This is an article on printing with a big difference from what I’ve reviewed before. I’m not reviewing an inkjet printer, or inkjet papers, or indeed anything you do in your house for print-making other than preparing the image file. I’m reviewing a recently developed printing process in the UK that is offered commercially – you prepare the image files, you send them to LumeJet and they return finished prints to you, along with an output report on printing accuracy. LumeJet approached LuLa to have a look at their process and consider a review. Having examined the input and output characteristics of this process, we decided it would be of sufficient interest to LuLa readers that I should delve into it and write it up. So here we are.
The details (print sizes, photo-book making service, prices) of what they provide are readily available on their L-type website (aimed primarily at professionals and users looking for a higher-volume service) and their consumer-oriented site. Our primary focus will be on the technology and the results. The immediate questions are what their printing process is, and how it compares with high-quality inkjet prints in respect of gamut, colour accuracy, smoothness of tonal gradations, deep quartertone shadow detail, highlight retention and apparent sharpness.
The LumeJet Process
Their team of engineers and lab technicians has spent more than 15 years developing their “L.Type” silver halide print process. Their intention was to apply the very latest digital technology to silver halide, which is what they consider to be the best traditional photographic medium.
Before going further, it’s useful to review the basic differences between half-tone and continuous tone printing, as inkjet printers use the former and the LumeJet process uses the latter.
Ink or toner-based printers use a half-tone process. This was originally developed in the 19th century as a way to reproduce black and white photographs in newspapers. It defines shades of gray with patterns of dots of different sizes, shapes or spacing, which could be printed using one ink. Halftones simulate continuous tone images by optical illusion: the human eye blends the dots into single tones. Colour halftone printing groups different coloured dots close together into rosettes to fool the eye into seeing colours in a way that avoids unwanted visual effects such as moire patterns. So to print a single image pixel representing one hue may require many ‘dots’ of different-coloured inks using the half-tone method.
Continuous tone photographic images (Contone) on silver halide are produced via a wet chemical process that uses the reaction of light on substances in the photo-sensitive paper. No ink is involved. This produces smooth pictorial graphics – specifically with the LumeJet process, photographs capable of over 4 billion colours (32-bit colour) within its gamut and without a dot structure. With Contone, there is no need for creating optical illusions using rosettes of dots, as done for inkjet printing. Each image pixel is printed with a single printed ‘dot’, which can have an effectively infinite variety of tone and hue. As a result of this, LumeJet claims that its 400dpi print quality is greater than that of multi-colour inkjet printing at over 4000dpi. We’ll examine this claim in the evaluation discussion below.
According to LumeJet, Contone has had a softness that makes it less successful for text and fine detail. LumeJet says its “L.Type” print addresses this issue through exceptionally sharp imaging – giving traditional Contone benefits; but with greater accuracy, gamut and colour fidelity.
The technical basis of the printing process is the new LumeJet-designed digital print head and printer: the product of more than 15 years of development. This prints on the full range of Fuji Professional papers (including matte, glossy, silk and luster papers), using a standard C-Type wet chemical process. The front half of the machine consists essentially of a paper cutter (which cuts the sheet to a fixed length of 1m), a paper path and two print heads, each of which prints on a 6mm wide swathe. They are mounted together, thereby printing 12mm at a time. The print heads image at about 1.5m/s across the paper, the paper moves forward 12mm and the print heads come back. It’s like a single-pass inkjet with light. Once fully imaged, the paper passes into the rear half of the machine, where it goes through a film processor and is developed. It emerges fully dry.
According to LumeJet’s technical literature, printing accurately on silver halide requires that every pixel is produced individually and imaged directly from above. Their custom-designed print head prints at 400dpi continuous tone in width and height for all print sizes and, according to LumeJet, pin sharp edge-to-edge down the length of a 1m print. The printhead is proprietary. It has an array of 576 individually-addressable R,G,B LEDs whose light is channeled through a patented optical fiber taper (Figure 1) to produce a 5:1 reduction. This means that the original LEDs, which are 300μm in size, are reduced to about a 60μm dot on the page. Because the light from each original LED passes down many very fine optical fibers in parallel, the end dot on the page is shaped very precisely. This allows LumeJet to create what they call the ‘pixel unit cell’ (Figures 2 and 3) in which each pixel of an RGB image exposes the 3 layers of the AgX emulsion one on top of each other in a vertical stack (‘dot-on-dot’), rather than dot-beside-dot that one obtains with a halftone printer.
Every pixel is created as a unique exposure for every one of the 576 RGB LEDs. This is finely controlled by digital circuits that give 2048 grey levels/pixel for Red and Green and 1024 levels for Blue (32bit levels). The individual RGB exposures for each of the 576 LEDs are delivered down the fiber taper and projected in parallel so that they image vertically onto the paper.
