Posts in Essays
Dynamic Range and Latitude Are Two Different Things

I've been meaning to write this one for a while. In fact it's been sitting as an empty draft for months, so it's about time we get on with it. One of my pet peeves is people who act like they know everything when they don't know what the hell they're talking about. A few years ago I wrote one of my favorite blog posts ever called Image Properties (Or how most people talk out of their ass), to tackle one of the common areas of confusion.

Today, in an attempt to help out some more of you we're going to talk about the difference between 'dynamic range' and 'latitude'. Two terms that many people use interchangeably which are actually two different things.

An image of trees in Olympic National Park which requires a lot of dynamic range.

Dynamic Range is a measurement of the size of the range of tones that a camera can record from completely black to completely white. In simple terms, the greater the dynamic range, the more detail you can see the shadows and highlights of a picture. In the old days, different film stocks had different dynamic range properties, especially from type to type. As a rule of thumb traditional black & white film has the most, then color negative film, and finally color slide film the least at about 7 stops. Now most of us use a digital sensor in the place of the film and so it's this sensor that now has the properties of dynamic range. So if you want better dynamic range, you've got to buy a new camera. The good news is that today's cameras are really pretty great in this regard, most within the 12-13 stop range which is as good as all but the best film with the best scanners. 

I'd like to take a second and discuss the dynamic range of film and why it's so hard to measure and compare with digital. You see, digital sensors are designed to be very linear, that is, if you add a stop more light the sensor will record the image a stop brighter. The down side is that when the light coming in overwhelms the sensors ability to record more light the highlights in your image "blow out". Digital doesn't do this very pleasantly and it's definitely one of the few things they've really got to work on. Film on the other hand (and this goes for audio recorded to magnetic tape as well) is mostly linear for most of the range. However at the highlight end of the scale the film pushes back, and when you add a stop of light and then another, the film records less and less of an increase. It's got sort of a built-in compression curve which transitions highlights to white very very smoothly. It's very pleasing to the eye, but that same effect also makes it very hard to measure how much dynamic range the film is accurately recording. Do we count those top few stops that get compressed as 'accurately' recorded? I'm not sure.

Either way, let's imagine that there's a ruler that has black at one end and white at the other. And since I live in America and there are about 12 stops of dynamic range in a sensor, I'm going to use a one foot ruler as a visual guide. Imagine that this ruler is broken up into 12 equally sized inches each representing one stop of brightness that the sensor can measure. The 0 on one end equals black and 12 at the other end equals white with grey in the middle. Easy right? Ok, hold onto that, we'll get back to it in a minute. 

Latitude is the other term which people use interchangeably when they typically mean dynamic range. "That new camera has better latitude than last year's model", I'm sure you've heard that kind of quote before. What latitude actually refers to is how much a picture can be over or under-exposed while still getting the image you were looking for, and that very much depends on the image you're trying to take. 

So for example if you're taking a picture which only has 8 stops between it's darkest and lightest areas, then your camera's 12 stops of dynamic range is overkill. That means that you've got 2 stops on both the shadow or highlight ends which are technically going unused. In fact if you overexposed by a couple of stops and then pulled it down by 2 stops in post, especially if you're shooting RAW files, you should end up with an image that looks just like one which was exposed correctly. In this instance you could say that you've got '2 stops of latitude'. 

To use the ruler analogy from above, if the picture you're trying to take is only 8 inches long, you could measure it from the 2" tick to the 10" tick on the ruler, or from 0" to 8", or 1" to 9", etc. As long as it all fits inside the twelve inches of the ruler, then you'll be fine.

Back in the film days, when you dropped off a roll of film at your local lab, a tech or a computer would make these kind of corrections for you when making your prints so that everything looked correctly exposed. It was the fact that film had a lot of dynamic range which allowed people to take all kinds of incorrect exposures and still delivered great looking prints. This fact was used to great effect in disposable film cameras which had a single shutter speed and aperture. They relied on the fact that you could usually get a 'usable' image regardless of how much light actually hit the film, especially when they could just pop a built-in flash.

