The File That Looks Nothing Like the Shot You Took

I pulled a raw file up on my laptop at a client walkthrough once and watched the art director’s face fall. The image looked flat, greenish, and lifeless compared to the JPEG preview she’d seen on the back of my camera two days earlier. That preview was the camera’s own JPEG rendering, cooked with the manufacturer’s in-body processing. The raw file was the actual data, unprocessed, waiting. She thought something had gone wrong. Nothing had gone wrong. But I learned that day that “shooting raw” means nothing without a disciplined processing workflow to back it up.

Raw files are promises, not photographs. If you hand a client a folder of unprocessed raw files and call it delivery, you haven’t delivered anything. You’ve handed them a problem.

What a Raw File Actually Is (and Why That Matters for Your Workflow)

A raw file, whether a Canon CR3, Nikon NEF, or Sony ARW, is sensor data captured before any in-camera processing. No sharpening, no noise reduction, no white balance baked in. The file records 12 or 14 bits of data per channel, compared to 8 bits in a JPEG, which means you’re working with somewhere between 4,096 and 16,384 tonal values per channel instead of 256. That’s not a marginal difference. That’s the reason you can pull two stops of detail back from blown highlights on a raw file and get clean, usable information, while the same recovery on a JPEG turns into gray mush.

But that latitude is only useful if your processing software knows what to do with it, and if you’ve set up your workflow to make consistent decisions at scale. On a 400-image commercial shoot, you cannot reinvent the wheel for every frame. You need a system.

Building a Processing Pipeline That Doesn’t Fall Apart Under Volume

My current workflow runs through Adobe Lightroom Classic, with DxO PureRAW handling noise reduction before import on any files shot above ISO 3200. Here’s the sequence.

Cards come in and go straight to an ingestion script using Photo Mechanic Plus, which copies the files to my primary working drive and simultaneously to a second drive. Right there, before I’ve opened a single file, I have two copies. That’s non-negotiable for me, and I’ll explain why in a moment.

From Photo Mechanic, selects go into Lightroom. I apply a camera-matching profile immediately, not Adobe Color, but the camera-specific profile that matches the manufacturer’s rendering. For Canon work, that’s the Canon Picture Style profile. This makes the raw look closer to what I previewed on the back of the camera and sets a more accurate baseline before any corrections.

My import preset does the following automatically: lens correction on, chromatic aberration correction on, sharpening at 40 with a radius of 1.0, masking at 25, and noise luminance at 0 (because I let DxO handle anything above ISO 3200 before it even gets here). Clarity sits at plus 10 as a starting point for commercial work. Portraits get their own preset with clarity at 0 and a slight negative texture value.

From there, I batch-correct exposure and white balance across similar lighting setups before touching individual frames. On a 400-image day, this approach cuts my processing time from six to seven hours down to roughly two and a half.

The Backup Paranoia That Paid Off

I worked at a daily newspaper in my twenties where the deadline was the deadline. You filed or you got replaced, and there was no version of “my files got corrupted” that saved your job. That environment burned a particular kind of discipline into me that I’ve never been able to shake, and I’m grateful for it.

I now keep three backup drives, and I test them monthly by actually restoring a file, not just checking that the drive spins up. One of those drives sits off-site at all times, rotated weekly. It sounds excessive until it isn’t. A few years ago my twins, who were seven at the time and deeply curious about what buttons do, managed to delete a folder of client proofs from my working drive while I was on a phone call in the next room. I had them fully restored from backup in under two minutes. I was irritated for about 90 seconds and then genuinely relieved that the system worked.

When photographers tell me they back up “when they get around to it,” I hear someone who hasn’t lost anything yet.

Export Settings That Match the Delivery Context

Processing the raw is only half the job. How you export determines whether your work survives the journey to the client.

For print delivery, I export full-resolution TIFFs at 300 PPI, ProPhoto RGB if they’re going to a professional lab, sRGB if they’re going anywhere else, including most commercial printers who won’t tell you they’re not color-managed but aren’t. File sizes for a full-resolution 45-megapixel Sony file exported as a TIFF will run 130 to 180 MB each. Budget your storage accordingly.

For web and social, I export JPEGs at 80 quality in sRGB, with the long edge capped at 2,048 pixels. This gives you a file that’s sharp, color-accurate across browsers, and small enough to load without friction. I never hand a client a JPEG at 100 quality exported from a full-resolution raw. The file is needlessly large and the quality difference compared to 80 is invisible at delivery resolution.

For client galleries, I use Pic-Time, which accepts JPEGs and handles its own compression on the platform side. Knowing that, I don’t add extra compression at export. I hand it the best JPEG I have and let the platform do its job.

Raw processing is not a creative afterthought you figure out after the shoot. It’s the infrastructure that determines whether your images are actually usable, whether your delivery is consistent, and whether your business can scale past the point where you’re grinding through every file by hand. Set up the system once, and let the system do the work.