The Epson SC-P5000 Printer
What the Printer Is:
It’s a replacement for the Epson SP4900, incorporating the new Ultrachrome HDX ink set and some design improvements described in the review. It is a production printer designed for use without many days of interruption.
What the Printer is Not:
The basic design and much of the printer’s operational features are the same or nearly the same as those of its predecessor. Epson did not intend this printer to be a fundamental re-design of the 4900.
Should I Prefer an SC- P800 or an SC-P5000 (Key Distinguishing Factors)?
Both are 17-inch professional printers with roll holding and poster-board / stiff media handling capability. You would prefer the SC-P5000 if one or more of the following are important enough to you:
- The customer for an Epson SC-P5000 printer prints regularly; the printer is designed for a production environment, and not ideal that it sits unused for weeks at a time; for those who print infrequently, the SC-P800 would be preferable;
- Automated, built-in roll-feed with integrated rotary paper cutter;
- Ability to add a Spectroproofer (optional and not reviewed here);
- There are several versions of the SC-P5000 printer. The Commercial Edition replaces Light Light Black ink with Violet ink for prepress and contract proofing applications. This allows it to print 99% of PANTONE® PLUS solid coated colors, but sacrifices both the advantages of the LLK ink and Advanced Black-and-White Photo Mode; if you are not doing prepress and contract proofing, prefer the Standard edition with LLK ink;
- Roughly 20% more gamut volume on PK papers, useful for rather limited areas of the color spectrum that some photos may occupy;
- The substantially lower cost per ml of ink due to the much larger ink cartridges;
- Industrial strength production capability and higher speed throughput;
- Epson will be providing a tracking mechanism of ink costs per job from the LCD panel as now in the SP4900.
Why Upgrade from an SP4900 to an SC-P5000 (Key Factors)?
- If the 4900 is encountering performance issues that would require many hundreds of dollars of repair work;
- If the additional tonal range from a deepening of minimum Black capability (e.g., from roughly L*4 to L*2 on PK paper, or below L*2 for ABW mode) is important;
- If the stated improvement of Yellow ink longevity/stability is important;
- If the more efficient handling of printhead maintenance is important.
What I Liked Least About this Printer:
- PK and MK inks still share a single channel in the printhead, making it necessary to evacuate small amounts of the one ink or the other when switching between PK and MK papers.
- Common to all printers: some “driver education” is necessary, (but largely avoidable with the new version of the Epson Print Layout application).
What I Liked Most About this Printer:
- Added dust protection and improved algorithms for keeping the head clean;
- The very high degree of printing accuracy achievable with good profiles;
- The respectable speed of the highest quality throughput;
- The superb quality of Black and, in general, B&W tones, especially on the new Legacy papers;
- The very wide gamut, even though not that many photos may require it;
- Overall superb print quality whether color or B&W (fine detail rendition, smooth tonal gradations, excellent tone and hue reproduction).
The Epson SC-P5000 Printer
Back in January, we advised that Epson had announced this new printer and that we would be reviewing it in depth once Epson made a unit available to us for the purpose. It is a replacement for the 4900 (about the same size/weight) with the new HDX ink set and printhead as in the SC-P7000/9000 models.
“The Beast” (yes, it’s built like a tank) was delivered the day before April Fools Day. My son and I carefully moved it into place and I installed it (Figure 1- seamless process if you follow the instructions exactly) and started playing. One of the very first things I noticed is that this printer is quieter than was the 4900 when idling. It makes a tolerable hum rather than a grating screech. I’m pleased that Epson addressed this matter, though even quieter would be nicer – of course.
It should sit on a sturdy table or cabinet. I had, from my 4900, the sturdiest work table IKEA produced at the time, which I further reinforced with under-fitting cabinets and a wood bar, all of which press very firmly against the legs (Figure 2). There are a number of furniture options – key thing is stability and minimal transmission of vibration.
Much of this printer is/looks/feels like a 4900, so one of the key questions 4900 owners may wonder about is whether they should upgrade. The bottom-line on this is stated in Executive Summary above. The value of those differences to you is a personal matter. For those pondering a choice between this printer, a P800 or a Canon Pro-1000, I’ll provide some comparison, and as well, see above. The SC-P5000 is a larger, heavier, more expensive printer than the other two, though all three are 17-inch models and all three produce fine prints.
Like the 4900, the P5000’s ten inks at 200ml capacity each, influence its dimensions. This ink set contributes to the printer’s very wide gamut (as for the 4900), while the high-capacity ink cartridges (compared to the P800) lower the P5000’s per ml ink cost. Like the 4900, Epson calls this a “production” printer and it is constructed accordingly, including its automated roll-holder and integrated rotary cutting mechanism.
Several points on set-up:
- The connection is via USB or Ethernet. There is no wireless (and thank goodness, no AirPrint driver).
- During the time I was preparing this review, the printer needed a firmware update. As issues are discovered, Epson updates the firmware. Therefore I recommend when setting up the printer to use the latest available firmware version from the Epson website and to check for subsequent firmware updates monthly. The firmware version was originally N002GC on the Epson website. The latest firmware version is NO24H3.EFU (OSX), which should be downloaded and installed with the Epson Firmware Updater.app included in the firmware package.
