | 36 |      8 | Chez moi |      8 | 20060204 |     11 | 160930+0000 |      2 | .. IDAT |    8192 | x...Y.$Wv&v.{.{l.T.......[w.=m$G IDAT |    8192 | .4X.y.AR...4....:Ue..U.|1..:..D. IDAT |    8192 | 'g.!... ...n...s..Jdz.........6. IDAT |    8192 | ........k....CY/75I..u;.. .z...} IDAT |    8192 | .f>..]....UKqD2s.(.q....=x.l.\V. IDAT |    8192 | .i.{!!B0...C!4.p..`D g`......... IDAT |    8192 | .*.].4..Q..}(9...S0&.......T.9.+ IDAT |    8192 | ..k...6....g.1..}.].&.H........ IDAT |    8192 | .j..S.........z..!U.G0*.m%..09<, IDAT |    8192 | .....t.>!.....6^.<..;..?$I..M.9 IDAT |    8192 | W.&5.5J........FW`....3.N.9Pk;.s IDAT |    8192 | .....d.z".`...v=g-..-.c8...Z.5.x IDAT |    8192 | .."...o<&."....1M....1&. ..5.... IDAT |    8192 | k........!..B*.....\*.(!..0.s..B IDAT |    3346 | .Y.L@I$M.Z[.0A ...K#.t.0+.G(.jX. 0 | test/data/imagemagick.png length | data IHDR |      13 | ...@........ iCCP |    1404 | icc..x...i8.......af\...w_3...+ 3 | ... zTXt |      87 | Software..x...A.. ......B....}.N ( libpng version 1.2.8 - December 3, 2004 (header) ) tEXt |   24482 | Raw profile type exif..exif.   1 |      type |    count |    offset | value |     SLONG |        1 |       320 | 320 |     SLONG |        1 |       211 | 211 |     ASCII |       18 |       146 | NIKON CORPORATION |     ASCII |       10 |       164 | NIKON D70 |     SHORT |        1 |     65536 | 1 |  RATIONAL |        1 |       174 | 174/0 |  RATIONAL |        1 |       182 | 182/0 |     SHORT |        1 |    131072 | 2 |     ASCII |       18 |       190 | digiKam-0.9.0-svn |     ASCII |       20 |       208 | 2006:02:04 16:09:30 |      LONG |        1 |       228 | 228 |      type |    count |    offset | value |  RATIONAL |        1 |       546 | 546/0 |  RATIONAL |        1 |       554 | 554/0 |     SHORT |        1 |    262144 | 4 |     SHORT |        1 |  13107200 | 200 |     ASCII |       20 |       562 | 2006:02:04 16:09:30 |     ASCII |       20 |       582 | 2006:02:04 16:09:30 |  RATIONAL |        1 |       602 | 602/0 |  RATIONAL |        1 |       610 | 610/0 |  RATIONAL |        1 |       618 | 618/0 |  RATIONAL |        1 |       626 | 626/0 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 |  RATIONAL |        1 |       634 | 634/0 | UNDEFINED |     6989 |       642 | Nikon.....MM.*.....+........0210 ... FILE (MM): MemIo |      type |    count |    offset | value | UNDEFINED |        4 | 808595760 | 0210 |     SHORT |        2 |       200 | 0 200 |     ASCII |        8 |       530 | RAW |     ASCII |       13 |       538 | AUTO |     ASCII |        7 |       551 | NORMAL |     ASCII |        7 |       558 | AF-S |     ASCII |       13 |       565 | NORMAL |     ASCII |       13 |       578 | ProcessingSoftware        |    SSHORT |        1 |         0 | 0 | UNDEFINED |        4 |     67072 | ... | UNDEFINED |        4 |     68608 | ... |      LONG |        1 |      7006 | 7006 | UNDEFINED |        4 |     67072 | ... GPSDestLatitudeRef        |     SHORT |        2 |       200 | 0 200 | UNDEFINED |        4 |     67072 | ... | UNDEFINED |        4 |     67072 | ... GPSDestDistanceRef        | SRATIONAL |        1 |       591 | 591/0 |     ASCII |        9 |       599 | NORMAL |      BYTE |        1 | 100663296 | . |  RATIONAL |        4 |       608 | 180/0 10/0 500/0 10/0 |      BYTE |        1 |         0 | | UNDEFINED |        4 |  16777217 | .... |     SHORT |        1 |         0 | 0 | UNDEFINED |        4 |1208028160 | H.. | UNDEFINED |     4160 |       640 | I.............................. ... |     ASCII |        9 |      4800 | MODE2 |     ASCII |       12 |      4809 | NATURAL | UNDEFINED |      465 |      4821 | 0103........................... ... |    SSHORT |        1 |         0 | 0 |     SHORT |        1 |     65536 | 1 |     ASCII |        5 |      5286 | OFF | UNDEFINED |     1412 |      5291 | D..H.H.H.H...................... ... | UNDEFINED |      140 |      6703 | 0103.................:...z...... ... | UNDEFINED |       31 |      6843 | ............................... |     SHORT |        2 |  99615728 | 1520 1008 |  RATIONAL |        2 |      6874 | 3/296 1/0 |     ASCII |       21 |      6890 | .........2........... | UNDEFINED |        4 | 808595504 | 0200 |      LONG |        1 |      5670 | 5670 | UNDEFINED |       20 |      6911 | ..........a~NIKON CO |     ASCII |       16 |      6931 | RPORATION.NIKON |     ASCII |       16 |      6947 | D70....,......., |     ASCII |       16 |      6963 | ....Bibble 4.5. |     SLONG |        1 |       320 | 320 |     SLONG |        1 |       211 | 211 |      BYTE |        1 |  33554432 | . |      BYTE |        1 |  16777216 | . |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 FocalLengthIn35mmFilm     |     SHORT |        1 |   4915200 | 75 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 |     SHORT |        1 |         0 | 0 SubjectDistanceRange      |     SHORT |        1 |         0 | 0 |     SHORT |        1 |    393216 | 6 |      LONG |        1 |      7673 | 7673 JPEGInterchangeFormatLeng |      LONG |        1 |      4376 | 4376 tEXt |     471 | Raw profile type iptc..iptc. | Length | Data |      4 | ovni |      1 | 0 |      3 | I - |     13 | SPCL Specials |      4 | E.T. |      9 | Spielberg |     21 | L'ovni de la Bastille |      5 | Paris |     13 | ..le de Franc |     12 | FRA - France |      3 | 750 |     10 | Y'en a pas TransmissionReference    |      2 | 36 |      8 | Chez moi |      8 | 20060204 |     11 | 160930+0000 |      2 | .. IDAT |    8192 | x...Y.$Wv&v.{.{l.T.......[w.=m$G IDAT |    8192 | .4X.y.AR...4....:Ue..U.|1..:..D. IDAT |    8192 | 'g.!... ...n...s..Jdz.........6. IDAT |    8192 | ........k....CY/75I..u;.. .z...} IDAT |    8192 | .f>..]....UKqD2s.(.q....=x.l.\V. IDAT |    8192 | .i.{!!B0...C!4.p..`D g`......... IDAT |    8192 | .*.].4..Q..}(9...S0&.......T.9.+ IDAT |    8192 | ..k...6....g.1..}.].&.H........ IDAT |    8192 | .j..S.........z..!U.G0*.m%..09<, IDAT |    8192 | .....t.>!.....6^.<..;..?$I..M.9 IDAT |    8192 | W.&5.5J........FW`....3.N.9Pk;.s IDAT |    8192 | .....d.z".`...v=g-..-.c8...Z.5.x IDAT |    8192 | .."...o<&."....1M....1&. ..5.... IDAT |    8192 | k........!..B*.....\*.(!..0.s..B IDAT |    3346 | .Y.L@I$M.Z[.0A ...K#.t.0+.G(.jX. 0 |

