Compare the Metadata of two Files

[JanK]

Is there an option to compare the Metadata of two files (jpg)?

When I manipulate a file for example with Adobe Bridge many Metadate is added to the file. To see what changes are made I will compare the original file and the edited file with Exiftool. I hope that Exifool can list the differences "deleted" "modified" "added" data to the file.

[Phil Harvey]

I do this all the time, but I use the "diff" utility (standard on Mac/Linux) to do this:

exiftool a.jpg b.jpg -a -G1 -w txt
diff a.txt b.txt

- Phil

[No fine pix with my FinePix]

Hello, I had some problems with differing meta data, so - because I did not found anything - I wrote a small batch (running in XP, Vista). Perhaps you find it helpfull.

Code Select Expand


@echo off
:: Compare_EXIF.cmd Compare EXIF-Meta-Data V1.6
setlocal enableextensions enabledelayedexpansion &  if errorlevel 1 ( echo Can not set local & pause & exit/b 99 )

:: ARGUMENTS
:: either two dirs with same named pics, Wildcards are not possible, choose mask for dir comparison
set WILD=*.jpg
:: or two pics

:: NEEDS
:: exiftool http://www.exiftool.org/
set EF="G:\Programme\GeoSetter\tools\exiftool.exe"
:: Winmerge http://winmerge.org
set WM="G:\Programme\PROGRAMME (USB)\WinMergePortable\WinMergePortable.exe"

:: for DEBUGGING uncomment
:: set DEBCMD=echo ~

:: START BATCH

:: Functions
set EXITE=echo. ^& pause ^& exit/b 99

:: Initialization
cls
if not defined DEBCMD ( set DEBCMD=rem ) else %DEBCMD% Using debug command: %DEBCMD%

:: exiftool commandline
if not exist %EF% ( echo ExifTool %EF% not found % %EXITE% )
:: Instead of excluding here, may use LineFilter in Winmerge
set EF=%EF% -groupHeadings --FileModifyDate
%DEBCMD% exiftool (found) %EF%

:: Winmerge commandline
if not exist %WM% ( echo ExifTool %WM% not found % %EXITE% )
:: /s single instance
:: /e for end with escape
:: /x for message and end if equal
:: /ul /ur don't add to history left & right
:: /wl /wr read only left & right
set WM=%WM% /s /e /x /ul /ur /wl /wr
%DEBCMD% Winmerge (found) %WM%

:: Check Argument/1
set D1=%~dp1
set N1=%~nx1
set ISDIR1=%~a1
set ISDIR1=%ISDIR1:~0,1%
if ~%ISDIR1%~==~d~ (
set D1=%1
set N1=%WILD%
)
%DEBCMD% Arg/1 %1%
%DEBCMD% Path/1 %D1% & %DEBCMD% Name/1 %N1% & %DEBCMD% Attributes/1 %ISDIR1%
if not exist %1 ( echo 1st File/s %1? & %EXITE% )

:: Check Argument/2
set D2=%~dp2
set N2=%~nx2
set ISDIR2=%~a2
set ISDIR2=%ISDIR2:~0,1%
if ~%ISDIR2%~==~d~ (
set D2=%2
set N2=%WILD%
)
%DEBCMD% Arg/2 %2%
%DEBCMD% Path/2 %D2% & %DEBCMD% Name/2 %N2% & %DEBCMD% Attributes/2 %ISDIR2%
if not exist %2 ( echo 2nd File/s %2? & %EXITE% )

:: No Argument/3
if not ~%3~==~~ ( echo Only 2 args, either 2 files or 2 dirs, NO WILDCARDS &  %EXITE% )

:: Check constency of arguments
%DEBCMD% Matched Attributes %ISDIR1%%ISDIR2%
if ~%ISDIR1%%ISDIR2%~==~dd~ ( echo Comparing two directories & goto compare )
if ~%ISDIR1%%ISDIR2%==~--~ ( echo Comparing two files & goto compare )
%EXITE% Can only compare two dirs or two files

:compare
:: Temporary Files
if not defined TEMP (echo TEMP-Directory defined & %EXITE%) else %DEBCMD% Using TEMP: %TEMP%
set T1="%TEMP%\%~n0_1.tmp"
set T2="%TEMP%\%~n0_2.tmp"
%DEBCMD% Temp 1: %T1% & %DEBCMD% Temp 2: %T2%

