by W.A. Steer PhD
Back to contents About...

Last year I generated a series of photographs and test-images which I took around a number of local stores and photographic shops offering digital photo-printing services, to evaluate them from a technical perspective. This page is an update on new findings and understandings revealed since then.

<<< Original RESULTS (previous page) See previous page for the original technical results.

Further results, one year on

[I have done my best to match my scanner settings to those used for the original pictures, but I cannot guarantee absolute like-for-like comparison of the images shown here.]

Jessops

Shortly after doing my original write-up, I found that by handing my CD over the counter to be loaded directly onto the Frontier (instead of going via the Pixology kiosk), the second resampling step could be avoided, resulting in marginally sharper pictures, and one less step of variability. Of course the flip-side is that by not using the serve-yourself kiosk it's not very practical to pick and choose which prints you want, or which you want multiple copies of etc.

Unfortunately the excellent Frontier-operator at my local shop moved on to better things in November 2005, and on the few of times I've had prints since then, their prints have emerged with varying degrees of green cast - particularly noticeable on greyscales or black-and-white prints. :-(

Jessops / Frontier
Same store, same Frontier: top image from July 2005 and representative of consistent results with that operator, bottom image from a recent visit in February 2006. Note the horizontal smearing as pictured (especially from the white bar) is caused by veiling glare in my scanner and does not exist on the prints themselves.

Jessops / Frontier
As above, Colour diagrams from July 2005 (top) and February 2006 (bottom). The recent greeny tint is obvious in the grey background and greyscales.

Photo-Me (Kis / Kodak)

In March 2006 I followed up an invite to visit a Photo-Me service base and have a look at some of their machines, and even see inside some!

Photo-Me/Kis/Kodak DKS1550 mini-lab machine

This is the Photo-Me/Kodak answer to the Fuji Frontier. In contrast to the Fuji Frontier which uses a laser scanning system to expose the photographic paper, the DKS1550 uses a halogen lamp and LCD exposure system, making it a bit more like a traditional enlarger. Using piezo transducers it steps the LCD around in fractions of a pixel and makes multiple exposures to enhance the resolution. It uses as optic element which Photo-Me describes as "MicroLens Sharpening Beam". I can only imagine this device leads to pixels from the LCD array casting proportionately smaller points of light on the photographic paper than the true LCD pixel aperture... this would then make sense with the LCD array piezo-stepping, allowing a real increase in resolution rather than just blurring the gaps between the pixels. My gut engineering feeling is that Fuji's laser exposure system is a nicer approach and more easily scalable to larger print sizes, but the Photo-Me/Kodak LCD system appears to be quite capable and I'm told it's much cheaper and less prone to breakdowns!

The DKS1550 was loaded with Kodak paper for all the prints featured here.

At present, this page represents INITIAL COMMENT ONLY on the DKS1550.
I've presented the key images, but have not yet had time to do a fair side-by-side comparison of the actual prints with those made by the other machines last year.

Photo-Me DKS1550
Reasonable: the image looks a bit low-contrast here, but I need to put the original prints side by side before making any judgement. Also I only have one print to hand; we made several with different colour settings - I don't know for sure which version this one is.

Photo-Me DKS1550
This pair of images is very interesting and shows at the top the colour diagram treating my source image as if in the native DKS1550 colour-space, while the version at the bottom was supposed to be printed using a 'proper' conversion from my original sRGB colour-space.
Curiously, the colour diagram image has revealed an unexpected printing issue: fine lines or white text are sporting halos against the black background (more on this below).

My original colour graphic is designed for emissive displays (rather than reflective prints), and exercises the absolute boundaries of sRGB colour-space - an accurate conversion to a photographic print may simply not be possible. As a general rule, ignoring colour spaces and treating any image as if in the native space of the printer (top image) will result in the most saturated and 'punchy' print, even if the colours are inaccurate. The gamut-mapping (lower) approach will try to keep the colours accuarate where possible (particularly for less-saturated colours) but very often compromises inevitably have to be made with more saturated colours. [This comparison suggests that the Kodak Digital Photomaker kiosk (see original results page) may also be trying to do accurate gamut mapping.]

Photo-Me DKS1550
Close-up (300dpi) scan showing halos around my fine text and lines on the Colour Diagram print

Fuji Frontier (July 2005)
Close-up (300dpi) scan showing very clean reproduction of detail
In both cases please forgive the spots/fluff on the images - I forgot to clean the glass of my scanner before doing the scans!

Now my Photo-Me guide suggested this artifact may be caused by the LCD stepping. I wasn't entirely convinced, it being a fairly serious shortcoming - and bigger than 1-pixel radius anyway. Since the photographic paper is a negative medium, again you wouldn't normally expect bleed from bright to dark areas in the final image. Looking at a number of my graphics prints later, it appears that the halo-ing is much worse on prints from lower-resolution source bitmaps (eg the Colour Diagram at 1800×1200, and less), and much less obvious on those from higher resolution originals (eg 2100×1500 upwards). It's a bit of a long-shot, but I'm tending to think this artifact may originate in software or software image-scaling algorithms?