- Likewise, during the time I was preparing the review, a couple of driver updates were required. The current Printer Driver and Drivers and Utilities Combo Package have been updated on the Epson website. The current combo package for OSX is Epson17979.dmg and the Printer Driver 9.91 is Epson18518.dmg. It is likely that the driver will be updated again, and we also recommend checking for updates on a monthly basis.
- There is new “Epson Print Layout” software, version 1.3.4 (18643.dmg), which I’ll describe at the end of the review.
- Print Head Alignment can be an automated process (no need to compare patches on a printed sheet, unless using the optional manual alignment method), necessary if you see banding, grainy or blurry prints. As I did not see any such problems I determined that the alignment out of the box is satisfactory. If you do it, it takes about 10 minutes according to the manual. Epson recommends using the LCD panel for this process.
- When printing the Test Page, one needs to make sure that the choices of Sheet versus Roll and “Manual” versus “Cassette” are consistent between the LCD panel and the driver menu on the computer, otherwise, you get an error message. If all goes well, you get a printed sheet offering you “Congratulations”.
After a bit of discussion below of the comparative feature sets and some differences between the 5000 and the 4900, I’ll go to the heart of the matter, which asks how well does it print. Papers and profiles are important to print quality and are factors over which we have some control; hence, I shall dwell considerably on these aspects of performance.
Notable Features, Differences, and Operational Tips
Choice of Printer Version and Corresponding Ink Set
According to Epson, the new SureColor P5000 UltraChrome HDX 10-color ink set includes newly developed pigments, new Orange and Green inks, as well as a Black ink that is 1.5 times denser than the previous generation, delivering a wider contrast ratio and improved resin encapsulation for better gloss uniformity and clearer, sharper images. There is a new Violet ink for the Commercial Edition.
A very important difference between the 4900 and a major feature of the 5000 that comes right up-front before you even buy it is the choice between the Standard/Designer Editions and the Commercial Edition. The difference between them is one ink: the Standard/Designer Editions use LLK (Light, Light Black), best for photography, while the Commercial Edition replaces LLK with V (Violet) ink, best for prepress and graphic design. You need to decide up-front which Edition to buy, for which Epson provides some guidance, explained just below.
The difference between the Standard and the Designer Edition is that the latter comes with the EFI Fiery RIP software supporting CMYK, Spot, RGB, and PostScript. This review focuses on the Standard Edition, which is the one Epson provided for this review and would be of greatest interest to most prosumer and many professional photographers who don’t need the additional Designer Edition options.
When specified with the Epson Ultrachrome Violet HDX ink set (Commercial Edition) the Epson SC-P5000 is a pre-press proofing printer for reproducing packaging, marketing materials and colors for other commercial purposes with, according to Epson, 99% Pantone Solid Coated Colour accuracy when used in conjunction with Epson Proofing Paper White Semi-matte and Epson ink.
When specified with the Epson LLK HDX ink set, the printer is purposed to photographers and artists with, according to Epson, exact reproduction of 98% of Pantone Solid Coated Colours when used in conjunction with Epson Proofing Paper White Semimatte and Epson ink.
The choice between the LLK option and V option needs to be made up-front and it is final. For photographers using the printer primarily to print “fine-art” and other ordinary photographs, Epson’s strong recommendation is to use the LLK option, because LLK contributes much more to desirable photographic image properties than would Violet. No-one would notice the difference of Pantone-based accuracy between an LLK print and a V print; but depending on the photo and the observer, one could notice the difference with versus without LLK to the quality of tonal rendition, especially for B&W work. As well, with Violet, there is no Advanced Black and White (ABW) mode because ABW needs LLK ink. Bottom-line advice from Epson: buy the LLK option unless you are printing for graphic design or prepress applications.
The set-up procedure in the LCD screen (which is nice and crisp) will nag you about three times over whether you are sure you want to install with the selected ink (LLK versus V), warning that once the ink install starts it’s irreversible. You’d have to be really daft to install the wrong ink by mistake. ☺
Initial Ink Supply and Ink Consumption
The printer comes supplied with “starter cartridges” (80 ml each). Once the printer is charged, the Supply Levels utility showed as follows (Figure 3), I believe scaled to the volume in these cartridges. People who want to complain about being supplied with 80 ml starter cartridges rather than full 200 ml cartridges are of course free to do so, but should consider that the Epson SC-P800 and Canon Pro-1000 printers have 80 ml cartridges as their maximum capacity ever. Over the life of these printers, due to volume economies the cost per ml of ink will be much lower for the 200 ml cartridges of the SC-P5000.
On the Printer Status sheet I printed immediately after set-up, it showed Maintenance Box 1 with three stars out of ten consumed. As I don’t know the capacity of the box, this is useless information, except to know 7/10ths remain, whatever that means. Most of this fill is shipping fluid packed into the lines, damper and printhead before shipping; it needs to be evacuated on set-up. A small amount of ink is also evacuated along with the shipping fluid, but we don’t know exactly how much. In any event, most of the ink sucked out of the cartridges is used to charge the lines, the damper and the printhead, so it gets used for printing, not wasted.
While on the subject of ink usage, I could find nowhere in the driver software or on the LCD panel where ink-use by job is reported. The 4900 had this information for the most recent ten prints and the Canon Pro series has it in gobs of detail (stored in the printer and accessible for download using the Canon Accounting Manager software). Epson advises that LCD access to ink usage data for the last ten prints will be included in a future firmware update, estimated timing of which could be during this summer.