By placing the RGB pixels to sub-micron accuracy, and with no cross-talk, LumeJet claims key benefits: very sharp text and graphics; very dense blacks using just CMY dyes in the emulsion; very accurate tone control (32bit) and over 4 billion unique colors possible for each printed pixel. LumeJet goes on to point out that the process prints almost double the density of the vast majority of silver halide printers (160k pixels/square inch for LumeJet vs 90k for the others). They chose 400dpi print resolution on the basis (in their view) of it being the finest resolution that a young human eye can resolve at arm’s length.
In this process, there is no distinction between PPI and printer DPI. Each image pixel is represented by one dot, which is a unique colour. Hence a photo having 4000 image pixels by 4800 image pixels prints on a 10″x12″ surface. One pixel in the digital file would be represented by one pixel or dot on the paper. Each dot/ pixel abuts the next one perfectly, and each dot/ pixel is one of the 4 billion colours available. However, they can resize image files of differing resolution so that the input resolution is made to match the required 400 PPI output resolution within the dimensions required for the print. If they need to resample, usually they would take any larger/ smaller image and pass it through their ColorGATE Production Server 10 RIP, which resamples it and converts it to bitmap for the printer.
As the printers print a continuous tone image at 400 PPI, a pixel-for-pixel print from a 30MP camera will natively print 30 million pixels at A3 size (297mm x 420mm). Maximum print size at present is 305mm x 1000mm (11.8 x 39.4 inches). On the full sheet, they print up to 75 million pixels (for the most detailed panoramas). LumeJet advises that they have also printed many portfolios from the latest 100MP Phase One cameras and although these images are down-sampled from their native size, there is no visible loss of image quality.
LumeJet uses Fuji Crystal Archive Professional DPII silver halide papers that are exposed by firing light at them. These are Fuji’s top-of-the-range exhibition papers. They do contain OBAs, but no acid or lignin. Fuji is currently undergoing the formal longevity certification process for these papers. They also print on Fuji HDX Pro papers, which are slightly thinner, designed for lay-flat books and are more fingerprint and scratch-resistant.
The process is fully colour managed; their machines are calibrated every shift. In addition, they print calibration patches alongside every print, which allows them to confirm print accuracy for each print, and they send the results of the per-print testing to the client along with the print. The accuracy data metric is dE(00), comparing the file values of a 31 patch check wedge with the read values from the print of that check wedge (in Absolute Rendering Intent), provided along with the client’s print.
They use BasicColor software for profiling and spot measurement. Their preference for BasicColor over i1Profiler stems from extensive testing of both from which they determined – and this in their view – that subtle advantages favor BasicColor in respect of producing more natural skin tones, smoother tonal gradations, improved shadow detail and improved neutral gray.
According to LumeJet, they can accurately print “pin-sharp” text down to 2-point size, including reversed-out text (a white-coloured text on a black or dark background) smaller than the human eye can read, such that from the densest blacks to the palest whites and grays, very fine details will be reproduced perfectly, and the grayscale will be neutral.
The foregoing discussion draws heavily on the literature and technical advice from the company, which we very much appreciate given the uniqueness of the technological innovations they have been crafting over the past fifteen years. For us now to turn to a comparative evaluation of results – what we see on paper.
Results – LumeJet and Epson SC-P5000 Compared
(1) Papers and profiles
Figure 4 shows the gamut volumes, Black point and White point for the four main LumeJet papers, and for inkjet comparison: Red River San Gabriel Semigloss Baryta and Epson Legacy Fibre, the latter two for the Epson SC-P5000 printer. These are profile readings using ColorThink Pro. Note the considerably wider dynamic range of the Red River SG Baryta for the Epson SC-P5000 (Black point 2 versus 10; White point 98 versus 92). This observation will resurface below in the B&W print comparisons.
Figure 5 is a two-dimensional representation of selected gamut volumes and shapes, to put the LumeJet paper characteristics in perspective.
There’s no doubt that in the non-matte papers here, the Red River paper vastly outperforms the LumeJet papers in respect of Gamut volume, save for a portion of the Blue-Magenta part of the spectrum (Figure 5 lower right corner). Red River’s White point is also considerably more neutral, as we discuss below indicating a much lower OBA presence. For the two matte papers, Epson Legacy Fibre’s gamut volume vastly outperforms that of the LumeJet Type H Velvet Matte, again except for a portion of the Blue~Magenta part of the spectrum.
This information would lead one to believe that prints from the LumeJet process would have none of the brilliance or vibrancy one could expect from traditional inkjet printing using high gamut papers and printers. But save this judgment for later – surprises are in store.
