Paul Hammond's Blog

Last week, through a friends blog, I stumbled across a Flickr group called 365 Days.  Since writing a novel for National Novel Writing Month last November, I have been looking for something new and creative to challenge me.  I don't feel that I have been spending enough time with my camera of late, so this particular challenge appealed to me.

Very simply, to quote the 365 Days rules page: "the 365 Days group is a project in which members submit one self portrait each day for a year".  There is further clarification on the site on what counts etc., but that is pretty much the essence of the project.  Again, I quote, "the main thing is that you are both the photographer and the subject. For the purposes of this group, any photo which you took that contains any part of your body counts".

If you'd like to follow my progress, here is my 365 Days Flickr photo set.  Alternatively you can follow in your feed reader, via the RSS feed here.

For day 1, I couldn't decide on a specific photo, so I ended up creating a collage.  It looks a bit like this:

As the year goes on, it'd be great to hear your feedback.  You can comment on the photos over at the Flickr photo set.

When I first started putting photographs online, I spent ages looking for an album creation tool that would give me some control but would automate a lot of the creation.  I eventually settled on JAlbum - a nice tool written in Java that has many downloadable skins to choose from.  You can see an album that uses JAlbum here.

More recently however I discovered two Flash based image viewers both from Airtight Interactive - AutoViewer and SimpleViewer.

AutoViewer allows you to create a line of images spreading from left to right.  It will show the images as a slide show, or allow the user to click through them one by one.  The images are shown at full size, unless the browser window is too small at which point the Flash viewer does a great job at scaling them down.  Example AutoViewer album.

SimpleViewer creates a thumbnail matrix for your images, and allows you to view full size versions by clicking the thumbnails.  Again, the images scales if necessary.  Example SimpleViewer album.

One nice feature is that the images are intelligently cached when the page loads.  This minimizes wait time for the end user and it is a great user experience too if you click on images that are not yet downloaded and cached.

Both are controlled by a single XML file that includes all of the options for the album as well as the image list.

I spent a little time writing a very crude application to allow me to create AutoViewer XML files very easily.  For SimpleViewer, I actually used to use a JAlbum skin, but this only supports version 1.7.1.  The current version is 1.8 and the XML format has changed a little.  I just copy and paste an old album file and fix it up by hand right now.  I hope to extend my "crude application" to cover SimpleViewer very soon!

A while ago, I signed up to take part in the beta of riya.com; I had forgotten about it until last week when I got an invite mail to try out the service.

In a nutshell, riya.com is a facial recognition service for uploaded photos.  It allows you to submit your photos to a website, and it will then start picking out portions of those photos that “might be faces”.  It then asks you to spend a little time identifying the people in the pictures - training it if you will.  After you have told it a few times which people are which, it can then start “auto-recognising” people from any new photos you upload.  It seems to be able to pick out words too (so, maybe road signs or shop names etc) which might allow you to better identify your pictures.

Much like other photo sites, it offers the ability to tag pictures.  It auto creates a number of tags from your Exif data, and allows creation of your own tags.  You can search your photos by person name, location, date range, album (it auto-creates albums based on your uploaded file structure), or tags.  You can also mark pictures and albums as “private”, or for sharing.

One nice feature is that you can link up with other people in your contacts list, and use their recognition data to help feed your recognitions.  When you train the software, you tell it both a name and an email address, so I presume it uses this data to cross reference my recognition data with others.  This could be a pretty powerful feature if enough of your friends and family begin to use the service.

Some comments I have on the service so far:

• Firstly, it requires you to install an uploader (written as a Java application) to upload your pictures.  This can run in your tasktray, and you can set it to monitor a folder so it uploads and auto-recognises your pictures as you add them.  It will also auto-start on when you start Windows.  I was a little unsure about installing yet another tasktray application, but it doesn’t seem too bad on resources, and it does make uploading to Riya.com easy.
• The website itself is very well put together.  It uses a lot of “Windows-like” funtionality (such as drag-and-drop and mutliple selection using the CTRL key) and talks to the Riya.com services via some sort of AJAX funtionality.  The whole user experience seemed to work well for me once I was actually inside of a specific task.
• That said, there are a number of ways to get into the differing “training modes” and it took me a little while to figure out what was going on.  Not being one to read manuals, there is probably an FAQ that tells me all I need to know, but you know how it is…
• It doesn’t work in IE7 (yet), although with the level of interaction I think they can be forgiven for now.  Works great in FireFox.

I have uploaded maybe 500 photos, and it seems to have gotten pretty good at recognising those people who have 10 or more “sample pictures” that you trained it with.  All in all, it seems to do what it says on the tin really quite well.