PK and MK Ink Switching:
You will be relieved to know that the much beloved time and ink wasting process of switching between MK and PK channels when changing between MK and PK papers lives on in the SC-P5000, but at least it can be automated (optional choice in Printer Set-up>Auto Black Ink Change). Black Ink Switching is an Epson institution by now, and as those of us of a certain age would appreciate, old habits don’t die easily even if they are annoying ones – to some.
The two main reasons why this approach remains in the SC-P5000 are: (1) it was not Epson’s intent to do a major re-design for this model, which adding a channel to the damper and print-head would have entailed, and (2) most of the market for this printer is commercial (recall, it’s considered a “production” printer), meaning the firms using them would have more than one and can dedicate different printers to either PK or MK workflows, so their need for switching is minimal. I know, irrelevant considerations to photographers using this printer for one-off jobs on a variety of papers and who just want to get on with it at minimum waste of time and materials.
Ink and clogging:
Epson determined that dust is an important cause of clogging, so they have added new features for protecting the inside of the printer from dust by redesigning part of the roll holder cover and the front paper tray. Epson says that the ink-repellent printhead coating will also enable reduced printhead nozzle maintenance.
One nice new feature in the firmware is a “Paperless Nozzle Check” that one commissions from the Maintenance tab in the LCD panel. It uses a sensor instead of paper. Unless the printer itself does an Auto Nozzle Check, I have been using it at the start of my printing sessions and it seems to work fine. If it detects no clogs there are no clogs; if it says cleaning is needed, it shows on the LCD the channel that needs to be cleaned and allows the user to trigger the cleaning. I have verified these cleaning notices by doing several paper-based nozzle checks just after they appear, and sure enough, the channel that the LCD said needs cleaning does need cleaning. So paperless is trustworthy. Use it.
If a head cleaning were needed – the Maintenance menu allows for one more option than was available in the 4900: one may select either “Normal” or “Heavy”; then if those fail, the option for a “Powerful” clean still exists, but has been moved from the Maintenance tab where it sat in the 4900 to the “Administrator” tab in the 5000. This was probably designed to discourage non-essential use. “Normal” does not engage the wiper blade and is very gentle on the printhead. The automatic cleanings discussed below are “Normal” ones. Epson strongly recommends not proceeding to “Heavy” unless there have been several unsuccessful attempts with “Normal”, and not to use “Powerful” unless “Heavy” has failed. This is in the interest of preserving the printhead.
Other Cleaning Cycle Options:
There is also a policy choice of whether or not to enable automatic nozzle-check/cleaning cycles, and if enabled, their interval. The options are “Periodically”, “Every Job” or “Off”. The default the printer ships with is ”Periodically”, and I would advise leaving it active. Over the month + that I’ve been using the printer (in 22%~37% humidity), making about six to ten letter-size prints per day it periodically detected the need for cleaning, did the cleaning cycle(s) and then allowed the print jobs to proceed – very handy and useful.
Allied with this option is another option selected from the LCD panel (Printer Set-Up>Auto Nozzle Check>Auto Cleaning Times) which allows the user to set the number of automatic print head cleanings the printer performs when clogged nozzles are detected. The range is from 1 to 3. I strongly recommend setting it to 3.
There is a reason for this recommendation. Auto nozzle checks often happen after paper is loaded and the Print button is pressed, so all is engaged to print, then the Auto Nozzle check is performed and it may detect that cleaning is needed. In these situations I’ve found that normally one cycle suffices. However sometimes it doesn’t suffice and it detects that at least another cycle or two are needed. If the setting is put to 1, and more cleaning is needed, the printer tells us this after it checks the state of the nozzles from the first cleaning, but then it will proceed to make the print regardless that it “knows” there is still a clog. Hence the user needs to cancel the print job, which if the paper is at all curly can eject the paper back into the printer rather than out the front, causing one to fish it out, do a manual cleaning or two and re-commission the print job.
Setting the “Cleaning Times” to 3 will give the printer three attempts to clear the nozzles automatically without user intervention before it will proceed to print. In my experience to date, this has been more than enough for the nozzles to be properly cleaned and avoids all the above-mentioned trouble.
My recommendation to Epson for improving this procedure is to avoid the risk of this trouble sequence by configuring the Auto Nozzle Check to occur only after paper is loaded but BEFORE the print is commissioned. Once the print command is given, nozzle checking should not intervene.
Anyhow, bottom line on all this: periodically the print head needs cleaning (in my experience to date, infrequently), but how often in general would depend on usage pattern, the kind of paper used and atmospheric conditions); there are automated ways for both detecting clogs and cleaning them, and setting up the cleaning regime as recommended above will keep the process quite seamless. I could not find out how much ink a cleaning cycle uses – I’m told it depends on a number of factors, such that it can’t be quoted as a fixed amount.
