I went out last week to play with umbrellas at sunset and make a couple portraits of Anna. I don’t get much time to experiment or shoot against sunsets and grass fields, so every once in a while it’s good to practice, I think. I wish the grass field photo had been a little earlier to get the sky a little brighter. After shooting frames with the 24-70 2.8, 85 1.8, and 70-200 2.8, it’s kind of funny that these two are both on the 85 1.8. It is a monster of a little, relatively cheap, prime.

Either way, it’ll be happy and sad when I’m full time at school, full time at the Emerald, and trying to do some of that “sleeping” thing. To be honest, I really miss the daily epicness of the paper. It’s seldom dull, often terrifying, and it all gets done somehow, every day. Good stuff.

I’m taking Advanced Photojounalism in the Fall, as if everything else wasn’t enough. I think Sung Park is going to teach it. When he was in town a few months ago some of the Emerald staff went out and drank some beers with him. He’s really awesome. So whether it’s Sung or Lee Emmeret (he’s taught it in the past), I’m sure it’ll be awesome.

Well, it’s probably time to sleep. I’m getting up picked up at 6:30 a.m. to photograph the Keeley Butte Classic tomorrow morning. Then I should probably update my 365 Project, but instead I’ll make some photographs for Monday’s Emerald and assemble housing leads.

I brought my H4n to grab some ambient audio, thinking it’d be fun to throw it in behind the slideshow. The shutter firing was way louder than everything else, though, so I scrapped it. Next time I’ll focus on getting some clean ambient…

There are some good examples of using nginx to front-end Node (DailyJS, HowToNode.org).

I use name-based virtual hosting with Apache2 on Ubuntu 10.04 on one IP. One virtual host file for each domain, Apache pulls them all in… pretty straightforward.

I’m also running ‘ufw’ to restrict traffic to ssh and http over port 80, so port 8100 is not accessable from the Internet.

And it works, so yay.

I need to do some more editing. Soon, I hope.

The following 42 line snippet is the whole application! I thought I’d post it because it demonstrates quite a lot of capability that these kind of applications can crank out in not-too-much code.

Lines 1-5 load some libraries, particularly Connect (framework), Mongoose (MongoDB) and Underscore (for HTML templating).
Lines 7-11 connect to the database server, set up the database connection and define a model (similar to a MySQL table).
Line 13 reads our Underscore template in. Writing lots of inline HTML is no fun at all.

Lines 35-40 set up the server with some Connect middleware layers. The Connect module bodyDecoder makes the data from the form POST available without extra work — very handy. The router module makes URL-based parsing super easy.

Line 16 is called by the router when a client sends a GET request to ‘/’. The anonymous function outputs a header (Line 17); queries the MongoDB via Mongoose and sorts by time, descending (18); debugs the query to the server output (19); and outputs HTML after the query object is fed through the template. The HTML includes a form before any of the query outputs — this form POSTs to ‘/’ with the new “chirps.”

Line 24 is called when that form POSTs to ‘/’. The request is then logged to the server output (25). After testing that the POST makes at least a little sense (26), it builds a new “record” to be “inserted” into MongoDB, then saves it (27). Line 30-31 issues a redirect to get the browser to do a GET on ‘/’, showing the new data.

And that’s pretty much it!

Side note: Paul and I spent a long time trying to figure out how to sort in Mongoose. For some reason, it takes sort arguments completely differently that Mongo does. The difference is shown in commented-out line 18 and line 19.

Here’s the template file, if you’re curious:

Gear 5D Mark II, 24-70 2.8, 70-200 2.8 IS

Edit Transcode to 720p and ProRes422 with MPEG Streamclip. Open in FCP 6, sync the audio tracks, make a multiclip, cut really fast between the angles (maybe even ‘wildly’) and try to be aware (at least a little) of the beat and the lyric phrasing. No grading. Not very thoughtful, not very methodical — mainly just learning and practicing shooting and workflow. Good times with Drew, too.

Head over to Vimeo to see it in HD.

Drew Phillips: “I Touch Myself” from Ivar Vong on Vimeo.

Shooting

I shoot JPEG and RAW on my 5D Mark II. For the newspaper, JPEG is smaller, faster, and (usually) doesn’t make much difference on newsprint. I switch to RAW when there are different color temperatures in the shot, large dynamic range (or any difficult exposure), or anytime if I know I’ll want more flexibility in post. One could argue (and quite well) that I should shoot exclusively in RAW. Maybe I should. A lot of newspapers and newspaper photogs shoot work only in JPEG for size and speed. For now, it’s still a mix between the two.

Upload

I’ve got a Lexar Firewire 800 reader. It’s awesome. I usually use Photo Mechanic’s “Ingest” tool to do batch metadata-ing on the fly.

Edit

Photo Mechanic is incredibly good at doing the edit — it’s very fast and has very good keyword support. I use a folder- and filename-based organization system. Folders are named YYMMDD.initials.slug (July 19, 2010 is 100719, for example). This makes them sort correctly in the file system, removing dependence on catalogs and file locking.

Initials maintain easy photog-identification when dealing with multiple photographers (such as at the Emerald).

If I tag images at all in camera, I’ll pull those out first, then go through the rest and tag everything that’s useable. Finally, I’ll sort into three bins. If they’re for the paper, I’ll write cutlines for the ones I’m going to file (this is usually the top color class).