Of course in the real world there are consequences to incorrectly exposing your image. Even if it fits into the dynamic range of the sensor, a stop or two (or on the odd occasion three), is about all you're going to get back in either direction and you'll likely have some artifacts like shadow noise. We've also been talking about this from the very basic point of view of monochrome luminosity. However real images have three color channels of Red, Blue, and Green each of which have a 'ruler' in them. So you might blow out one of the color channels and not the others, and so forth, which further complicates things like recovering highlights. It all varies from image to image.

So to wrap up, Dynamic Range refers to the size of the range of light that a camera can capture from complete black to complete white, while Latitude is the amount of under or over exposure that a particular image can handle while still fitting into the camera's dynamic range.

Now you know, and knowing is half the battle. Go forth and school some friends.

Tag Team Backup - Digital Photography Workflow

Years ago I wrote a post on my file workflow. That is, what I do with my files once I pull them off the card to make sure that they don't disappear. Since then I've made some changes to my workflow so I thought I'd write a little update to that old post. One of the major problems with modern digital photography is that we tend to take a lot of pictures and need somewhere to put them. Strangely enough, I don't shoot that much in comparison to most photographers, even many amateurs. For instance, my typical editorial shoot is 150 images on average. I have some event shooting friends who take more pictures in a day than I shoot in a month. So all together once I do all the math, almost everything I've ever shot can fit on a little over 3TB. Nothing for most photographers, I know. I still need to work on those files and backup my data however, so here's what I've come up with that works for me.

The Jump to a RAID Array

I found myself waiting for 2GB heavily layered PSD files to be read and written to disk and so started looking for ways to speed up the process. The thing is that I'm a real stickler for noise and so moving to 7200RPM drives, which I find much more noticeable, was a no go. I'm also not made of money so the idea of swapping out all of my photo drives for SSD is not yet a reasonable solution (though it may be soon, more on that below). So the answer I came up with was to bond two of my WD 2TB Green drives together in an OS X software RAID-0. This doubled my throughput to around 180MB/s which is pretty good. Reducing my save/load times by almost half. Of course the big problem with RAID-0 is that if either of those drives died, all of the data on both drives dies. So when you play with RAID-0 make sure you have an extra special backup strategy in place.

Off Site First

Screen Shot 2014-01-06 at 12.16.51 PM
Screen Shot 2014-01-06 at 12.16.51 PM

The most important part of backup is to get data off site. So if your house burns down or gets pulled down a river, you've still got your data. Now for me with a 5Mbps upstream connection, having a true one to one backup of my data drives up on the cloud just isn't a reasonable thing to do, it would take months and months to upload. And honestly, if my house burns down, do I really need the RAW files for outtakes that didn't make the cut in the first place?

So a while back I instituted a system of exporting my final images as full-res jpegs at 85/100 quality and uploading them to Dropbox.  This both gets them out of the house AND allows me to access final print-ready copies of my work when I'm out and about or on vacation. I can send email links to any of the files right from my phone. So it's convenience AND backup for which I pay $100/year.

Plus, instead of 3.2TB of data, my entire 'Finished Images' folder, everything I've ever shot that I care to keep, totals a whopping 20.5GB  I could keep a copy local on my phone if I really wanted to. Or on a keychain USB drive I guess, that's not a bad idea actually...

Tag Team Backup

The proliferation of inexpensive USB 3.0 drives has been a great boon to backup users everywhere. For one thing they're cheap. Often cheaper than the bare drive that's enclosed within goes for, and they're seemingly always on sale somewhere. I've bought two 4TB Seagate drives in the past few months for about $150 each. That's just nuts. Here's one on Amazon right now for instance:

Seagate Backup Plus 4 TB USB 3.0 Desktop External Hard Drive on

The trick to my backup solution is to have two drives to backup to, but only backup to one of them at a time. One drive is on my desktop, the other is in my closet. Once a month (I switch them when I write my rent check) I swap the two drives so that the one in the closet becomes the one on my desk and vice versa.  You may ask 'Why?', but I assure you there is a good reason for this.