Paper Curl and Head Strikes:
I discovered the hard way (loading sheets of Legacy Etching paper straight out of the Epson sample pack) that it’s important to avoid loading heavy, even very moderately curled cut sheets into the top sheet loader of this printer without taking adequate precautions. I suggest manually setting the platen gap to Wide or Wider for thick heavy papers that have a bit of a curl. Also, gently bend the paper in the opposite direction of the curl to flatten it as much as possible just before loading it. Otherwise, there could be head strikes. Head strikes can damage the print head (didn’t happen to me), or they could be an important cause of nozzle clogs (did happen to me, and took a few cycles to completely clean them). Once I implemented these two precautions, there were no more head strikes and no nozzle cleaning on this account.
Where to Make Settings (LCD versus Computer):
A number of paper and paper handling settings are available on both the driver menus (computer) and the printer LCD. Users may be uncertain about what is best to set where, and which settings in the one place take precedence over the same settings in the other place. The general rule is that software (driver menus on the computer display) overrides hardware (the printer LCD) except for LCD items that the user adjusts off their default positions. Once the user moves an LCD setting from the default using the LCD controls, that setting overrides the computer-based driver setting for the same item.
If on the printer LCD for paper type you change the setting to “No Paper Selected”, ALL paper-related settings (e.g. media type, thickness, platen gap, etc.) need to be set in the driver menus on the computer. This is good, because for any given paper it allows the user to create and save a custom preset with complete certainty that the settings in this preset will prevail from one print job to the next using the same paper.
Printing Canvas Rolls:
Epson provided to me a roll of Epson Exhibition Canvas Satin for reviewing the performance of the roll holding unit. Epson provides very clear and complete instructions for setting it up and using it in the Quick Reference Guide, so I shall not repeat any of that here, as this is not the purpose of a product review. I do however have a few bits of advice regarding its usage.
Firstly, people sometimes find it challenging to set up roll printing so that they achieve the image placement they want on the roll. There is an approach for securing reliable results, which is to create a Custom page set-up in the driver. For example, in Lightroom, in the Print Module>Page Set-Up, select “Manage Custom Sizes” and you will see a screen looking like that in Figure 4.
In this example, I made a custom page set-up for a print fitting on a segment of the roll whose total page size would be 13 inches wide (the full width of the roll I was provided) and 25 inches high (or “long” – a mini-pano). So I set these dimensions in the Paper Size panes. For the “Non-Printable Area”, select “User Defined”, and enter zeros for all four dimensions. You can save these dimensions as a custom preset if you think you will be using them again. This set-up then allows you to revert to the main right-side menu of the Print Module (if printing from Adobe Lightroom), where you can adjust the Margins and the Cell Size to suit the image and your margin requirements, provided all fits within the Paper Size for that preset. The print will emerge according to your instructions.
Secondly, it is worthwhile following the guidance provided with the canvas roll about whether or not to allow the printer’s cutter to cut the canvas after it prints. For example, with Epson Exhibition Canvas Satin, the recommendation is to cut it manually. The reason is that the canvas is very heavy, it could prematurely dull the blade if done often enough, and while the blade is user-replaceable, Epson advised me that the part is expensive. To facilitate manual cutting there is an option in the driver set-up to print a faint cut line near the trailing edge of the print. I recommend using this option and very sharp scissors.
Thirdly, there is an option on the LCD to select whether or not to print “Remaining Roll Information” on the canvas after the print is finished. I thought this was a neat idea, so I turned it on only to discover that it consumes (wastes) roughly a foot of paper. The reason for this is that the feature is most useful for people using the optional Epson Spectroproofer, as that instrument needs 10.5 inches at the trailing end of the print. So for those not using the Spectroproofer, the recommendation is to leave this option OFF unless you don’t mind wasting material.
With these pointers in mind, and following the instructions in the Quick Reference Guide for managing the mechanism, printing on rolls is quite automated and easy. All worked as advertised, as I found from the three prints I made with it for this review (one of which is Figure 5).
Talking of Image resolution in Pixels per Inch (PPI), Epson America advises that the native resolution of the half-tone driver is 360 PPI, which is the maximum needed for normal photographic purposes. Epson’s “Precision Dot” screening creates the continuous tone image appearance. Selecting “Finest Detail” in the driver changes the driver native resolution to 720PPI, intended for graphic arts, fine type, and fine lines. It also quadruples the amount of data being sent to the printer, and for normal photographs, Epson America advises it is unnecessary. Epson America recommends sending to the printer a minimum resolution of 180 PPI, and especially with low resolution to not over-sharpen the photo (keep Radius at or below 1 pixel to avoid visible halos). When the user sets resolution in the photo editing application to levels other than 360 or 720 PPI, the print pipeline will resample the data so that the printer works with the native resolution.
Image PPI discussed just above is a completely separate kind of resolution setting than the choice of, say, 1440 dpi or 2880 dpi in the printer driver. These latter settings are for printer dots, not image pixels. The notes in the manual regarding the quality to expect from each of these dpi settings provide guidance on which to use. 2880 is for the highest photographic quality the printer can deliver (but it is slower); therefore I normally opt for quality over speed and print at that setting. There is little difference of ink consumption between 1440 and 2880 because on the whole 2880 uses smaller dots than does 1440. Think of it as taking the same amount of ink and dividing it into bigger or smaller drops.
Color Gamut of the New Ink Set:
One of the more important characteristics of a printer is the color gamut it can reproduce on papers designed for wide-gamut printing. The addition of the revised Orange and Green inks, along with the improvements of Yellow and Black, give the SC-P5000 a gamut advantage over models such as the SC-P800 and the Canon Pro-1000 which do not have these additional colors. But this was also true of the 4900. So the question is: how do the gamuts of the 5000 and 4900 compare.