If I shot in JPEG, I’ll copy (never edit the originals!) these selects to a new “edits” folder to do basic adjustments in Camera Raw / Bridge / Photoshop. They then get exported to another subfolder with the edits applied. Those files are then delivered via FTP, email, SmugMug, etc.

If I shot in RAW, I’ll put the edit through Adobe DNG Converter and save them in a subfolder and continue with the above workflow (ACR, Br, Ps) and export to full-quality JPEG or TIFF. After I’m done, I usually delete the DNGs for space, rather than the RAWs, so I always have the pristine originals.

Backup

I have two 1 TB LaCie eSATA drives. One is Calvin, one is Hobbes. Calvin is the working copy of the archives. When I write a project off of my laptop to Calvin, I then run the following script. It syncs any changes across the two drives in both directions.

Here’s the little BASH script that syncs the two drives together:

#!/bin/bash
echo "Syncing Calvin to Hobbes..."
rsync --progress -av --exclude=".*/" /Volumes/Calvin/ /Volumes/Hobbes/
echo "Syncing Hobbes to Calving..."
rsync --progress -av --exclude=".*/" /Volumes/Hobbes/ /Volumes/Calvin/

It functions similarly to a mirrored drive setup, except each drive exists perfectly happily by itself. For this reason, I can take Calvin with me into the field or the office, work on the drive, copy files, then come home and sync everything up. This allows physical separation of the drives for additional safety and two copies of the images at all times, except when I’m shooting.

Here’s the order, with where the images are stored during each step:

Shoot. (Images on CF)
Transfer to Macbook Pro. (Images on CF, MBP)
Edit, file, then copy shoot to Calvin (Images on CF, MBP, Calvin)
Sync Calvin and Hobbes (Images on CF, MBP, Calvin, Hobbes)
Format CF. (Images on MBP, Calvin, Hobbes)
Delete from MBP (Images on Calvin, Hobbes)
The images that were uploaded provide another layer of backup, whether an FTP server, email, or SmugMug, but I hesitate to put my full trust in them.

For me, the takeaway is: keep at least two copies at all times. Think about what would happen if right now, your hard drive died forever. It would be sad, but it shouldn’t be tragic. Plan for it. It happens, but that’s not a good excuse for losing important data.

I rode Highway 242 with Lisa Turnbull last Tuesday. It was at the top of my list in terms of rides in the region I haven’t done — and for good reason. It was an absolutely gorgeous ride, albeit it very hard and long, especially considering my lack of riding in the recent… well, long while.

We drove out to the ranger station at McKenzie Bridge after the end of the Tour stage, kitted up and rolled out. You get three miles of flat, hit 242, and it’s up and up and up for a full two hours (at my pace).

My PowerTap was giving me watt readings that looked reasonable, but I don’t trust it at all anymore. It’s been wandering randomly, refusing to stay zero’d, and needs to be serviced. (For those interested, it was saying 180-200 W most of the time I was looking at it) Anyway it seemed to behave a little. I concentrated on my HR, trying to keep it below 174 bpm. There’s no real magic with that number because there were so many variables. When I was (more) in form, my threshold was around 181-184 bpm with a 20-minute max around 300 W.

But this day was with only five training rides in the past few weeks, and two before that since January. Yowtch, needless to say. And there’s elevation, heat, etc. But 174 bpm up the climb seemed like a noble goal — it was giving that uncomfortable tightness in your quads, where you might be able to sustain it a hair harder, but you keep just keep it rolling.

The last 30 min gave some relief in the form of rollers and false-flat, but you still put in ~3500 feet of climbing in 25 miles. (Figures are only approximate, memory is only… approximate). It was definitely good that the hardest part of the ride was first.

The top! At last! I have never climbed that long.

We took a little break and put on some extra clothes on for the descent, even though it was comfortably warm at our 5300′ perch. I took a few pictures on my Droid, and while it’s put to shame by my 5D Mark II, it gives a glimpse into what it looks like up there:

It was a great time rolling down the other side and into Sisters. Not much that’s technical, pretty clean roads, no traffic. We hit up an awesome little bakery and got some cheese bread, coffee, and refilled our bottles.

The next hour and a half was the least enjoyable, especially the first part: it’s a half-hour slog back to the bottom of the climb on the straight, false-flat uphill. The scenery got a lot better when we got onto climb. I was definitely feeling my legs more and more. My HR was a little lower, right around 169 bpm, but I feeling pretty tired and probably about a cog slower. Lisa kept riding away from me, which was pretty funny. Again, lots of variables (fatigue, heat, altitude, lack of a baseline) make it hard to really analyze the numbers.

The good news was that when we got to the top, the work was over — just 25 miles of downhill to the car. Needless to say, I survived it up the climb and even gave it a little gas near the top. Success!

Looking down to see 65 miles down out of 80 was slightly disheartening, though. I felt like I’d ridden a lot farther than that. But the miles flew off as we were rewarded with one of the best descents I’ve ever seen. And it took a full hour. An hour, more or less, of optional pedaling. Just superb. Three miles back to the car, chamois off, home. One of the most amazing days on my bike ever, for sure. Highly, highly recommend.