tiff file tags not saved

I am able to create and save tiff files but the tags I create don't appear to be saved when i view the "properties" from Windows file explorer nor am I able to open the same file with

file = Tiff(tiffFileName,"r");  

Curiously, the data is there - when I drag the file from the file explorer to the command line and let the import fcn do its thing, the image data there.

Here's the code I'm using to write the file:

t =  Tiff([dataDir filesep saveFileName],"w");
setTag(t,"ImageLength",1024);
setTag(t,"ImageWidth",1024);
setTag(t,"Photometric",Tiff.Photometric.MinIsBlack);
setTag(t,"Compression",Tiff.Compression.None);
setTag(t,"BitsPerSample", 64);
setTag(t,"SampleFormat", Tiff.SampleFormat.IEEEFP);
setTag(t,"SamplesPerPixel", 1);
setTag(t,'ImageDescription', myMetaData);
write(t, data)

Answer 1

[steve solomon]

Yup, it was the ordering of the tiff tags. When i write the tags per the order per the TIFF specs found in the Library of Congress website, all the tags appeared.

Happy trails and thanks to all for the help & feedback.

Answer 2

[DGM]

I think the main problem here is that you're not closing the file. There are also some annoyances you'll run into if you're trying to set tags this way using setTag(). If you do it this way, you need to set the tags in a certain order in order for them to work right. Certain tags need to be set before others can be set. Alternatively, you can put all the tags in a struct in any order is convenient, and then you can write that tag struct to the Tiff object at once. That's how I'd do it.

This is mostly copy-pasted from another answer I posted. It's more than what you need, but it's generalized enough to cover most cases while being flexible. FWIW, the Deflate compression is lossless. Given that we're making giant DPFP TIFF files that probably nothing else can read, there's little risk in complicating the affair, but fair gains to be had over using uncompressed data. Like anything, it depends on the image content.

% a DPFP test image
inpict = imread('cameraman.tif');
inpict = im2double(inpict); 

% metadata
myMetaData = 'banana';

% output file name
filename = 'test.tiff';

% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% image can be I/RGB, 8/16/32b uint/int or 32/64b float
% get to work
sz = size(inpict,1:3);
inclass = class(inpict); 
t = Tiff(filename, 'w');
tagstruct.ImageLength = sz(1);
tagstruct.ImageWidth = sz(2);
tagstruct.Compression = Tiff.Compression.AdobeDeflate;
tagstruct.SamplesPerPixel = sz(3);
tagstruct.PlanarConfiguration = Tiff.PlanarConfiguration.Chunky;

% I/RGB?
% i'm choosing to presume that 3ch inputs are RGB
if sz(3) == 1
    tagstruct.Photometric = Tiff.Photometric.MinIsBlack;
elseif sz(3) == 3
    tagstruct.Photometric = Tiff.Photometric.RGB;
end

% what is the numeric type?
if isinteger(inpict) && inclass(1) == 'u'
    tagstruct.SampleFormat = Tiff.SampleFormat.UInt;
elseif isinteger(inpict) && inclass(1) == 'i'
    tagstruct.SampleFormat = Tiff.SampleFormat.Int;
elseif isfloat(inpict)
    tagstruct.SampleFormat = Tiff.SampleFormat.IEEEFP;
else
    error('this demo only supports integer and float inputs')
end

% what's the sample width?
switch inclass
    case {'uint8','int8'}
        tagstruct.BitsPerSample = 8;
    case {'uint16','int16'}
        tagstruct.BitsPerSample = 16;
    case {'uint32','int32','single'}
        tagstruct.BitsPerSample = 32;
    case {'uint64','int64'}
        error('LibTIFF does not support 64b outputs except for double-precision floats')
    case 'double'
        tagstruct.BitsPerSample = 64;
    otherwise
        error('this demo only supports integer and float inputs')
end

% additional tags
tagstruct.ImageDescription = myMetaData;

% write everything and close the file
t.setTag(tagstruct);
t.write(inpict);
t.close();