Before showing that comparison, I would like to remind readers of observations I made in my P800 and Canon Pro 1000 reviews: there are only limited parts of the color spectrum that need the full gamut of the Epson 4900/5000 printers for some kinds of photos, especially those with very brilliant green-yellow hues. Nonetheless, for the benefit of those of us who may be gamut-obsessed, let’s see what changed between the two models (Figures 6 and 7). I’m demonstrating these gamut volumes with Epson’s profiles for its Legacy Baryta paper, this being a very wide gamut paper. I assume Epson has optimized gamut in its profiles for both printers.
The gamut volume of the 4900 is about 1.6% (one-point-six percent) larger than that of the SC-P5000; an insignificant difference. We may also look at the difference of gamut shape. The wireframe (black grid with tiny squares) is the SC-P5000. It has slightly more gamut at the brighter end of the Luminance scale and slightly less at the darker end (Figures 6 and 7). These differences are too small to be practically significant.
So the take home on gamut is that the 4900/5000 printers are very comparable.
Turning to gamut comparison of the SC-P5000 versus the P800 (Figures 8,9, 10 and 11), the SC-P5000 gamut is on the whole considerably larger than that of the P800, the P800 gamut poking through that of the SC-P5000 only in a range of light blue and magenta. By the way, the Canon Pro-1000 shows slightly larger gamut in the Blue-Magenta quadrant of the two-dimensional portrayal in Figure 11, but on the whole, the Epson SC-P5000 is capable of wider overall gamut than either the P800 or the Pro-1000.
These differences of gamut volume have no bearing on the printing accuracy tests I discuss below because the measurements are limited to the colors in the Color Checker target, which do not challenge the gamut boundaries of any of these printers. Nor do the gamut volumes have any particular implications for the deepest Black (i.e. minimum Black in L*) that either the P800 or SC-P5000 can reproduce, as we’ll see further below.
Now, just to bring the gamut discussion down to practical photographic reality, I’ve mapped a couple of “real world photographs” to the gamuts of Ilford Gold Fibre Silk (very similar gamut to Epson Legacy Baryta) in both the SC-P5000 and P800 printers.
Wagner Bird Photo (Figure 12):
This is a very wide gamut photo with brilliant green/yellow hues that challenge printer/paper gamut boundaries.
The technique of illustration here is to decompose the photo into its individual pixel/colors and map them against the printer/paper profile of interest. (It would be convenient to blame the time it took to produce this review on the months required to select each of these 230 thousand + pixels and individually place them where they belong in or around the gamut map, but that would be facetious- a ColorThink Pro Color Worksheet does this in less than a minute; fabulous application.) Having done that, select the angle of view of the 3D diagram in ColorThink Pro’s 3D mapper that best shows the out of gamut (OOG) colors (the squares that fall outside the shaded area – Figures 13 and 14).
Comparing these two Figures, you can see very clearly how more of the greenish-yellow colors are OOG for the P800 compared with the SC-P5000, but in both cases, the vast majority of the colors are in gamut for both printers. Side by side prints of this photo I prepared for the Canon Pro 1000 review indicates a slight visual advantage for the 4900/5000 type printers in this limited area of gamut, but these differences are subtle and hard to portray over the Internet.
There isn’t much Blue or highly saturated Red in this photo, so I did another emphasizing natural blue sky and some deep red. Blue sky didn’t challenge either printer, but the Red did (Figures 15, 16, 17, 18, 19.)
The takeaway for the Waterloo set is that there is just a bit more saturation clipping of the Red hues in the P800 relative to that in the P5000, while the Blues are no challenge for either printer. Red is a small share of this photo, so it would look very much the same coming from either printer.
The overall bottom line on gamut is this: not much difference between the SP4900 and the SC-P5000. These are the widest gamut desktop inkjet printers on the market. In the fine-art photography market, how often one would need the extra gamut compared with a P800 is a big open question, depending very much on the kind of photographs one prints and how sensitive one is to the one or two areas of more limited gamut that could impact certain kinds of photos printed in the other printers. There are photos that could benefit from the widest possible gamut a printer can offer, many others not.
The inset just below provides another graphic comparison of gamut volume between the SC-P5000 and SC-P800 for a completely different kind of photograph.
Gamut Comparison – Another Real Photograph
Figure INS-1 is a photo I made with my Sony a6300 in London’s Leake Street Tunnel (underneath Waterloo Station) last November at ISO 8000, f/4.5, 1/20th second handheld.
I selected it for this demonstration because it has particularly challenging colors in the yellowish-green, red and blue parts of the color spectrum. I wanted to see how it performs between printers and between papers.
To do this, as above, I converted the file to a Color List in ColorThink Pro, yielding some 174,500 total colors, of which some 42,000 unique colors. The list is then graphed yielding 174,500 square dots distributed to where they belong in the L*a*b* color space. This is overlaid with a semi-opaque view of the relevant printer/paper profile’s 3D gamut volume, and the composite tilted to whatever position best shows the dots that spill out of the profile’s gamut boundaries.