RE: Extracting/reading PNG iTXt chunks with Exiv2 - Added by Arnold Wiegert about 5 years ago

Thank you, Robin,
but I should have been clearer, I want to do this from within C_++ code, not the command line.

RE: Extracting/reading PNG iTXt chumnks with Exiv2 - Added by Robin Mills about 5 years ago

The API that performs this work in PngImage::printStructure(). The output of -pS is captured in a stream which you can easily parse. If you wish to process the chunks in some complex way, you may wish to copy the code from printStructure() to your application.

   void PngImage::printStructure(std::ostream& out, PrintStructureOption option, int depth)
   {
        if (io_->open() != 0) {
            throw Error(9, io_->path(), strError());
        }
        IoCloser closer(*io_);
        // Ensure that this is the correct image type
        if (!isPngType(*io_, true)) {
            if (io_->error() || io_->eof()) throw Error(14);
            throw Error(3, "PNG");
        }

        char    chType[5];
        chType[0]=0;
        chType[4]=0;

        if ( option == kpsBasic || option == kpsXMP || option == kpsIccProfile || option == kpsRecursive ) {

            const std::string xmpKey  = "XML:com.adobe.xmp";
            const std::string exifKey = "Raw profile type exif";
            const std::string iptcKey = "Raw profile type iptc";
            const std::string iccKey  = "icc";
            const std::string softKey = "Software";

            if ( option == kpsBasic || option == kpsRecursive ) {
                out << "STRUCTURE OF PNG FILE: " << io_->path() << std::endl;
                out << " address | index | chunk_type |  length | data" << std::endl;
            }

            long       index   = 0;
            const long imgSize = io_->size();
            DataBuf    cheaderBuf(8);

            while( !io_->eof() && ::strcmp(chType,"IEND") ) {
                size_t address = io_->tell();

                std::memset(cheaderBuf.pData_, 0x0, cheaderBuf.size_);
                long bufRead = io_->read(cheaderBuf.pData_, cheaderBuf.size_);
                if (io_->error()) throw Error(14);
                if (bufRead != cheaderBuf.size_) throw Error(20);

