Q) Which is better for scanning – negative or positive film?

A) Generally speaking, color negative (C-41) film is a dream to scan relative to transparency (E6) and black and white films. Color negative not only captures a huge dynamic range but compresses it to a very small dynamic range that’s easy for scanners to extract.  Color negative film allow allows the photographer to print in either color or black and white. Images can be scanned in color mode and in front of the lens filtration (yellow filter, red filter, etc) can be simulated during the process of converting the image to a grayscale mode. Fuji has concentrated recent efforts on tweaking their new color negative films to be even more optimal for scanning. Some B&W neg films (like TMAX 100) can be problematic because they can exceed the dynamic range of the scanner so you have to loose some of either the highlights or shadows.

Now, there are always exceptions. Some scanner software applications aren’t great at handling color from color negatives. Some drum scanner applications, for example, just don’t know what to do with the orange mask and it would take a super skilled operator to overcome the software limitations to get good scan out of it. Most drum scanning applications were designed to scan transparency films so you might hear an operator say E6 is best for scanning (at least on *his* scanner system). Finally, some people might prefer to shoot their favorite B&W neg film for the sake of the unique grain structure.

Imacon/Hasselblad and Nikon scanners are great for color negative scanning and with those scanners you’ll probably want to shoot color neg, especially Fuji’s latest flavors.

The sRAW (small RAW) format is gaining popularity among 20+ megapixel camera owners that occasionally don’t want as much resolution or such a large file size. sRAW files are half to a quarter of the resolution of a full RAW and this equates to smaller storage requirements and faster processing for projects that don’t require as much resolution. sRAW files do maintain a nearly all the development flexibility of a true RAW file, and therefore are often more desirable than a less flexible JPEG.

sRAW files aren’t actually true RAW files. The sensor’s true RAW data is demosaiced and rendered to a lower resolution file and saved in the YCC color mode. For this reason the final file size isn’t quite as small as one might hope (IE: quarter resolution sRAW files are about half the file size of a full RAW file). Unfortunately, when these files are currently converted to DNG, images are converted to less efficient RGB space and the file size increases – negating some of the file size and processing benefits.

The chart adove illustrates the file sizes of a full res (resolution) raw file, a half res sRAW1 file, a quarter res sRAW 2 file before and after conversion to DNG, and a full res large JPEG.

Perhaps the DNG file format will be updated to include YCC sRAW files at their smaller file size. In the meantime, if you wish to keep file sizes small when shooting sRAW, I recommend sticking with the proprietary sRAW format that your camera uses (CR2, NEF, etc).

Apple’s long anticipated 10.6 “Snow Leopard” operating system is now shipping. Snow Leopard’s modern cocoa code base promises system-wide responsiveness, reliability, smarter multiprocessor utilization and multitasking on existing Intel based hardware. Geek Mac and Windows users alike should read this in-depth article on Snow Leopard’s under the hood technologies. As with any major operating system upgrade it is important that we make a list of our applications and printer drivers and check for compatibility before making the switch.

Printer driver compatibility is probably the main issue that will force some users to delay upgrading. Snow Leopard’s CUPS printing structure is different enough from the previous version that existing printer drivers may not work. While Apple has their generic “Gutenprint” driver for many common printers, professional will want to wait for the official drivers from the printer manufactuers. Printer manufacturers are weary of releasing 10.6 compatible drivers before they’ve had the chance to test them throughly with the final version. I’ve been told to expect 10.6 compatible printer drivers for currently shipping printers to be released 2-6 weeks after Snow Leopard’s ship date. Expect a slightly longer wait for printer drivers to older, discontinued printers.

Caution is advised when upgrading to Snow Leopard. So far I have seen several color management issues that involve printing in Snow Leopard. Check for compatibility of your applications and printer drivers at the Snow Leopard Compatibility Wiki prior to upgrading. Users should test 10.6 throughly on an extra computer before upgrading production workstations. At the same time, mac users shouldn’t wait too long – this new operating system is ground breaking and will bring significant advantages that will be further realized in future applications (Creative Suite 5 for example). 100 Gigs of RAM, 100 Terabytes of hard drive space and 100 core processors might sound crazy today but Snow Leopard foresees this future.

Having used several cameras for a lot of night photography lately, I think the Canon 5Dmark2 is quite likely the best camera on the market for night photography right now. Here are some of the highlights:

6400 ISO for testing
Being able to take photos at 6400 ISO has huge benefits for the night photographer. Night photographers tend to waste a bunch of time taking long exposures only to discover they need to reshoot with a better exposure. Testing exposures at high ISOs can save huge amounts of time but doing so has been clunky as it requires complicated math when using a camera that has a ISO ceiling of 1600. Now that that ceiling has been lifted to include 6400 ISO we can make a direct translation from ISO 6400 exposures in seconds to ISO 100 exposures in minutes. For example, a 15 second exposure at ISO 6400 is the same as a 15 minute exposure at ISO 100 (Canon’s native ISO). Nikon users can figure that a 15 second exposure at 12,800 ISO is the same as a 15 minute exposure at ISO 200 (Nikon’s native ISO).

