fix646_NikonAF22

[clanmills]

There is a long discussion in #646 about this issue and my investigation into the how the makernotes are decoded. I will summarise here. Fix #646

History

The tag for Nikon's AutoFocus data is 0x00b7

Nikon encode their version of tag in the first 4 bytes. There was a 40 byte version of AutoFocus which decodes as Exif.NikonAf2.XXX. This new version (1.01) is 84 bytes in length and decoded as Exif.NikonAf22.XXX.

The two versions (NikonAF2 and NikonAF22) are now encoded as a set with the selector in tiffimage_int.cpp

    extern const ArraySet nikonAf2Set[] = {
        { nikonAf21Cfg, nikonAf21Def, EXV_COUNTOF(nikonAf21Def) },
        { nikonAf22Cfg, nikonAf22Def, EXV_COUNTOF(nikonAf22Def) },
    };

The binary layout of the record is defined in tiff image_int.cpp. For example, AF22 is:

    extern const ArrayCfg nikonAf22Cfg = {
        nikonAf22Id,      // Group for the elements
        littleEndian,     // Byte order
        ttUndefined,      // Type for array entry
        notEncrypted,     // Not encrypted
        false,            // No size element
        true,             // Write all tags
        true,            // Concatenate gaps
        { 0, ttUnsignedByte,  1 }
    };
    //! Nikon Auto Focus 22 binary array - definition
    extern const ArrayDef nikonAf22Def[] = {
        {  0, ttUndefined,     4 }, // Version
        {  4, ttUnsignedByte,  1 }, // ContrastDetectAF
        {  5, ttUnsignedByte,  1 }, // AFAreaMode
        {  6, ttUnsignedByte,  1 }, // PhaseDetectAF
        {  7, ttUnsignedByte,  1 }, // PrimaryAFPoint
        {  8, ttUnsignedByte,  7 }, // AFPointsUsed
        { 70, ttUnsignedShort, 1 }, // AFImageWidth
        { 72, ttUnsignedShort, 1 }, // AFImageHeight
        { 74, ttUnsignedShort, 1 }, // AFAreaXPosition
        { 76, ttUnsignedShort, 1 }, // AFAreaYPosition
        { 78, ttUnsignedShort, 1 }, // AFAreaWidth
        { 80, ttUnsignedShort, 1 }, // AFAreaHeight
    };

The two versions of the data are encoded in tiffimage_int.cpp

        { Tag::root, nikonAf21Id,      nikon3Id,         0x00b7    },
        { Tag::root, nikonAf22Id,      nikon3Id,         0x00b7    },

Binary Selector

The code to determine which version is decoded is in tiffimage_int.cpp

       {    0x00b7, nikon3Id,         EXV_COMPLEX_BINARY_ARRAY(nikonAf2Set, nikonAf2Selector) },

When the tiffvisitor encounters 0x00b7, he calls nikonAf2Selector() to return the index of the binary array to be used. By default it returns 0 (the existing nikonAf21Id). If the tag length is 84, he returns 1 for nikonAf21Id

    int nikonAf2Selector(uint16_t tag, const byte* /*pData*/, uint32_t size, TiffComponent* const /*pRoot*/)
    {
        int result = tag == 0x00b7 ? 0 : -1 ;
        if (result > -1 && size == 84 ) {
            result = 1;
        }
        return result;
    }

The decoder

EXV_CALL_MEMBER_FN(*this, decoderFct)(object);

This function understands how to decode byte-by-byte from const ArrayDef into the Exiv2 tag/values such as Exif.NikonAF22.AFAreaYPosition which it stores in the ExifData vector.

This is ingenious magic. I'll revisit/edit this explanation in the next few days when I have more time to explain this with more clarity.

Test files

_DSC8437.exv and test_issue_646.py Boiler plate. The .exv was extracted from a sample test files on Nikon's web site (location in #636)

[codecov]

Codecov Report

Merging #900 into 0.27-maintenance will increase coverage by 0.01%.
The diff coverage is 100%.

@@                 Coverage Diff                  @@
##           0.27-maintenance     #900      +/-   ##
====================================================
+ Coverage             62.86%   62.88%   +0.01%     
====================================================
  Files                   156      156              
  Lines                 21598    21606       +8     
====================================================
+ Hits                  13578    13586       +8     
  Misses                 8020     8020

Impacted Files	Coverage Δ
src/tiffvisitor_int.cpp	87.31% <ø> (ø)	⬆️
src/tiffimage_int.cpp	91.76% <ø> (ø)	⬆️
src/makernote_int.hpp	50% <ø> (ø)	⬆️
src/tags_int.cpp	86.41% <ø> (ø)	⬆️
src/tags_int.hpp	76.92% <ø> (ø)	⬆️
src/makernote_int.cpp	72.83% <100%> (+0.31%)	⬆️
src/nikonmn_int.cpp	58.73% <100%> (+0.21%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1eab4e0...db3a260. Read the comment docs.

[piponazo]

Thanks for the contribution!

Although I could not invest much time following the previous discussions, most of the changes look good to me. I just made a bunch of minor comments and questions.

Pull request has been modified.

Pull request has been modified.

[piponazo]

@D4N do you want to take a look to the PR before we merge it ?

[clanmills]

Thanks for dealing with this, @piponazo . I've been working very hard yesterday and today on 582 (Andreas Sneider's request to support Sony FocusPosition), then I want to get Exiv2 0.27.2 RC1 done and released by Thursday before going to Normandy on Friday to run in the D-Day Half Marathon this weekend.

I'd like to discuss 882 FileIO::eof() calling stat with you a little more. I can't remember what I want to say to you about this, however it's on my TODO list. If I don't get that done this week, I will deal with it next week and we can decide if it's going in 0.27.2 or more-likely defer for 0.27.3.

[D4N]

Overall this looks good to me. I've added just a few comments

[D4N]

[Suggestion] For consistency I'd suggest to call list & enumeration element *Af2 instead of *Af21.

[clanmills]

You're right about this. We have two things which are very similar nikonAf2ID (AutoFocus version 1.00) and nikonAf22Id (AutoFocus version 1.01).

[clanmills]

Dan: As always you are very sharp eyed. When I developed the code, I changed nikonAf2Id to nikonAf21Id and cut'n'pasted/morphed nikonAf21Id into nikonAf22Id. Later I discovered this broke the test suite, because there's a test that uses Exif.NikonAf2.something. I hacked that table to pass the test suite. I've restored nikonAf2Id.

It's really hard to understand the tiff parsing code. It's very abstract. I've been trying for 2 days to fix #582 and haven't got it work yet. I'll defer #582 until v0.27.3.

Pull request has been modified.

Pull request has been modified.

[mallman]

As the originator of the request for and a beneficiary of this functionality, I want to offer my (long overdue) and effusive thanks to @clanmills and collaborators on developing and merging this PR.

I'm working to test support on my end, building 0.27.2-RC2, but having some problems with the build. I'll open an issue if I can't resolve them myself.

[mallman]

I'm working to test support on my end, building 0.27.2-RC2, but having some problems with the build. I'll open an issue if I can't resolve them myself.

First of all, I was able to build successfully by updating my version of Cmake from 3.13.2 to 3.14.5. Seems like such a minor update, but it worked.

Second, I've verified that the exifprint command is producing what appear to be correct values for the Exif.NikonAf22.* keys when I run against my D850 NEF files.

Many thanks again!