The test printers are the Epson SC-P5000 and SC-P800, the test papers Epson Legacy Baryta and Legacy Etching (a PK versus an MK paper). The two papers are compared for the SC-P5000, and the two printers are compared for Legacy Baryta. I looked in particular at the yellowish-green, red and blue portions of the color spectrum as these are the most exposed areas for gamut issues.
Starting with greenish-yellow (Figures INS-2/3/4), the SC-P5000 is able to contain most of the colors in this range, the SC-P800 a bit less so, and the Legacy Fibre paper much less so (red arrows point to Out of Gamut (OOG) colors):
A similar story applies for the red-blue parts of the spectrum:
The takeaway from the study of these gamut overspill illustrations is three-fold:
(1) The SC-P5000 with a good quality PK paper has the best chance at reproducing most if not all of the gamut one is likely to require from highly saturated photographs.
(2) The SC-P800’s color rendition is not really too far behind, despite the large statistical impression one may have of its gamut differences relative to the SC-P5000, and
(3) For some photos with highly saturated color, paper matters much more than the printer model, insofar as a high-quality matte paper, even with the wide gamut SC-P5000, sacrifices a considerable amount of useful gamut relative to that achievable with high-quality PK papers.
This section is about determining how accurately the printer/paper/profile combination in an ICC color-managed workflow can reproduce on paper the file color values sent to the printer. The core methodology is the same as that I used in my Epson SC-P800, Canon Pro 1000 and Canon Pro 2000 printer reviews. A summary of the procedure is just below. For more detail please see explanations in the earlier articles.
- Start with an image of a synthetic Color Checker with known reference values for each patch (the printer evaluation target).
- Select the paper, Media Type and ICC profile being tested.
- Print the evaluation target with Absolute Rendering Intent, and let the print dry.
- Use MeasureChart in i1Profiler to measure the printed patches with an i1Pro2 spectrophotometer.
- Save the measurements as a text file in CGATS format and import that file into my Excel template created for measurement of the dE between each of the measured printed patch values and their reference values. Take an average of the 24 dE measurements for deriving an Average dE figure.
- The lower the Average dE and the less the dE variance of each color around the average the better.
Given the considerable interest in Black and White printing, in this review, I am also using several enhanced tests for Black and White printing, described further below.
Given that this printer bridges the prosumer, professional and commercial markets, I decided to test it with quite a range of media and profiles. Hence, I am presenting results for ten papers, using both OEM and my custom profiles. Of the ten papers, five are for PK ink (luster) and five for MK ink (various matte finishes).
Before we get into data and detailed comparisons, I feel compelled to emphasize, as I have before, some important factors for interpreting the meaning of these results.
The primary distinction for profiles is that between OEM (original equipment manufacturer) supplied profiles and my custom profiles. Readers will notice from the dE (delta-E) measurements presented below that invariably, my profiles are more accurate than OEM profiles. This is normal and expected.
When I do my own profiling and profile testing, the instrument, the measurement software and the printer used to create the profile are the same as those used for printing and measuring the target used to evaluate profile performance. This is a tight loop assuring consistency of materials and processes occurring between profiling and evaluating the profiles.
Notwithstanding the basic intent and huge efforts made in the field of ICC profiling to assure cross-device consistency of outcomes, variances can and do creep in because of possible or inevitable differences between an OEM supplier and me in respect of:
- measurement devices being used,
- the performance of different units of the same measuring device model,
- software packages being used,
- profiling approach (e.g. profiling software settings) and
- perhaps printer performance from one printer to the next of the same printer model due to factors in the usage environment, though Epson advises that the printers are calibrated to a set standard at the factory, such that unit-to-unit variation is rare.
Hence, we should expect higher dE outcomes when I measure their profiles with my set-up than obtained from my internally consistent set-up wherein I am using the same spectrophotometer, the same software, and the same printer in the same environment for both profiling and profile evaluation.
So for example, when we see OEM profiles with a dE result of say 2.5, while mine for the same paper may be 0.6, it doesn’t mean there’s anything wrong with the OEM profile. It’s just less accurate as seen by my measurements for the above-mentioned reasons. The OEM profile may still yield results that are fully satisfactory to the user on his/her printer. Users who do not have their own profiling equipment in-house should visually evaluate their results using OEM profiles before opting for custom profiling.
As well, there are tolerances of human visual perception in respect of what we think we are seeing when we look at photographic prints in isolation. Depending on the color (we are more sensitive to differences in some colors than to others), a departure of several dE from the file value may be unremarkable, or at worst remain acceptable in the context of ordinary photography where strict accuracy is not only non-essential but also hard to detect. So in this example, while I prefer an accuracy result of Average dE 0.6 over Average dE 2.5, it’s only because 0.6 gives me more assurance of color accuracy (relative to the image file values) than 2.5 does. Compared with Average dE 2.5, Average dE 0.6 also has less variance by color from reference values and from the Average.
For the most part these days OEM profiles are pretty good, and unless you are using or engaging services using high-quality professional equipment for profiling, you may not necessarily do much better. For avoidance of all doubt, this discussion does not recommend that you need custom profiles to get good results from this printer. Getting “down to earth” about it, I made test prints of real photographs with OEM profiles and they look fine, regardless of statistical differences in accuracy.
