Wouldn‘t it be nice …
…to have an automatic image workflow from the photographer to the publisher?
… to know all the time, when and where and by whom a photo was created?
… to get your revenue share?
… not to rekey again, what others have already done earlier?
… to create high quality colour images for different types of publications?
… without manual intervention

Don’t throw away metadata – conserve it. IFRA has done extensive work regarding image processing issues. They found that conservation of EXIF-data is important for correct and efficient image processing. Specially ICC-related data. But it’s also important to not limit metadata standards to images descending from cameras – it must work for any type of image.

Conclusions and open issues:

• Well defined mapping between metadata properties of different schemas
• Write-Once metadata values
• Versioning of metadata values
• Cameras should allow easy-to-preset metadata values
• Cameras should deliver more than Exif metadata
• Improved support for controlled vocabularies as the source of metadata values on the user interface level
• Consistent implementation and use of photo metadata standards must be improved
• Imaging devices not restricted to cameras anymore
• A key feature of metadata for digital assets are globally unique identifiers

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Michael started with telling the background of IPTC. The IIM standard was introduced 1994. The current ITPC Core was published year 2005. Today those live in parallel. Proper synchronization of data in-between IIM and Core are very important. This must be handled by software.

There are several common misuses of IPTC-fields. One that used to be very common are that the date field is populated with scanning date instead date of the actual capture. Today several fields have different name in different applications, sometimes event with different software application versions. With localization this even get worse due to varying translations. It is important to have consequent naming.

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Reuters have 600 employed photographers and image editors. They produce 1 500 images per day. The photographer are responsible to enter basic metadata as:
• headline
• caption
• category code
• urgency
• supplemental category code (not keywords)
• byline
• credit
• object name
• date created
• city
• state (USA only)
• country
• original transmission reference

The photographer does NOT add keywords!

Add metadata early and don’t throw it away. It would be an advantage if you could upload categories to the camera for quick selection by the photographer. Information like time and place should be handled automatically. Another wanted feature would be if the photographer could enter routing information in the image before shouting so the images automaticallay can be distributed via selected channels.

Reuters currently strips away EXIF data. One major client complained that this information broke his workflow. Reuters adds rights information to the caption and to the special instructions field. But rights management is one important part to simplify. The information that are entered for news purposes doesn’t work well for stock use. There are also several customers asking for RAW-files. Another special request came from MSN – the wanted square thumbnails to make the images appear correct in their template.

Summary:
• Industry Standards allow us to concentrate on what makes us different
• Common standards are great for the client
• What are we discarding today that will be invaluable tomorrow?

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Most stock agencies only use the older IIM standard for IPTC information. They can not properly handle XMP based metadata like IPTC-core.

For historical images you often don’t know exact creation date. But he current date field requires you to specify year, month and day. Then the editor enters a fake month and day just to get past, this is not very good. We need to change the definition of this field to allow less precise dates.

We also lack relevant fields for:
* model release
* type use that are allowed
* With what rights the image was bought
Metadata that are entered with use of controlled vocabularies are easier to translate. It’s also important to enter the information in the correct field. Image buyers expect to find certain information in a specific field independent of image source.

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Thanks to EXIF information it’s now easier to process images in newspaper workflows. It is important to keep EXIF and ITPC metadata through the whole process.

Today images goes through several adjustments and conversions. Each of conversion gives a quality loss. We should adjust the image as early as possible and then do as few corrections and conversions as possible. We at the newspaper can’t fine-tune every image. We must be able to use automatic conversion.
Keep the image in any of the standard color spaces (AdobeRGB or sRGB) and go straight from there to CMYK close to the press. Today most newspaper workflows are based on the use of the IFRA26 ICC profile instead of custom press color profiles.

The photographer should enter as much information as possible shortly after the shoot. Camera manufacturers need to ensure that information is embedded in a non-proprietary format

A problem today is that only 10 to 15 % of all incoming image files contain EXIF information.