Once you have a backup system working, the nightmare scenario is that data on your main library drive becomes corrupt or something gets accidentally deleted and an automated backup goes and clones those mistakes to your backup drive before you realize it. So now you're left with not one but two drives which don't have your data on them. By having two backups that you swap in an out, you always have a backup that's not going to be automatically overwritten which is no greater than a month old (or a week old if you swapped them weekly, or a day old if you swapped them daily, etc). Another ancillary benefit is that the drive in your closet is not connected to power, so that if that random power surge or lightning strike kills your electronics, your data is covered.

Image Library on SSD

I mentioned above that the idea of putting all of my images on fast SSDs had occurred to me. The prices of the drives has fallen A LOT in the past year or so. To the point where you can currently buy a 960GB drive for $500. Still a little too rich for my blood, but if I archived the old stuff to a couple of external drives and kept my library tidy, I could probably get it to fit within 2TB or so. And that would only cost about a grand. A lot of money? Sure, but not completely astronomical like it would have been a couple of years ago.

It's certainly to the point that the next time I build myself a new computer, I'll probably make the switch. Hopefully by then the price will be down to $250/TB.  Imagine two or three of those drives as your RAID-0 array. Loading images at 1.6GB/s would be pretty nice. Necessary? Nah. But pretty nice.

On Eleanor

I got some terrible news last night. My good friend and muse Eleanor had died suddenly while visiting family for the holiday. No warning, no sickness, no time to say goodbye. She was one of the outliers the entire time I knew her. I would often use her as an example in conversations when I needed an example of someone who confounded the world's expectations. Sassy to the nines and always so incredibly full of life that I can hardly believe that she could possibly be gone. If you needed another reason why you should never wait to do that thing you've been meaning to do, or tell someone that you love them, here it is.

Below are a few of the photographs I made with Eleanor over the years.







Family Photo Restoration


My partner brought a little present home for me from her mother's house. Less a present really, more of an assignment. Her mother had found this picture of HER mother and sister and grandmother in some drawer or other. It had been battered and beaten and probably put through the wash a few times, but they wanted to see what I could do with it.

Step one was to scan it in order to get as much information as possible out of the original. Pulled my old Epson flatbed out of the closet and plugged it in, fired right up. Scanned it to a TIFF file at 1200dpi. Not that there was anything near that much information in the print, but I find that when doing restorations like this, the higher resolution let's you more easily discern between the image and any physical flaws that have befallen the print. If you don't have a scanner, you can also take a well exposed picture with your camera and start with that. Just make sure you light it from the side so you don't get reflections in the image.

Most of this kind of work can be done in Photoshop using things like the spot-healing brush and stamp tools. Certainly when it comes to creases across a largely white sky those techniques work pretty flawlessly. The problems come when you need to invent information. The places on the print where the image has been torn away for example, that's information I have no way of getting back. For things like the pattern in their dresses, you can use the healing tools to mimic or clone in the pattern from elsewhere to good effect. Other areas like the swing set to the right of the central girl, I can't accurately recreate that. Best I can do, within reason, is to use the surrounding image to guess, and that's just what I did.



Overall not bad considering I only spent an hour or so on it. Is there room for more work? Sure, but you're quickly approaching the limits of your return on time invested. We want to save and perhaps restore the memory a bit, no need to go all Ken Burns.

I also could have increased the contrast even more and desaturated the whole thing entirely. That would probably deliver an image that was much closer to how the print looked in the 1930's, but also kind of loses some of the Age that the print has imbued on the memory. So in the end I pulled back the contrast and desaturation layers to let it feel a bit more like the original.

Either way it's a good 'waiting for the snowstorm' project, and a great way to get your feet wet in PhotoShop. You'll learn how to use layers and healing/clone tools, as well as adjustment layers and color. I highly recommend you get a tablet to do work like this. Trying to do this with just a mouse would be like doing a fine pencil drawing while wearing ski gloves. The less expensive Wacom tablets are a great deal, and a good place to dip your toe in the water.

So go rummage through some drawers and give it a shot. Your children and your children's children will thank you some day.