:: Run first arg
echo.
echo exiftool/1
echo File/s: %N1%
echo in: %D1%
cd /d %D1%
%DEBCMD% "%EF% %N1% > %T1%"
%EF% %N1% > %T1%
if errorlevel 1 ( echo ExifTool/1 Error on %1 & %EXITE% )
if not exist %T1% ( echo Not found: %T1% & %EXITE% )

:: Rund second arg
echo.
echo exiftool/2
echo File/s: %N2%
echo in: %D2%
cd /d %D2%
%DEBCMD% "%EF% %N2% > %T2%"
%EF% %N2% > %T2%
if errorlevel 1 ( echo ExifTool/2 Error on %2 & %EXITE% )
if not exist %T2% ( echo Not found: %T2% & %EXITE% )

:: /dl /dr Description left & right
%DEBCMD% start "WinMerge" /b %WM% /dl %1 /dr %2 %T1% %T2%
start "WinMerge" /b %WM% /dl %1 /dr %2 %T1% %T2%


[Alan Clifford]

exiftool a.jpg b.jpg -a -G1 -w txt
diff a.txt b.txt

- Phil

That looked interesting but I didn't like the thought of typing in the file names twice. So, still rather unrefined,


#!/bin/sh
# Usage: exifdiff.sh path/file1.ext path/file2.ext parameters
# eg:    exifdiff.sh dsc_7811.nef ./jpegs/dsc_7811.jpg -a -G1 -datetimeoriginal

B1=`basename ${1}`
B2=`basename ${2}`

exiftool "${@}"  -w! /tmp/%f.%e.txt \
  && diff /tmp/${B1}.txt /tmp/${B2}.txt  | less -i


I don't know if the /tmp directory is cleared out automagically on my mac. We shall see.

Alan

[Alan Clifford]

I've been looking at diff and side-by-side looks better.  It needs a wide window.

#!/bin/sh
# Usage: exifdiff.sh path/file1.ext path/file2.ext parameters
# eg:    exifdiff.sh dsc_7811.nef ./jpegs/dsc_7811.jpg
# eg:    exifdiff.sh dsc_7811.nef ./jpegs/dsc_7811.jpg -a -G1 -datetimeoriginal

B1=`basename ${1}`
B2=`basename ${2}`

exiftool "${@}"  -w! /tmp/%f.%e.txt \
  && diff -d -y --left-column /tmp/${B1}.txt /tmp/${B2}.txt  | less -i

[simonmcnair]

Hi all,
I naively thought that importing photos from my camera in windows would result in the images being imported in their unaltered state.  It seems now that it is not the case.  I _think_ that windows gallery/photo/import/whateveritscalledthisweek seems to add a load of schema junk to my files.  I have two files which look identical, and have identical exif and IPTC info but one is 3,065,614 bytes and the other is 3,641,151 bytes.  I originally thought that (tin foil hat time) a virus was storing itself in my files in a steganography type style.  After some research I suspect this is not the case, but I cannot fathom what microsoft could store that would increase the filesize by nearly 600K.

Are you aware of a tool that could point out to me what the data is (decode it preferably) so I can just treat it as a dupe and delete the larger file ?

as an aside, is there any mileage in batch converting all my jpg's to pngs or an alternative format which is lossless and 'better' ?

tia :-)

Simon

[Phil Harvey]

If you send me both images I will analyze them to see what the difference is.  I don't know of any tool that is as good as me.

My email is philharvey66 at gmail.com

I don't recommend converting the JPEG format unless you plan to edit the files.  For intermediate files in an editing workflow, I suggest using the native format of the editor (PSD is what I use since I edit with Photoshop).

- Phil

[Phil Harvey]

Simon,

I got the samples, thanks.  Below is a diff of the metadata.  Basically, here is what Windows has done:

1) Deleted the JPEG MPImage trailer containing the PreviewImage

2) Corrupted the MakerNotes offsets

3) Added some XMP

4) Re-generated the ThumbnailImage

5) Changed the EXIF byte order <-- THIS IS A REAL NO-NO!!

6) Added EXIF OffsetSchema and XPKeywords tags

The bottom line is that Microsoft has made a real mess of the EXIF and deleted the PreviewImage completely.  Nasty.  There is no way to completely recover from this.