“What I’d like to see…
* Further development of software capable of using Exif metadata. Products such as Agfa Intellitune
* Education of photographers (in terms of manipulating images prior to us receiving them)”

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Many Photoshop XMP-fields lacks a corresponding IPTC field. One is the Copyright flag field – Peter suggests that this one should immediately be included into the IPTC standard.

One interesting problem is how to handle metadata for image collages. We would need the possibility to copy all metadata from all source images to the composite. To enable this we need a new XMP bag-container.

For keywords he would like to see parent-child relations (a type of three hierarchy). This could be implemented as an alternative field that shows the relation between keywords. By this implementation we can do this without breaking the current keyword function.
We also need a new type of field for all those keywords that are not true keywords. The keyword field is heavily used for process and handling instructions (archive, retouch, re-scan). Peter suggests a bag-container with note fields for this kind of labeling.

He also suggest alternate metadata sets for different uses, users and alternate renderings. More or less a type of versioning system. This makes it possible for the image editor or distributor to add and modify captions and keyword and still leaving the original metadata from the photographer.

You can read more about Peter Krogh here

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

There are large problems with clients that modify metadata without knowing what they are doing. Many clients believe that they can’t view embedded metadata without opening the image in Photoshop. Another sad thing is that a lot of metadata are lost in many commonly used workflow systems.

David proposed a checklist to solve common problems

• Clients change filename. This makes it difficult for the supplier to find the image again. Always setting a unique image title is one solution would be a great help. A basic thing would be to copy the filename to this field, but this simple operation is difficult to do with ie Adobe Bridge. Many clients download composition images, and change the filename. Then they can’t find the source and image again.
• Save for web in Photoshop discards all metadata.
• Insert metadata as early as possible – preferable at the camera.
• Validate that your metadata are still there if you retrieve one of your images form the image agency. Many distributors routinely discard metadata.
• The standard are missing logical fields to use by image distributors, so they override some of the fields designated for use by the creator (ie copyright field).
• Many distributors, image editors and photographers do not follow standards. They write any type of info in the caption field.
• Custom information are a problem since it doesn’t fit in a regular image database.

85% of all SAA members shoot RAW. But it’s serious problems handling metadata during the raw processing. The most common software for image tagging is Adobe Photoshop / Bridge. The majority of all photographers leaves keywording to the image agency.

Training is the key to better photo metadata!

You can read more about David Riecks here.

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Over a normal weekend the imagedesk at Stern receives 25 000 images. Yesterday (Wednesday) they received 15 000 images from 12 different image suppliers. This day was a bit above normal due to the G8 summit. A search on “G8” results in 8691 images. If you limit to the last three days you get 4355 hits. If we narrow the search with the keyword “demonstration” you get 796 images. Only 0.03 percent of all incoming images are published.

Andreas spoke about the problem with incorrect information entered by photographers and image editors to give more search hits to their images. They believe that this increases their sales chance. But this is very frustrating for the image buyers trying to find an image of an specific topic. Very few image editors have proper education, they are self educated or has learnt their work through collegues.

Over half of all images in a normal picture archive show people. A search for Angela Märkel gives over 11 000 images at Getty. Since we ask photographers to enter journalistic text we also get hits for images of demonstrations against her politics.

LESS IS MORE
Have you ever viewed page 81 of a google search of “photo archive”? A search that term results in 258.000.000 entries. To check all those entries, then this search will take you 44,791 days (spending 15 seconds per entry), or just under 122 years – but that doesn’t include any coffee breaks!

It is more important to get better quality of image metadata content than better database software.

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

No tags Hide

Notes from the first Photo Metadata Conference. This conference was held in Florence, Italy june 7, 2007. It was organized by Cepic, IPTC and IFRA. These are my private notes published to give you some idea of what was discussed. I will later write some articles from the conference and publish those on BLF.se and DigitalFotografen.se, but probably only in Swedish. If there are any demand I might make a short translated version in English. You can find presentations and more information at phmdc.org and cepic.org .