Either way, this means a night photographer can quickly determine the optimal exposure at a high ISO before committing to a long, final exposure at the native ISO. This has provided me with significant time savings and more optimal exposures on the 5Dmark2.

Low noise
To make a long story short, the noise levels are shockingly low, even during very long exposures (like 30+ minutes at ISO 100). Even 6400 ISO images are less noisy than one would expect, (more…)

There have been lots of great reviews on the much anticipated Canon 5Dmark2 camera that’s hitting shelves this month. Phil Holland’s review is particularly impressive. Since the 5D has been a favorite among a number of night photographers, I thought I’d post a few notes about the 5Dmark2 from a night photographer’s perspective. I am considering upgrading to the 5Dmark2 from the 5D myself and have had a little time to play with both cameras side-by-side. Here are a few, quick, night oriented observations: (more…)

I’ve had some good conversations lately with a number color scientists and gurus about colorimeters and display calibration. Several of them perform exhaustive lab tests on a variety of devices, checking for repeatability and the like. The consensus seems to be that while the long time favorite DTP94/Optix device continues to be a solid performer, it’s been found to have some weaknesses when it comes to calibrating some of the recent large gamut LCD/LED displays. As for other devices, the EyeOneDisplay continues to have quality control problems, spectrophotometers have issues with shadow detail when it comes to display calibration and the Huey is an inexpensive solution that demanding users shouldn’t take seriously.

Surprisingly to some, the new 7-sensor Spyder3 has excellent repeatability and it performs remarkably well in a variety of situations, especially with the latest large gamut displays and projectors. It looks like the Datacolor Spyder3 has trumped the Optix colorimeter as the best and most versatile colorimeter made today. Eizo and others are starting to bundle rebranded Spyder3s with their high end monitors.  My own testing has shown that the Spyder3 performs just as well as the trusted Optix and clearly better than the others. I like the fact that the Spyder doesn’t require dark calibration prior to each use and it comes with a nice holder complete with a tripod mount for projector calibration.

While the Spyder 1 and 2 were disappointing for demanding users, I’ve got to hand it to them for turning this product around and making it the best in the industry. Datacolor sells a basic “Spyder 3 Pro” package for $169 and a more advanced “Spyder 3 Elite” package for $249. The hardware is not only the best thing made but the software is really easy for anyone to use and the “elite” version of the software is surprisingly capable with advanced features like L* tonal response curve, black luminance control and ambient light integration.

Integrated Color sells a display calibration product called Color Eyes Display Pro that’s been a favorite among demanding professionals for several years.  Color Eyes Display Pro is now bundled with the Spyder3 device and I find it to render slightly superior shadow tonality and gradations.

For common monitor calibration, I am currently recommending the basic $169 “Spyder3 Pro” package to consumers and the new $299 Color Eyes Display Pro / Spyder3 bundle for demanding professionals.

Note: CRT users should put a black cloth over the monitor and Spyder when calibrating to eliminate ambient light contamination. Those using Eizo, NEC and Quato displays should stick with the supplied software to calibrate their displays. Eizo’s Color Navigator, NEC’s Spectraview, and Quato’s iColor Display have all become quite good and can have advantages over 3rd party solutions. Use the Spyder3 device with these applications when possible.

Contrary to popular myth, white balance should bet set using a white card instead of a gray card. It is, after all, a *white* balance that is being set, not exposure for middle gray. Software developers including Thomas Knoll have confirmed this with me. While the distinction between using a gray or white card may not be huge, I have found it to be somewhat significant in some situations. While using a gray card is better than nothing, I encourage photographers to replace their grey cards with white reference cards like the ColorChecker White Balance, ColorChecker Gray Scale, orginal ColorChecker, QPCard or WhiBal. Non-reference quality white objects, such as a white piece of paper should be avoided because optical brighteners can negatively effect the white balance process.

One hidden benefit to using a white card is that you can also use it to find the optimal exposure. While there are details and exceptions worth discussing, I’ll simplify here by encouraging users to use the lightest exposure that doesn’t blow out the white card as indicated either by a camera’s RGB histogram or it’s flashing, blown out highlight indicator.

Determining optimal exposure and white balance in-camera is a common training topic for my business, espicially for high volume JPEG shooters that seek optimal print results with zero color correction. Naturally, the same concepts apply for RAW format shooting as well. So switching to a white card and not only helps improve white balance accuracy but also helps photographers fine tune their exposures so as to minimize post processing.

G7 is a proof-to-print process developed by the IDEAlliance that uses spectrophotometry to calibrate presses by the numbers to specifications like SWOP, GRACoL or FOGRA. Previous press calibration methods relied upon basic densitometry, dot gain and visual analysis and could take days of press runs and tweaking (more…)