At this point, I should mention that in the LumeJet line, the three DPII papers do not include a true matte paper, even though one carries that name. The matte paper is more like a PK type eggshell luster, similar to the more usual luster papers, such as Ilford Gold Fibre Silk, Epson Legacy Baryta, Red River San Gabriel Semigloss Baryta and the like. The luster paper is a rather beady version of luster, while the gloss paper is very glossy, much like the ferrotyped prints of yesteryear. LumeJet’s real matte paper is the Type H Velvet Matte, which has an extremely low gamut volume and a very weak Black Point. This isn’t a paper I focused on for further testing, as experience tells me that with profile numbers this weak no amount of tweaking in an image editor would produce really satisfying prints from a wide variety of photographs; however it could be good for low key/low contrast photos.
Figures 6 and 7 show where the image file values of the 24 patch GMCC sit relative to the gamuts of the LumeJet Glossy and H Type Velvet papers respectively.
For the Glossy paper, 20 of 24 are well within the gamut boundary, whereas for the Velvet Matte paper, only 15 colours are in gamut. Most of the GMCC patches are of fairly low saturation, all of which easily fit into the gamuts of papers such as the Epson Legacy Fibre or Baryta products. We’ll see below how all this plays out with real-world prints of vibrant subjects.
For testing this, I selected three photographs that I printed firstly in the Epson SC-P5000 (Figures 8, 9 and 10) in the usual way (Lightroom, custom profile) and secondly prepared to LumeJet’s specifications for printing with their process. For the SC-P5000 prints, I used Red River San Gabriel Baryta Semigloss and adjusted the photos under softproof using my custom profile for this printer/paper combination. For the LumeJet files, I prepared them softproofing with their Glossy profile (the objective being to match the Red River appearance as closely as I could despite the much lower gamut volume), resized them to 400 PPI, applied output sharpening using Photokit Sharpener’s Contone 400 Output Sharpener (yes, there is one in that marvelous application!) and sent them off to LumeJet for printing.
When faced with lower gamut alternatives to the Epson SC-P5000/Luster paper combination, the key to optimizing those alternatives is first to edit them appropriately in Lightroom (or other application) in order to replicate the appearance of the dynamic range and vibrancy of the wide gamut option as well as possible. There are of course compromises in doing this. We can pump the clarity, vibrance and contrast, adjust the Black and White points, etc., always being careful to avoid saturation clipping that would smash important image detail, but the histogram shows there will be OOG colours regardless. That said, image appearance between the high and lower gamut options can be made to look quite similar with careful editing. This shows in Figures 8 to 10, where you can observe the differences in the histograms and the adjustments between the wider and narrower gamut options.
Photo 4568 was selected to show primarily green magenta, pink, purple and some orange and lime, while 4591 was selected to show primarily various shades of blue, red, yellow and orange. Both of these photos (made with my Sony a6300 & Zeiss f/4 18~70mm lens) contain a lot of fine micro detail and subtle tonal distinctions, which I wanted for evaluating the relative ability of these printing systems to render fine detail and subtle tonal distinctions.
You may observe that notwithstanding the gamut clipping in the LumeJet softproofs, the photos can be made to have quite similar overall image appearance in respect of vibrancy and dynamic range. Interestingly, the B&W file values are all neutral, but under the influence of softproofing, one sees a somewhat cooler temperature of the LumeJet rendition compared with the Epson SC-P5000/ Red River SG Baryta option. We’ll come back to this observation later.
Figures 11 to 15 provide a graphical portrayal of the gamut comparisons, allowing one to see the extent of the adjustment challenge trying to conform the image appearance between renditions of differing gamut. The positioning of the 3D gamut volume in each illustration is optimized to reveal the maximum number of out-of-gamut (OOG) colours (the little dots). The technique is to use the ColorThink Pro Color Worksheet to extract all the unique colours in the photos and map them onto the paper gamuts. As in Figures 6 and 7, these are mappings of unique colours of the photos against the backdrop of the profile gamut in ColorThink Pro. Where the colours over-spill the backdrop, it means that those colours are OOG for that printer/paper combination.
In both cases, we see that the OOG challenge is much greater for the LumeJet option. While I am not testing or using the paper, I did prepare one such diagram (Figure 15) to illustrate the problem one would encounter trying to print wide gamut photos on their Velvet Matte. Here you can see that the OOG overspill is daunting.
I sent the files prepared to LumeJet’s specs (400 PPI, TIFF format, output sharpened to Contone 400) and waited for the results to be mailed back to me, so I could directly compare what’s on paper between the Epson SC-P5000 output and the LumeJet output. I didn’t know what to expect. On the one hand, I sent them files prepared to look as similar as possible to my Epson prints (see Figures 8 ~ 10 above), but on the other hand, I knew the gamut of their materials was much less than what I have available from the Epson printer.