Yesterday I attended the Photo Metadata Conference in Florence. A full day on the topic of how to mark and find images in digital image archives. This conference was a success. The conference room were crowded and those who arrived late (still meaning before start) had to sit on the floor. A quick count in the morning resulted in more than 120 participants. The quality of speakers and discussions was very impressive. Many great suggestions on how to make image mark-up and searching came up during this day. The camera manufactures Canon, Nikon and Hasselblad attended and were very open to suggestions. The achievement of getting people from Nikon and Canon to travel all the way from Japan specifically for this event is worth a warm applaud to the organizers.

To summarize the day in a few bullets:

• To days enormous flood of images requires accurate mark-up
• Keyword spamming are a sever problem for those trying to find a specific image. Many photographers and image editors writes as many keywords they can possible come up with, in many cases even totally wrong ones just to give the image more search hits.
• Many images are missing unique image identification and owner information
• It must be easier and more efficient to tag and mark images
• It is important that no metadata are lost in workflows and distribution systems. Interoperability is a serious issue.
• The demands are very different for press, stock and heritage images
• The is a huge demand for proper training of photographers and image editors

This article will soon be followed by summaries of each speech as separate posts.

I am a member of Cepics technical committee that recently released a white paper on how to improve the current IPTC-standard. Well to be honest, many of the other in this group did the major work. I have more been a reviewer. But I will soon start a more active part with this work – one of my roles in this group will be to formulate a suggestion to the camera manufactures on what IPTC-fields that are the most important to set already in the camera. But more on that later…

No tags Hide

Tisdagen den 5/6 hade EU-projektet Miles konferens i Florens. Jag blev inbjuden av Staffan Teste att delta i detta seminarium. Här är korta sammanfattningar från några av dagens tal. Tyckte en del var väldigt intressant så därför delger jag er läsare detta.

Paul King (BAL), GLS Migration Project
Berättar om hur de nu bygger ett nytt system för att hantera sina samlingar. Man har haft sitt gamla system i nästan 20 år. När man nu utvecklar det nya systemet har man utvärderat befintliga system samt standarder. Men de hittade inget som täckte hela deras behov. Därför bygger man et eget system som baseras på en egen metadatastruktur. De tittar dock noga på befintliga standarder för att de ska vara enkelt att importera och exportera data enligt befintliga standarder. Det är lätt att mappa de egna fälten mot motsvarande fält enligt olika standards. Det som skapar problem är varierande vokabulär samt namn på konstnärer, platser och tekniker. Om platsen är ena gången är namngiven med ett engelsk namn och andra gången med samma orts franska namn går de inte hitta med samma sökning. Detta samma gäller färger. För att kunna hantera detta har man byggt upp synonymordlistor där man definierat vilket som är primärt ord (red, Firenze) och sekundära (scarlett, Florence). Med hjälp av detta går det bra att behålla sökbarhet även i importerat material – om bara materialet har en bra struktur går importen bra. Mardrömmen är när all information är inskriven i ett och samma fält.

Nick Pole (MDA) SPECTRUM
MDA – UK standardisation body for Collections management
Museer förändras från att vara utställningshallar och lager för samlingar till att mer handla om att berätta historier och leverera innehåll. Detta innehåll använda sedan utbildningsmaterial, original till souvenirer och bilder till externa utställare. Men som kund är det svårt att hitta materialet när det tillhandahålls av 1000-tals olika museer. Det är viktigt att tjänster som samlar ihop detta till centrala sökbara kataloger. För att detta ska vara möjligt krävs att metadata är standardiserat så denna aggregation kan ske helt automatiskt.

SPECTRUM Standard ProCedures for CollecTions Recording Used in Museums
Licenserat till över 4000 museer över hela världen. Det är ett metadataramverk för varje del i att ”managing” ett objekt i en samling. Det innehåller över 400 element.