However, now I have checked it out for the geek factor, my interest has waned a little.  I have about 20,000 photos (so far), and I use Microsoft’s Digital Image Suite to manage them.  This gives me similar features (such as tagging etc) so that I can organize everything, and it is all done on my own machine using full resolution photos.  If I do want them online, I currently share a subsection of them with friends and family via my own website.  Uploading that many pictures *just* so I can search them seems a bit much.

Now, if you could build this sort of functionality into my desktop application so that I can have an extra level of classification, then I would be very happy.  Right now, I tend to flag pictures by hand with the people who are in them, think how much time would be saved auto-tagging my pictures with names.  Of course, this doesn’t necessarily take advantage of the collaboration aspect of using your contacts’ recognition data, but to be honest I am not sure how much cross-over my photos have with theirs anyway. 

Anyway, photographers and AJAX developers out there, check it out – http://www.riya.com.  For a good overview, the tour is a great place to start – http://www.riya.com/learnMore

Ever taken what you thought was that perfect “family photo” only to develop it (or download it to your computer) to find Granny is grimacing, Uncle Joe looks like he has indigestion, baby Alice is looking the other way, your sister looks asleep or your mum looks wild eyed on some hallucinogenic drug?

For me, the idea behind the family group shot is almost always spoiled by the physical effect produced.  You want to capture that happy moment for all to treasure, to try and relay to the viewer the joy of the occasion.  Instead you get one (or maybe more!) scenes of sleeping, farting, spaced out grimacing.

Maybe you take 5 or 8 shots; surely Granny is getting close to a smile in ONE of them.  Baby Alice decides to grace you with her smile on a couple of pictures, but not the same ones as when Granny found her grin.  Mum and sister both sort their eyes out for the last picture, but fail on the rest.

That leaves you with EIGHT shots that all fail dismally to convey the core atmosphere you were trying to capture in the first place – a glorious record of the happy occasion.

Enter Microsoft Research’s “Group Shot”.

This software lets you take a number of “similar” photo’s, and later on select the best parts of each to merge into that single perfect shot.  You can highlight Granny’s smile, baby’s face, Mum and sister’s eyes, all in different versions of the shot.  Then the software will pull together a great faimly portrait you can be proud of.

Obviously I am only listing the one obvious use for this software, I am sure there are many more scenarios where this would be pretty useful.

I tried out an earlier internal copy of this and was quite impressed.  In a later post I will try and pull out some of my own photo’s and give an example of the results this software will produce.

My colleague Matt sent round a link the other day to a Stanford paper on Light Field Photography.  These geniuses have built a plenoptic hand held camera that takes a 4D light field picture in a single exposure.  It is all very technical, but it essentially means you can re-focus a single picture to different depths in post-production.

Rather than duplicate the images from the site here, why not click through and check it out yourself.  There is a sequence of 5 images of a crowd on the homepage here.  On the gallery page, there are a number of videos of the images being changed right before your eyes.

For the technically inclined, there is also a copy of the paper there too: Light Field Photography with a Hand-held Plenoptic Camera.

I am happy to be returning to Ireland for a week in August to visit my friends Simon and Esther.  They live just outside Clonakilty which is in County Cork.  Ireland is a very beautiful and peaceful place, and I am looking forward to a week of relaxing with friends, running along the beach in the morning, and taking my camera out to try and capture some of the beauty.

One of Simon's friends owns a cafe in Clonakilty called Betty Brosnan's (which does killer breakfasts btw).  In addition to the food, he also uses the space to display his own black and white photography taken in and around the West Cork area.  It is great to be able to relax with cup of coffee and admire the great pictures.  You can even buy the pictures if any take your fancy.

I am happy to say that it is no longer necessary to travel all the way to Southern Ireland to enjoy the pictures.  Simon recently helped Dermot put a website online to showcase and sell the pictures.  The site can be found at http://www.westcorkinblackandwhite.com.  Take a look, see what you think.

A couple of weeks ago I posted about Microsoft's announcement of RAW support in Windows.  They have now released the RAW Software for Windows XP - you can grab a copy here.

There is also a Whitepaper: Viewing and Organizing RAW Images in Windows XP.

Nice.

Microsoft is going to be adding RAW support to Windows in a few different ways.  This is very cool for those of us that have Digital SLRs that support the RAW format.

• They are adding RAW support into Longhorn, which is a great move IMHO.  However, this is not greatly exciting "now". 
• Better news is that they are soon going to release a free download for Windows XP that will allow viewing and previewing of RAW files within Windows Explorer (specifically Canon and Nikon format RAW files).  This will be invaluable.
• Additionally, the new version of Digital Image Suite will support RAW in both the library and editor applications.

Check out this article for more details.