WoW Embedded Data in Screenshots - Need Info

[Nikachelle]

This is kind of WoW involved, and kind of not but I'm hoping there are people that know how to do this because it's something that has direct impact on a project I'm working on at work. (Not the WoW part of it, but rather the way the technology was adapted.)

About a month back we discovered that all our WoW screenshots were being encoded/watermarked with our account IDs, IP address, etc. ( http://www.tomshardware.com/news/waterm ... 17522.html ) What I'm curious to know, is how this is done and if this is something that once an image has been watermarked, if it can be read by some type of software without going through elaborate Photoshop extraction methods (which this thread seems to indicate: http://www.ownedcore.com/forums/world-o ... shots.html ).

Now where am I going with this? Essentially, we're trying to develop a platform that can tag meta data into documents without corrupting them so that they can be reuploaded at a later date without the user having to re-align their documents to match the previously existing document that they're replacing. So far we've had developers indicate that this is possible for all Microsoft products, but that JPGs and image files run the risk of becoming corrupted due to the watermaking process. We were told this was the case because image files do not have a lot of extra data space in order to inject tagging data.

The situation with WoW barcoding their image files is what made me think that this is actually possible without corrupting files.

Does anyone have any insight? Thank you!

[Koatanga]

Not sure if you are looking for a digital or optical solution for this, but here is a company I found using a very quick search for invisible metadata on pictures:

http://www.digimarc.com/digimarc-for-images

[Sabindeus]

So actual in-image watermarking is Not What You Want. What you want is a metadata field.

http://en.wikipedia.org/wiki/Exchangeab ... ile_format

When Exif is employed for JPEG files, the Exif data are stored in one of JPEG's defined utility Application Segments, the APP1 (segment marker 0xFFE1), which in effect holds an entire TIFF file within. When Exif is employed in TIFF files (also when used as "an embedded TIFF file" mentioned earlier), the TIFF Private Tag 0x8769 defines a sub-Image File Directory (IFD) that holds the Exif specified TIFF Tags. In addition, Exif also defines a Global Positioning System sub-IFD using the TIFF Private Tag 0x8825, holding location information, and an "Interoperability IFD" specified within the Exif sub-IFD, using the Exif tag 0xA005.

You should be able to store arbitrary some data in one of those blocks in a JPEG. However...

Exif metadata are restricted in size to 64 kB in JPEG images because according to the specification this information must be contained within a single JPEG APP1 segment. Although the FlashPix extensions allow information to span multiple JPEG APP2 segments, these extensions are not commonly used. This has prompted some camera manufacturers to develop non-standard techniques for storing the large preview images used by some digital cameras for LCD review. These non-standard extensions are commonly lost if a user re-saves the image using image editor software, possibly rendering the image incompatible with the original camera that created it.

Only 64 kb of it.

[Nikachelle]

Awesome, this is helping!

Sab - Yep, meta data is what I want, but I wasn't sure if the watermark/barcode from the WoW screenshots operated the same way as the meta data tags.

I know what 64kb means to me in terms of MP3s, txt files, jpgs, word documents, etc. But I don't know how big 64kb is in terms of embedded meta data. Can you give me some sort of reference point for size? Maybe a character length example? (i.e., holds 100 characters).

Our developers are saying that IF this is possible it would require two extra days of work PER file type in order to develop this (outside of Microsoft products). Which means that all our tiff, gif, bmp, jpg and png files will add an additional 10 days to our development timeline. Is this a realistic ask from our developers?

Koatanga - Thank you for looking into it. However, I'm not specifically looking for something in order to establish copyright.

To explain a bit more/clarify - I'm looking for a method to tag documents, specifically image files since they've indicated this is the problem point for them, to enable for easy matching when re-uploading (so that the user does not have to manually match their new documents to the old documents they've since modified). This cannot be done via the file name, as it is entirely likely that they would rename the file.

[Sabindeus]

1 character = 1 byte in ascii, 2 bytes in unicode/UTF-8


However you originally asked about JPGs, but those other formats have different deals -

GIFs don't have metadata regions afaik

PNGs can have an arbitrary number of arbitrary length text regions

TIFF is the same as JPG for this purpose

As a developer myself, I would advise you to listen to your developers on how long it will take. Even if they're padding, they're padding for a good reason. I can assure you it IS possible, and they are definitely right about each image format you choose to support taking additional time, and if they say it'll take 2 days then I'm sure that is a good estimate. There's plenty of open source libraries they can use to modify a bunch of files, so their efforts should be spent learning and integrating them. If they need a push in the proper direction, it would depend on what language(s) you're working in. But I'd start here: http://libexif.sourceforge.net/

[Sabindeus]

i feel like given a few hours and google, I could write a tool to arbitrarily manipulate metadata for whatever images in ruby...

[Nikachelle]

i feel like given a few hours and google, I could write a tool to arbitrarily manipulate metadata for whatever images in ruby...

lol can we hire *you*?

(I totally don't have that kind of control over this project.)

[Andurin]

But you can print out this topic and show them to the ones who are in charge? :p And without going deep into the links posted, I think Sab got a really good way there, and I think that is also a bit more safe then tampering with images the "WOW" way

[Boyfriend]

I'd concur with sab that very likely meta data as additional fields; Watermarking the wow way, even with a more advanced method will likely give you far less space than 64kb. Advantages of watermarking is that they can survive format conversion by systems unaware of the existence of the watermark (might even still be present if a very high quality camera takes a picture of a high quality display with the watermark on it) and the capability to be relatively well hidden. If these are not things you require EXIF metadata is much better and gives you much more space. (Wow embeds around 256 bytes into screencaps)

64kb is a lot, 65'536 characters (~30 pages) uncompressed and you can compress or 'zip' any data first before putting it in the metadata field so I'm pretty sure you could get up to a hundred page document of pure unformatted text into the EXIF, or alternatively a relatively small image or icon would work too.