                // Decode chunk data length.
                uint32_t dataOffset = Exiv2::getULong(cheaderBuf.pData_, Exiv2::bigEndian);
                for (int i = 4; i < 8; i++) {
                    chType[i-4]=cheaderBuf.pData_[i];
                }

                // test that we haven't hit EOF, or wanting to read excessive data
                long restore = io_->tell();
                if(  restore == -1
                ||  dataOffset > uint32_t(0x7FFFFFFF)
                ||  static_cast<long>(dataOffset) > imgSize - restore
                ){
                    throw Exiv2::Error(14);
                }

                // format output
                uint32_t    blen = dataOffset > 32 ? 32 : dataOffset ;
                std::string dataString ;
                DataBuf buff(blen);
                io_->read(buff.pData_,blen);
                io_->seek(restore, BasicIo::beg);
                dataString  = Internal::binaryToString(buff, blen);

                if ( option == kpsBasic || option == kpsRecursive )
                    out << Internal::stringFormat("%8d | %5d | %10s |%8d | " 
                              ,(uint32_t)address, index++,chType,dataOffset)
                                    << dataString << std::endl;

                // chunk type
                bool tEXt  = std::strcmp(chType,"tEXt")== 0;
                bool zTXt  = std::strcmp(chType,"zTXt")== 0;
                bool iCCP  = std::strcmp(chType,"iCCP")== 0;
                bool iTXt  = std::strcmp(chType,"iTXt")== 0;

                // for XMP, ICC etc: read and format data
                bool bXMP  = option == kpsXMP        && findi(dataString,xmpKey)==0;
                bool bICC  = option == kpsIccProfile && findi(dataString,iccKey)==0;
                bool bExif = option == kpsRecursive  && findi(dataString,exifKey)==0;
                bool bIptc = option == kpsRecursive  && findi(dataString,iptcKey)==0;
                bool bSoft = option == kpsRecursive  && findi(dataString,softKey)==0;
                bool bDump = bXMP || bICC || bExif || bIptc || bSoft ;

                if( bDump ) {
                    DataBuf   dataBuf;
                    byte*     data     = new byte[dataOffset+1];
                    data[dataOffset]=0;
                    io_->read(data,dataOffset);
                    io_->seek(restore, BasicIo::beg);
                    uint32_t    name_l = (uint32_t) std::strlen((const char*)data)+1; // leading string length
                    uint32_t     start = name_l;

                    // decode the chunk
                    bool bGood = false;
                    if ( tEXt ) {
                        bGood = tEXtToDataBuf(data+name_l,dataOffset-name_l,dataBuf);
                    }
                    if ( zTXt || iCCP ) {
                        name_l++ ; // +1 = 'compressed' flag
                        bGood = zlibToDataBuf(data+name_l,dataOffset-name_l,dataBuf);
                    }
                    if ( iTXt ) {
                        while ( data[start] == 0 && start < dataOffset ) start++; // crawl over the '\0' bytes between XML:....\0\0<xml stuff
                        data[dataOffset]=0;                 // ensure the XML is nul terminated
                        bGood = (start+3) < dataOffset ;    // good if not a nul chunk
                    }

                    // format is content dependent
                    if ( bGood ) {
                        if ( bXMP ) {
                            out <<  data+start;             // output the xml
                        }
                        if ( bExif ) {
                            const char* bytes = (const char*) dataBuf.pData_;
                            uint32_t    l     = (uint32_t) std::strlen(bytes)+2;
                            // create a copy on write memio object with the data, then print the structure
                            BasicIo::AutoPtr p = BasicIo::AutoPtr(new MemIo(dataBuf.pData_+l,dataBuf.size_-l));
                            TiffImage::printTiffStructure(*p,out,option,depth);
                        }

                        if ( bSoft ) {
                            const char* bytes = (const char*) dataBuf.pData_;
                            uint32_t    l     = (uint32_t) std::strlen(bytes)+2;
                            // create a copy on write memio object with the data, then print the structure
                            BasicIo::AutoPtr p = BasicIo::AutoPtr(new MemIo(dataBuf.pData_+l,dataBuf.size_-l));
                            out << Internal::indent(depth) << (const char*) buff.pData_ << ": " << (const char*) dataBuf.pData_ << std::endl;
                        }

                        if ( bICC ) {
                            out.write((const char*) dataBuf.pData_,dataBuf.size_);
                        }

                        if ( bIptc ) {
                            IptcData::printStructure(out,dataBuf.pData_,dataBuf.size_,depth);
                        }
                    }
                    delete[] data;
                }
                io_->seek(dataOffset + 4 , BasicIo::cur);
                if (io_->error()) throw Error(14);
            }
        }
    }

RE: Extracting/reading PNG iTXt chunks with Exiv2 - Added by Arnold Wiegert about 5 years ago

Again, many thanks for this clarification.
For some reason I had expected to find this metadata under metadata in the Exiv2 code :-)
I'll give this sample code a try.

(1-4/4)

Project

General

Profile

Exiv2