Discussions
Categories
Groups
Community Home
Categories
INTERNAL ENABLEMENT
POPULAR
THRUST SERVICES & TOOLS
CLOUD EDITIONS
Quick Links
MY LINKS
HELPFUL TIPS
Back to website
Home
Web CMS (TeamSite)
Image quality reduced when stripping metadata
nate25
I set up a retreival task to strip all metadata (excluding one or two) for rendering on the web via the "Metadata Remover" transormer, but apparently when doing so it's also reducing the quality of the image; you can easily see the degeredation. Is this intentional? Am I doing something wrong?
On a related note, what is the difference between doing a "(get original)" retrieval task vs a "(none)" retreival task? I see that the image size is reduced when doing the latter. Is there some compression going on?
Any assistence appreciated.
Find more posts tagged with
Comments
msrinivas
Please attach the task that you have created.
From page 113 of the MB 4.6.1 admin guide:
Use the (None) and (Get Original) Tasks When Retrieving Assets
When an image asset is inserted into the repository, MediaBin automatically adds a tracking tag to the file which allows the image to be traced back to MediaBin.
If you use the (None) task to retrieve an image asset, you will get a file with exactly the same image data as the original, however it will also include any metadata edits and a tracking tag. In addition, the image format may have been “normalized” such that the file is usable in applications that were unable to read the original file.
If you use the (Get Original) task to retrieve an image asset, you will retrieve the file that was originally inserted. This file will not reflect any metadata edits and will not contain a MediaBin tracking tag.
MediaBin only makes minor modifications to standard image assets; no changes are made to other asset types including Generic, Multimedia, Microsoft Office, PDF, vector EPS, Quark, and Flash files. Therefore, for all asset types other than image assets, there is essentially no difference between the None and Get Original tasks. Using either task will result in the retrieval of the exact same file that was originally inserted into the repository.
HTH
nate25
Thanks msrinivas. It's interesting that with the (None) taks it reduces the file size quite a bit, I'm seeing between 15 and 45% reduction in size.
In any case, I attached the metadata remval task - note it includes custom meta data I added to MB. To creeate the task I did as mentioned - used the Metadata Remover tranformation, and selected all meta data excluding 1 meta data I added ("Credit").
I also attached the phyical images generated via the (None), (Original) and the strip_all_metatags tasks.
Migrateduser
Are you comparing the files in Photoshop or in the browser. I didn't see any change when comparing the files in IE. In Photoshop you can see the lighting is different, but that's because it's using the embedded ICC color profile in the original file.
lyman
If you want the images to look the same in Photoshop, strip all metadata except the ICC profile! While that is technically metadata in our architecture (and in many instances such as web deployment an important thing to strip), you would not want to remove it if the intended destination is a program that performs color correction.
No other metadata removal that I can think of would affect image quality. Neither should there be a distinction between (None) and (Get Original) in image quality. The distinction between the two tasks is first that (Get Original) is usually much faster but does not preserve metadata edits.
Cheers,
Lyman Hurd
nate25
Thanks all for your input...making more sense now.
nate25
Since you've been so helpful...I'm still surprised that doing a "(none)" task shrinks the file size significantly - from what you're saying, if anything it should be slightly larger then the original (w/ the additional meta data) - is there some compression going on? Also with the meta data remover taks, I'm not a big image guru, but I thought jpegs are already efficiently compressed, and seems odd that removing metdata reduced file size so much.
lyman
To simplify somewhat, JPEG's consist of two parts, metadata (caption, ICC profile, etc.) and compressed image data. Only the latter is compressed and both (Get Original) and (None) keep this part of the file intact.
The (None) should only really be smaller if metadata is removed. Metadata is reconciled between MB and the outgoing image in a (None) retrieval which means that metadata I have removed in MB no longer appears in the image (i.e., reconcile can add, subtract, or edit).
The elephant in the room for JPEG's is often the ICC profile which determines the color mapping. It is basically a big equation for mapping colors. There are some simple ICC profiles out there but they can get quite large and software such as web browsers ignore it. That is why removing it often goes a long way towards reducing filesize. A typical use case is that you preserve in MB the "big" version for editing and deploy to the website the stripped down one.
Hope this helps.
Cheers,
Lyman
nate25
got it, thanks!