My particular interest here in custom profiling is to test just how accurately a properly colour-managed printing pipeline featuring the Epson SC-P5000 can replicate the reference image color values. Any outcomes superior to the OEM profiles would point to the possible contribution of high-quality custom profiles for being more reliably assured of optimal results, all else copasetic.
Figures 20 and 21 show the results of the analyses done, using ColorThink Pro i1Profiler and Microsoft Excel.
Columns A to C: Identification of the Media
Column D: Gamut Volume (from ColorThink Pro)
Columns F, G, H: The L*a*b* coordinates of the RGB values forming the gamut boundaries (from ColorThink Pro).
Column I: Average dE accuracy of the printed profile evaluation target(from i1Profiler, i1Pro2, and Excel).
Columns J, K: The L*a*b* coordinates of the White Point in the profile (column J) and measured on the print (column K).
Columns L, M: The L*a*b* coordinates of the Black Point in the profile (column L) and measured on the print (column M).
These tables contain a considerable amount of analytical information. I’ll provide here a summary of the key takeaway observations for each.
- For the OEM and the Custom profiles, gamut volumes for the luster (PK ink) and matte papers (MK ink) fall within the expected range for this class of printer.
- For the OEM and Custom profiles, the luminance range is very wide, with darkest printed Black in the range of L* 2.2 to 2.7 for the luster papers and L* 13.5 to 15.7 for the matte papers; printed white points range between L* 92 and L*98 depending on the paper.
- For the OEM profiles, accuracy relative to the reference image is satisfactory, the range [measured in dE(76)] being from 1.25 (Epson Legacy Platine – exceptionally good for an OEM profile) to dE 2.88 for Hahnemuehle Bamboo. I have no explanation for why accuracy varies between papers, but it does.
- For the Custom profiles, accuracy relative to the reference image is stellar for all the Baryta papers and Epson Legacy Platine (range of dE 0.44 to 057), and very good for Epson PLPP, at 1.00. Accuracy for most of the matte papers also falls well below 1.00, except for Hahnemuehle Bamboo at 1.5.
These accuracy ratings are averages of the dE values for the 24 colors in the evaluation target. It’s instructive to stop for a moment and consider the difference it makes whether the average is the best of the lot (0.44) or the “worst” (2.88).
These are charts of the dE per color for the most and least accurate averaged renditions in the set. When comparing them, be sure to note the difference of the dE scale on the left side, and remember: lower and more uniform is better. For the Hahn Photo Rag Baryta (Custom profile), the average dE is 0.44, the “worst” colors being yellow and orange-yellow at dE 1.29 and 1.07 respectively (both still very good). So the maximum variance from the Average is 0.85 (1.29 – 0.44). Most of the other colors show a dE below 0.5, which for most people would be visually undetectable.
Turning to Figure 23 for Hahnemuhle Bamboo (OEM profile), the average dE is 2.88, the worst colors being yellow, orange-yellow and orange at 4.62, 4.37 and 4.27 respectively. Half the colors on the chart have dE above 3.0. The maximum variance from the average is 1.74 (4.62 – 2.88). Once dE exceeds 3, it is fairly safe to say that an astute observer would notice the color isn’t quite right relative to parallel perception of the (correct) reference value. However, depending on the tolerance allowed by the photograph and the photographer, a print with this profile could still look pretty good.
Before I turn to Black & White printing with the SC-P5000, let’s move off of targets and measurements, and have a look at another real-world photographic result – well, to the extent I can help you visualize it over the Internet.
This is the soft proof (screen-grab) for Ilford Gold Fibre Silk (the paper I had on hand of the requisite size); the print looks very much the same. I made the photo of this piece of mural art that resides on the curvature into the ceiling of the Leake Street Tunnel underneath Waterloo Station with my Sony a6300 (APS-C 24 MP sensor). It’s quite dark in there, so this shot required ISO 12,800, f/4.5 at 1/20th of a second, hand-held with Image Stabilization on. The print is on a 17×22 inch sheet of Gold Fibre Silk, with a one-inch border for the long dimension. The native resolution of the file at this size was 300 PPI, upsampled in Lightroom’s print module to 360 PPI. Sharpening Amount is 50, and noise reduction is set to Luminance 20, all the rest at Lightroom default. Output sharpening is Standard Glossy. This is challenging stuff to both photograph and print to an acceptable standard of quality, so I was pleased with it, as were several people who viewed it and do tell me exactly what they think.
Black and White Printing:
Given strong reader interest in Black and White printing and the special place of this printer straddling several niches of the imaging market, I thought it perhaps useful to focus on Black and White printing in some depth for this review.
The key questions of interest for B&W printing, hence what we want to test the printer and related components for, are:
- How accurately does the printer/paper combination reproduce the tonal values of the image file, especially neutral ones?
- What is the darkest Black that the printer/paper combination can reproduce?
- How well does the deep shadow detail of the image file reproduce on paper?
- What difference does Rendering Intent make, especially for Matte prints?
- How does Epson’s Advanced Black and White mode perform relative to a standard application-managed ICC workflow, again especially for matte prints?
- How well does highlight detail visible on the monitor reproduce on paper?
To address these questions except for (f), I developed three tools for examining grayscale tonal rendition that are more granular than I had used previously. I complemented these data based tests with visual observation of the Roman 16 low-key B&W image. The answer to (f) partly depends on the 21-neutral test, but more importantly on visual observation of the Roman 16 high-key B&W image.