% read the image back
newpict = imread(filename);
isequal(inpict,newpict) % test that it's the same

ans = logical
   1

% check the tags
imfinfo(filename)

ans = struct with fields:
                     Filename: '/users/mss.system.wrwxh/test.tiff'
                  FileModDate: '24-Jun-2025 22:24:40'
                     FileSize: 73863
                       Format: 'tif'
                FormatVersion: []
                        Width: 256
                       Height: 256
                     BitDepth: 64
                    ColorType: 'grayscale'
              FormatSignature: [73 73 42 0]
                    ByteOrder: 'little-endian'
               NewSubFileType: 0
                BitsPerSample: 64
                  Compression: 'Deflate'
    PhotometricInterpretation: 'BlackIsZero'
                 StripOffsets: 8
              SamplesPerPixel: 1
                 RowsPerStrip: 256
              StripByteCounts: 73697
                  XResolution: []
                  YResolution: []
               ResolutionUnit: 'Inch'
                     Colormap: []
          PlanarConfiguration: 'Chunky'
                    TileWidth: []
                   TileLength: []
                  TileOffsets: []
               TileByteCounts: []
                  Orientation: 1
            AutoOrientedWidth: 256
           AutoOrientedHeight: 256
                    FillOrder: 1
             GrayResponseUnit: 0.0100
               MaxSampleValue: 1.8447e+19
               MinSampleValue: 0
                 Thresholding: 1
                       Offset: 73706
             ImageDescription: 'banana'
                 SampleFormat: 'IEEE floating point'

EDIT:

It might not be relevant here, but I forgot to mention that signed-integer outputs cannot be RGB. It's a niche concern, but this demo code doesn't check for that. I also made no accomodation for logical images. Take that FWIW.

[Walter Roberson]

By the way:

The image created above by @DGM code, is considered an invalid or corrupt image by MacOS Sequoia 15.5. But imwrite(inpict,'test2.tiff') works. (imwrite() writes it as uint8 instead of as double. imwrite() does not handle single precision or double precision for tiff images.)

[DGM]

Is that because I'm doing something inappropriately, or is it just because the OS tools can't decode a float TIFF? At least on my ancient Linux system, nothing else accepts it either. I've been under the impression that we have to accept that most things (other than maybe MATLAB or ImageJ) will treat it as garbage.

I think in the past, I've suggested that the limited portability of DPFP TIFF should be cause enough to consider that saving data using a MAT-file might be a reasonable and potentially more convenient alternative.

[steve solomon]

Thanks guys, appreciate the quick feedback. I am closing the file, just didn’t include that command in the code snippet. Tried writing the tags to a struct per DGM’s suggestion and that solved one problem, namely it let me open the files from inside Matlab but I still don’t see any of the tags from Windows file explorer.

Went looking for documentation re the ordering of “tag writes” and found tihs on the library of congress website:

The TIFF specification requires that tags be encoded in numerical order and that is the sequence used in the table below.

https://www.loc.gov/preservation/digital/formats/content/tiff_tags.shtml

Will give that a shot and report back.

[DGM]

I have no idea how Windows treats anything, but at least with exiftool, the tags all show up.

>>$ exiftool test.tiff

ExifTool Version Number : 9.46

File Name : test.tiff

Directory : .

File Size : 72 kB

File Modification Date/Time : 2025:06:24 17:23:17-05:00

File Access Date/Time : 2025:06:24 17:23:21-05:00

File Inode Change Date/Time : 2025:06:24 17:23:17-05:00

File Permissions : rw-r--r--

File Type : TIFF

MIME Type : image/tiff

Exif Byte Order : Little-endian (Intel, II)

Image Width : 256

Image Height : 256

Bits Per Sample : 64

Compression : Adobe Deflate

Photometric Interpretation : BlackIsZero

Image Description : banana

Strip Offsets : 8

Samples Per Pixel : 1

Rows Per Strip : 256

Strip Byte Counts : 73697

Planar Configuration : Chunky

Sample Format : Float

Image Size : 256x256

Other tools (e.g. gthumb) may not decode the image data, but the tags are all present.