When I opened the envelopes with the prints and compared them with the output from the Epson printer I was impressed. First of all, the prints handle very nicely because they are perfectly fused to a substrate of about similar thickness, which makes them really robust. The image appearance, especially for the colour photos, between the two was so close that you wouldn’t be able to tell them apart were I to scan them into JPGs and post them here. Tones, tonal gradations, colours, image detail, all very, very close. This means that my colour management is good, their colour management is good, the image preparation customized to each printing technology was good and as a result, not much was lost appearance-wise from the relatively smaller gamut of the contone process and papers.
The prints they returned to me included their report on colour accuracy using the BasicColor software mentioned above. The report for each photo has three components: the step wedge (which includes 31 colours, 24 being the GMCC plus 7 additional grayscale), a summary dE(00) analysis report, and then a detailed dE(00) calculation for each patch in the wedge.
Typical of the results in the Summary Analysis:
Average dE(00): 0.65
Peak dE(00): 1.14 (for 31% gray)
Paper White dE(00): 0.61
Average dE(00) for the Colours: 0.56
Average dE(00) for the Grays: 0.76
These are excellent results by any standard, partly explaining why the prints look so good. Be aware, however, that these results don’t necessarily apply to the rendering intent of the print itself. The “measuring stick” is the step wedge using absolute colorimetric rendering intent, which is appended to the print. The step wedge is used as a print process control strip to validate the accuracy of each printer.
Turning to the Black and White results, the first point of comparison is the aforementioned difference in temperature between the Epson/RRSG Baryta result and the LumeJet, which I illustrated and remarked on above. Theirs’ looks bluer than mine. The reason is pretty straightforward: looked at side by side the LumeJet paper itself looks bluer than the Red River, or in the obverse, the Red River looks more yellow than the LumeJet. To properly assess these perceptions I hauled out the i1Pro 2 and made three measurements on each of the B&W prints: Paper White, Blackest spot in the print and about Middle Gray. I also looked at evidence of OBA content from the paper white measurement (Figures 16 and 17).
The LumeJet paper shows evidence of relatively stronger fluorescence from OBAs.
In the above data, M conditions 0 and 1 include UV reading, while condition 2 cuts UV. Focusing on the conditions 0 and 1, it is readily apparent why the B&W print on the LumeJet paper looks bluer than it does on the Red River paper: it is the influence of the UV as seen in the more pronounced negative b* values especially for Paper White and Middle Gray.
The influence of the lower dynamic range is also evident in the data, where the Maximum Black from the same spot on each print is 10.3 for LumeJet versus 3.3 for the Red River and the Paper White is 92 for LumeJet versus 98.2 for the Red River. Therefore it’s not surprising that a Black and White print would show more “pop” in the latter, regardless of one’s skill in image editing to compensate for lower dynamic range. It would also explain why very deep quartertone shadow detail is somewhat better preserved in the Epson print.
Figures 19 and 20 provide further comparative insight into grayscale rendition for the Epson and LumeJet options, using my 21 Neutrals test which looks at dE(76) readings for the grayscale from L*0 to L*100 in 5 L* increments per step. To remind, 0 values are optimal and the further the departure from 0 the less optimal, so in the best of all worlds, which never happens, both the luminance and chroma curves would hug the very bottom of the graph. For this test the relevant range for the LumeJet paper is 15~90, and that for the Epson/Red River paper is 5~95.
Over their relevant ranges, the average dE is 0.51 for the Epson option and 1.14 for the LumeJet option. Variances from step to step are contained within a fairly narrow range in both cases. In comparing the graphs, it is necessary to be mindful of the scale difference between them (left side).
As to be expected from this data, the gradation smoothness of the tone ramp in the Atkinson printer test page is very satisfactory for both options. I won’t both scanning them and reproducing them here – please take my word for it, having looked at hundreds of these.
The final subject I’ll cover here is sharpness. I was keen to see first-hand what difference it makes to perceived image sharpness whether the print is a half-tone from an Epson SC-P5000 versus a contone from LumeJet, both on luster paper. The short answer is “none”. If one is starting with sharp photographs to begin with, looked at with the naked eye (and even under a 7X aspherical magnifier) one sees the same image detail in both renditions. This means that LumeJet has succeeded in one of its main technical objectives, which was to overcome what they consider to be the natural softness of the contone printing process they started out with. Their technical achievement has been to match the quality of detail obtainable from a top of the line inkjet printer such as the Epson SC-P5000.
Summing up, I think this is a fine print service. They have put a huge effort into achieving a high degree of fidelity in image reproduction and it shows in the quality of the prints, the accuracy of colour rendition and the quality of image detail. I find the Epson has a slight edge for Black and White photos, but for a substantial range of colour photos, one would be hard put to tell the difference between high-end inkjet and LumeJet contone, notwithstanding the narrower gamut of the contone process.