Toujours Frais

Wherein Lefty Does A Good Deed

I regularly peruse the photography subreddit on Reddit.Com and earlier tonight, while taking a break from day job-related stuff that overflowed into my non day-job timeframe, I came across a post from a fellow Redditor asking for help reconstructing a RAW file.  Specifically, he asked posted the following:


Being in a generous mood, I decided to help him.  Of course, as he stated, Photoshop was going to be difficult, so I wasn’t about to give that a shot.  (I’m not that altrustic!)

So I gave it a shot in Lightroom 4.  No dice.  Then I remember that I’d played around a bit with an app called Raw Photo Processor.  So I loaded it up, imported the file and…

 

…Success!

I converted the file to a jpg, uploaded to imgur and posted the link.

Sadly, my fellow Redditor didn’t recognize the subject of the photo, so who knows what or where it came from…it’s just some random kid:

And while it didn’t really help this guy, it was fun doing some detective work, if only for a few minutes.

 

Update! It might be the father’s ex-girlfriend’s kid!

Retraction

One of the problems with running a daily photoblog with almost 1700 entries is sometimes it’s hard to remember what you’ve published already.  Last night, I posted today’s (original) image, went to sleep, got up, took my dog to the dog park then realized that today’s image seemed really familiar.  Returning home, I searched my archives and discovered I had posted a remarkably similar photo in July of last year.  I quickly updated the post with a new image, but am slightly disappointed in myself.  Anyhow, see for yourself:

 

Today’s original image:

 

 

 

 

 

 

 

 

 

 

 

Balconies and Glass, published July 8, 2011:

What RAW Will Get You

A lot of people are confused when it comes to RAW vs. JPG, so I just wanted to show you a quick before/after of what kind of dynamic range you can get out of RAW. This is today’s shot on my photoblog. The before is what it looked like out of the camera, while the after is with the exposure boosted. And while I’d never publish the “after” without some serious post-processing to clean up the banding and noise, you can clearly see that an amazing amount of detail and color information is hidden away in the dark areas of the photo. I, of course, settled on a more sensible final exposure that is more interesting than the “after” shot.

before
after

My 500px Randomizer

I’ve mentioned this before, but I thought I’d revisit it since I’ve recently rolled out a few changes.

In the last year or so, 500px has exploded to become the “serious” alternative to Flickr—a place where you show off your best work while not being crowded out by the kinds of people on Flickr that upload every lame snapshot of their kids or dump their entire CF or SD card into their stream everytime they come back from an outing.

Unfortunately, one of the big problems with 500px is that there’s no “good” way to randomly view images, save for a link at the bottom of their homepage inviting you to “StumbleThru 500px”, which simply uses user-submitted links on third-party service StumbleUpon.

I found this to be inadequate, as it only relied on photos that users had bothered to “Stumble”.  This meant that the number of images were somewhat limited as well as confined to mostly the more popular images—ones that the hive mind of the 500px viewership community had decided were “the best” and worthy of linking to.  After a few clicks on the Stumble button, I found that I was getting the same few images over and over again—some of them were even months old.

What I wanted to do was see truly random images from 500px, so I set out to make a tool to do the randomizing for me.

My first iteration was simple.  Since, at the time, 500px didn’t have a publicly-exposed API, I had to be “kludgey”.  A bit of poking around revealed that 500px simply numbers photos sequentially as they are uploaded—in fact, you can easily go see the first photo uploaded publicly thanks to the ridiculously-easy-to-reverse-engineer URL convention of http://500px.com/photo/<number of photo>.  This made it extremely easy to create a php application that would simply generate a random number between 1 and whatever the largest number used is.  The php code is easy:

$iPic =  mt_rand(  1 , $iMax );
return “http://500px.com/photo/”.$iPic;

Where $iMax is whatever the largest number used is. Since I didn’t want to have to bother writing an overly-long script to figure out what the latest number used was, I’d simply manually update $iMax every couple of days with the photo ID of the first image on 500px’s Fresh page.

In my app, I embedded an iFrame to hold the photo’s page while the app acted as a wrapper around it, keeping the “Randomize” button and “Permalink” link visible at all times.  And this worked great, but I wasn’t happy.  I quickly found that while 500px has tons of great work on it, there’s a lot of cruft.  Especially amongst the older photos submitted before the community really “took off”.  What I really wanted was to be able to randomly browse photos that were of a bit higher quality and were on the newer side (and this, I suppose, is where my application lost it’s complete randomness).  Unfortunately, this wasn’t really possible at the time, so I let the project sit on the backburner for a while.

Then 500px opened up its API.  And I leveraged it.  I quickly figured out how to use JSON queries to return arrays of recent photos in the Fresh, Editors’ Choice, Upcoming and Favorites categories.  Not really caring about Fresh since that’s the initial “dumping ground” for everything uploaded, I concentrated on the other categories.  I quickly produced some code that queried the API and returned a random photo from the user’s choice of categories, replacing my previous code that haphazardly propelled you around 500px and instead allowed you to randomly view more-curated photos.

A bit of Javascript trickery also allowed me to make a feature to view the photographer’s 500px homepage using a window blind effect so that a user can quickly check out other photos by that photographer as well as follow them without leaving my app.

So, there you have it, the story behind my 500px Randomizer.  Feel free to use it all you want;  just keep in mind that it’s still a work in progress and if you find any bugs, feel free to send me an email at lefty@leftyrodriguez.com to let me know.

Processing “The Last Supper”

Pedro Alves asked in a comment on today’s photo if I could explain the processing.  So I thought I’d give it a quick try.
The original raw photo was shot at f/10 at ISO 100 and a shutter speed of 1/100 of a second, using a polarizing filter to darken up the sky a bit.  After importing to Lightroom, I pre-sharpened and adjusted the white balance, giving me this:
Not very exciting, eh?  I decided to tone map it to bring out the shadow and highlights detail in a sort-of “faux” HDR process.  Since I hadn’t shot multiple bracketed exposures, which would be necessary to do true HDR, I faked it, relying on the pure dynamic range that shooting RAW affords a photographer.
In Lightroom, I created four virtual copies of the photo, giving me five copies altogether, including the original.  I left the original’s exposure value at 0, then set the others at values of +1, +2, -1 and -2 respectively, imitating the bracketed exposures I’d get with a “real” HDR shot.  I then exported these to Photomatix to do the tone mapping, which resulted in this image:
This gave me great detail in the shadows, but killed the sky.  I didn’t really care, though, because I still had work to do.  I imported the original photo with the dark sky I liked and the tone-mapped photo I’d created in Photomatix into Photoshop for further work.
First step was to copy the tone-mapped version into a new layer over the original.  I then created a layer mask which allowed me to use a black paintbrush to “punch through” the tone-mapped layer to the original photo below.  I used a brush with an opacity set to roughly 50% to slowly bring the original sky into the tone-mapped layer.  Once I was satisfied, I applied the layer mask, resulting in a photo that had tone-mapped statues and mountains, but original dark sky.  I then used Topaz Adjust to bring out a bit of detail in the mountains and statues, because I feel like the tone-mapping process leaves the photos looking a bit flat detail-wise.
My next step was to convert to black and white.  For this, I used Nik Software’s excellent Silver Efex Pro 2.  I started with the built-in “high structure” preset, then added a bit of extra structure and a little bit more contrast, while dropping the exposure down a notch or two.  Then, I used Silver Efex’s control points feature to darken up the sky just a bit more while leaving the mountains and statues unaffected. Once this was done, I saved back to Lightroom, did some final noise reduction and a bit of sharpening and posted it to the site.
Here’s a before/after:

before
after