- Phil

Code Select Expand

diff tmp/2011-01-08_14-49-59_1.txt tmp/2011-01-08_14-49-59.txt
2c2,4
< [System]        File Name                       : 2011-01-08_14-49-59_1.jpg
---
> [ExifTool]      Warning                         : [minor] Possibly incorrect maker notes offsets (fix by 3768?)
> [ExifTool]      Warning                         : [minor] Suspicious MakerNotes offset for DataDump
> [System]        File Name                       : 2011-01-08_14-49-59.jpg
4,5c6,7
< [System]        File Size                       : 4.4 MB
< [System]        File Modification Date/Time     : 2011:01:08 08:49:58-05:00
---
> [System]        File Size                       : 3.8 MB
> [System]        File Modification Date/Time     : 2011:01:08 09:49:59-05:00
9c11
< [File]          Exif Byte Order                 : Little-endian (Intel, II)
---
> [File]          Exif Byte Order                 : Big-endian (Motorola, MM)
24a27,28
> [IFD0]          XP Keywords                     : Misc
> [IFD0]          Padding                         : (Binary data 2060 bytes, use -b option to extract)
56a61,62
> [ExifIFD]       Padding                         : (Binary data 2060 bytes, use -b option to extract)
> [ExifIFD]       Offset Schema                   : 3768
66d71
< [Panasonic]     Data Dump                       : (Binary data 8200 bytes, use -b option to extract)
70c75
< [Panasonic]     Internal Serial Number          : (F52) 2009:11:26 no. 0059
---
> [Panasonic]     Internal Serial Number          :
84c89
< [Panasonic]     Baby Age                        : (not set)
---
> [Panasonic]     Baby Age                        :
96c101
< [Panasonic]     AF Point Position               : 0.7 0.5
---
> [Panasonic]     AF Point Position               : 0 0
115c120
< [Panasonic]     Baby Age                        : (not set)
---
> [Panasonic]     Baby Age                        :
117d121
< [InteropIFD]    Interoperability Index          : R98 - DCF basic file (sRGB)
139,140c143,144
< [IFD1]          Thumbnail Offset                : 10752
< [IFD1]          Thumbnail Length                : 4424
---
> [IFD1]          Thumbnail Offset                : 14670
> [IFD1]          Thumbnail Length                : 3809
155c159
< [MPImage2]      MP Image Start                  : 4014592
---
> [MPImage2]      MP Image Start                  : 4001936
157a162,165
> [XMP-rdf]       About                           : uuid:faf5bdd5-ba3d-11da-ad31-d33d75182f1b
> [XMP-microsoft] Date Acquired                   : 2011:01:10 10:51:36.348
> [XMP-microsoft] Last Keyword XMP                : Misc
> [XMP-dc]        Subject                         : Misc
159c167
< [Composite]     Base Name                       : 2011-01-08_14-49-59_1
---
> [Composite]     Base Name                       : 2011-01-08_14-49-59
169c177
< [Composite]     Thumbnail Image                 : (Binary data 4424 bytes, use -b option to extract)
---
> [Composite]     Thumbnail Image                 : (Binary data 3809 bytes, use -b option to extract)


[simonmcnair]

Phil, You're a Gem.  So Bottom line is to copy from SD card, readonly the sucker and then go from there ?

[simonmcnair]

Phil,
Sorry one last thing.  Please can you give me a copy of the script you used to generate that ?, it'd be really useful for me to understand what's changed in my metadata :-)

[Phil Harvey]

Hi Simon,

Yes.  Windows has a nasty habit of changing writing image metadata when you aren't expecting it.

The "diff" utility I used is a standard Unix utility.  I'm sure there is a free version for Windows somewhere that you could download.

- Phil

[simonmcnair]

Hi Phil,
I meant to ask at the time, but forgot up to now.  Is there any way (and I know this is asking a lot) to get exiftool to identify/rename each Microsoft picture and rename it (for the sake of argument) with the suffix '-Microsoft' so that I can see if I have two copies of the file and delete the microsoft wrecked version ?

tia
Simon

[Phil Harvey]

Hi Simon,

You can do this if you can identify a tag that is added when you edit with Microsoft software.  One possibility is OffsetSchema, which should be added only by Microsoft if your image contained maker notes:

exiftool -if "defined $offsetschema" "-filename=%f-Microsoft.%e" DIR

where DIR is the name of the directory containing the images.  Add -r to also process images in sub-directories.

- Phil

[ruth]

I do this all the time, but I use the "diff" utility (standard on Mac/Linux) to do this:

exiftool a.jpg b.jpg -a -G1 -w txt
diff a.txt b.txt

- Phil

Sorry for replying to such an old thread...

Is it also possible to compare the metadata between 2 (or more) files by exporting all the metadata fields into a csv file? Would there be any problems with fields in the two files possibly not being the same (in number of fields, location, name, etc)?

Thanks, Ruth (total newbie)

[Phil Harvey]

Hi Ruth,

The -cvs option handles the case of different tags from each file.

- Phil