The zoneplates and scaling tests will inevitably prove tougher DKS to manage as at 300dpi they are not so well matched to the 340dpi of the DKS as to the Frontier.

Photo-Me DKS1550
The grey is quite neutral and the contrast ok, but there's quite significant evidence of rogue rings. It's possibly not helped by the mis-match of resolution of my source-image (300dpi nominal) to the DKS1550's 340dpi, and this is particularly evident in the reproduction of the pixel-on/pixel-off lines along the top and left.

Photo-Me DKS1550
How would you describe it? The fine diagonal (curved) lines just don't have the precision of the Frontier prints from my original review, but it's difficult to know how much the DKS1550 has been handicapped by having to work from a 340/300dpi resolution mismatch. The disturbances to the finest row of black/white lines do seem to point to less-than-perfect resampling, and the vertical disturbance to the bar is very strange... I suspect this may be another manifestation of the same issue which causes the halo-ing around white text on the black background.

One final oddity I noticed with the DKS1550 is that some prints, under some circumstances, had detectable "raster" lines. You'd need to look very closely (or with an eyeglass) to see this, but it was particularly curious that it was present on some prints and virtually absent on others, even for duplicates of the same image. Again it's guesswork, but I'd look at the piezos stepping the LCD and check whether there's some 'creep' (or other mechanical/thermal/vibration problem) causing some of the multiple-exposures to slightly mis-align. Alternatively maybe something in the LCD or shutter is giving non-uniform exposure times for the sub-exposures?

"Raster lines" on a print (600dpi scan for illustration). In this case running parallel to the short (4") edge of the print.

We did scan two frames from a 35mm negative film. While not enough to necessarily be representative, results were fairly encouraging and the negative scanner may yet prove better than the (frankly disappointing) Fuji.

More information on the Photo-Me (and Kodak) kiosks

Photo-Me kiosks use thermal printers made by Mitsubishi, and apparently the 'vertical blurring' I noticed last time (on both the Photo-Me and Kodak kiosks) is a known issue with thermal printing technology. Apparently in high-speed prints the head does not have time to cool down properly as the image is made, resulting in blacks smearing slightly into whites in the direction of travel of the paper on abrupt edges! Illustration to follow (perhaps).

Where next?

I think I'm saturated with printing tests for now! New tests I'd design if I were repeating the exercise would include:

CIE chromaticity diagrams at the present and at reduced intensities (eg 100% 75% 50% 25% reflectance)
Simulated spectra ("rainbows") again at full and (maybe just one) reduced intensity. Again this drops hints about what's going on with colour-mapping. Maybe add R/G/B / C/M/Y patches at full/75% intensity to enable colour-points to be measured.
Try one or both of the above with files which are untagged, and files which have IEC colour-profile attached.
Zoneplate patterns rendered for nominal resolutions of 256/300/340dpi and 6Mpixel (unmatched, high) to all be sent to all machines, to ensure a level playing field
White blocks on black and not-quite black backgrounds, Black blocks on white and off-white backgrounds to check for printer smearing and for strange impulse-response functions in image-processing algorithms. Possibly at low/med/high file resolution to distinguish image-processing from fundamental printing issues.
Pairs of images with full vs clipped greyscale/photo-levels - to reveal which machines/processings use auto-contrast or image-dependent enhancement functions.
High-contrast black-and-white photos have proved useful.
Use some very high quality high-res pin-sharp photographs (negative-scans from my Nikon LS-50) to check general photo- quality and sharpness.

Feedback

I am interested to hear your feedback or comments on my analysis, or your own experiences with digital printing. Have you found these pages useful? I'd especially like to hear from anyone employed in the design of digital photo-printing equipment or software for it!

I'd also be interested to hear from any magazine which might be interested in collaborating with me to do a review...

Created: March 2006
Last modified: 14 March 2006
Source: http://www.techmind.org/photoprinting/resultsnew.html


Jessops / Frontier Same store, same Frontier: top image from July 2005 and representative of consistent results with that operator, bottom image from a recent visit in February 2006. Note the horizontal smearing as pictured (especially from the white bar) is caused by veiling glare in my scanner and does not exist on the prints themselves.

Jessops / Frontier As above, Colour diagrams from July 2005 (top) and February 2006 (bottom). The recent greeny tint is obvious in the grey background and greyscales.


Photo-Me DKS1550 Close-up (300dpi) scan showing halos around my fine text and lines on the Colour Diagram print

Fuji Frontier (July 2005) Close-up (300dpi) scan showing very clean reproduction of detail In both cases please forgive the spots/fluff on the images - I forgot to clean the glass of my scanner before doing the scans!