Nya riktningar inom metadata
Det är allt för dyrt att sitta och försöka fylla i allt som kan besvara framtida frågor om ett objekt.
Man bör titta på att skapa metadata ur det information som (social tagging). Genom att dra nytta av frivilliga intresserade som skriver om föremålen kan vi mycket mer information om objekten. Kvalitén på detta behöver inte vara sämre än nuvarande professionell märkning. Mycket av dagens information kommer från ganska okunniga människor som indexerar objekten – det är ofta ett nybörjarjobb eller i vissa falla voluntärprojekt. Problemet är att läsarna tror att det är fakta skrivit av en expert.
Det är viktigt att vi har ser på metadata från med mera affärsmässiga ögon. Det gäller att kostanden för märkning ligger i relation.

www.mda.org.uk
www.colectionslink.org.uk
www.culturalproperyadvice.gov.org

George Kordelis (Archetypon) – XLIFF
George berättade om XLIFF som är ett filformat för översättning och lokal anpassning av framför allt mjukvaror.
XLIFF – XML Localization Interchange File Format (OASIS-XLIFF)
En XLIFF fil kan innehålla allt källmaterial samt ordlistor samt färdigöversatt materiel för ett helt projekt. XLIFF hanterar inte bara översättning av text utan även grafik och textformatering.

Han tipsade om ett antal andra besläktade initiativ.
The Getty Research Institute  – Crosswalk of Metadata Element Sets
RDA Resource Description and Access – en ny standard för resursbeskrivning och access. Den är skapat för digitala användning redan från början. De flesta andra standarderna för beskrivning av föremål har ofta ett ursprung i manuella katalogsystem.

Xavier Castell (AISA) – Multilingual Thesauri
Xavier berättade om problemet med flerspråkiga system.

Han gav tips på bra ordlistor för digital översättning.
Getty Thesaurus – mest engelska och ej speciellt djup
Wordnet – bra men dyr och innehåller en hel del synonymer som ställer till problem.
Library of Congress

Andrea de Polo (Alinari)
De har en grupp som väljer ut vilka bilder som ska digitaliseras baserat på om de är säljbara. De digitaliserar och indexerar ca 300 bilder per dag. Om man skulle vara effektiv borde man starta med en specifik typ av bildmateriel och göre alla dessa bilder klara innan man byter till nästa. Men detta funkar inte praktiskt då externa händelser påverkar vilken typ av bilder som är intressanta för tillfället. En viktig bit är hur mycket bildförbättring som ska göras. Normalt tar man bort färgstick och repor men inte fingeravtryck som kan ja ett historiskt värde.
Texten skrivs på italienska och översätts sedan till fler språk. När det gäller nyckelord så är det intressant att titta på statistik från sökmotorer. Detta för att se vilka nyckelord som folk använder och vilken typ av material som är mest eftersökt.
Man arbetar en hel del med specialiserade miniwebbplatser riktade mot olika intressegrupper. Dessa små webbplatser kan skapas snabbt då de automatiskt plockar material ur det centrala arkivet.
Det är viktigt att använda skanners av bra kvalitet och sedan levererar bilder i en standardfärgrymd som Adobe RGB. Det måste vara lätt för kunde att använda bilden utan risk för färgfel. Man är noga med att tillhandahålla all information som slutanvändaren behöver för att kunna använda bildfilerna korrekt.
Man gör fortfarande kopiering enligt gammalt förfarande. De blandar fortfarande kemi enligt recept från de ursprungliga Alinaribröderna. Man är försökt med nya material och processer.

Alinari använder Microsoft Accessdatabas. Man är intresserad av open source lösningar men problemet är att rekrytera personal som är bra på dessa lösningar. Det är lättare att rekrytera personal som kan Microsfts lösningar.
Det är problem med att Apple och Microsoft försöker att driva metadata och sökfunktioner så de låser in användarna i sitt system. Detsamma gäller många applikationer. Många applikationer lägger till en massa privat information som egentligen inte är relevant för bilden.
Det är viktigt med ett öppet ramverk för metadata som kan är lätt att utöka och fungerar på olika typer av mjukvara och hårdvara (inklusive kameror).

No tags Hide