The three tests are explained in my article “Expanded Neutrals Testing”, published May 24, 2017, on this website. To remind, the first, “21 Neutral”, measures the accuracy of tonal rendition and neutrality of the grayscale in 21 L* levels from L*0 to L*100 (Black to White), each level being L*5 brighter than the previous one. The second is another 21 levels, but ranging from L*0 to L*20 in increments of one L* level, primarily a visual test looking for the darkest L* Black that one can see on paper as being separate from a Black surround in the print of the target; however, it is also accompanied by L*a*b* data. The third, “30 Neutrals”, is a graphical representation evaluating how accurately a printer/paper/Rendering Intent combination can differentiate shadow tones from L*1 to L*30. In this article, I focus on the first and third tests (results in Figures 25 and 26), plus corroboration with real photographs from the Roman 16 collection.
I have not applied the fullest possible extent of the “30 Neutrals” test to all nineteen profiles in this review because it would have required printing and measuring a very large number of targets; instead, I used a sample of papers for illustrating the capabilities of the printer in respect of B&W lower quartertone rendition and some interesting distinctions between profiles, Rendering Intent and workflows.
21 Neutrals Test (Full grayscale), Key Observations:
This test addresses questions (a), (b) and (e) above.
- For the OEM profiles set, average dE (including Luminance and Chroma) ranges between 1.0 and 3.0. These are very decent overall results for OEM profiles. Results for some of the Epson Legacy Baryta and Legacy Platine tests are around 1.00, which is excellent.
- For the Custom profile set, average dE (including Luminance and Chroma) ranges between 0.62 and 3.86, depending on the paper and the Rendering Intent. The lowest (best) dE results are achieved with the PK papers.
- For both the OEM and Custom results, the data for the matte papers includes deep shadow Luminance (L*) that the papers cannot reproduce. Column D shows L* adjusted data, which includes L* levels 5 to 95 for the PK papers and L* levels 15 to 95 for the MK papers. By excluding the absolutely unattainable values, the remaining L* dE numbers, on the whole, show that the printed outcomes are quite accurate. Depending on the matte paper it may be appropriate to start the series at Level 20, as shown (Figure 27) for Epson Legacy Fibre paper.
- No matter whether the profiles are OEM or custom, the neutrality of B&W printing is excellent (most of the samples dE well below 1.0), a bit better for the custom than the OEM profiles, but these are small differences of already very low departures from 0 (absolute neutrality) for both the a* and b* channels.
- The Deepest Black is very good for all the papers, generally falling within the range of L*2 to L*3 for the PK papers (and a bit deeper Black yet for the ABW mode) whether for OEM or custom profiles. This is more black than was achievable with the 4900 printer, where a minimum value of L*4 was more usual. Minimum Black for the matte papers falls with the range of L*13.5 to 15.5, which is also darker than previously exhibited for this class of papers. With previous ink sets, values of L*16 and above were more the norm.
Based on these results, with a properly colour-managed workflow, this printer consistently delivers quite accurate and neutral B&W results across the relevant tonal ranges of both PK and MK papers. I shall discuss ABW further below.
30 Neutrals Test: Key Observations (Question (c) above, Deep Shadows)
I implemented the 30 Neutrals test for the following papers, OEM and Custom profiles:
- Epson Legacy Baryta, Absolute R.I.
- Epson Legacy Platine, 3 Rendering Intents
- Epson Legacy Etching, 3 Rendering Intents
- Epson Legacy Fibre, Absolute R.I.
- Ilford Gold Fibre Silk, (Custom profile only; Absolute and Relative Colorimetric R.I.)
The numbers presented here (Figures 25, 26) provide general guidance, but are not necessarily the last word when assessing the fidelity of shadow detail. It is necessary to print photographs designed for this purpose and examine them, as I did using the Roman-16 images mentioned above. First, what do the graphs show us?
Particularly easy to appreciate on the PK papers, the linearity of L* printer response whether with OEM or Custom profiles is remarkably good (for example Figures 28, 29, 30; black line = linear reference; red line = measured results). The data written on the graph is the average of the Absolute departure of measured values from reference values for levels 1 to 30 in each of the three channels (L*, a*, b*). ABSCOL means Absolute Rendering Intent and RELCOL means Relative Colorimetric Rendering Intent.
For all three of the above Figures, the little triangle between the Red and Black curves at the lower left corner of the graph indicates that at L*1 in the image file, the printer/paper can produce at best L*2+; however, by L*3 they converge, which is very good performance.
For good reason (which we’ll see in the discussion of the matte papers), photographers would print with RELCOL + BPC or PERCEPTUAL Intent, not in ABSCOL as portrayed in the three figures immediately above (as I’ve mentioned before, ABSCOL is used primarily for proofing; if the proofing indicates high accuracy, the “Intent” of the Rendering Intent you use will reproduce pretty much as intended). Let’s focus on RELCOL + BPC for this demonstration of how RELCOL values diverge from ABSCOL rendering – as they should, using Epson Legacy Platine and my custom profile (Figures 35, 36), though the OEM profile would serve as well.
The ABSCOL red plot (Figure 31) shows that the L* channel of the print pipeline exhibits eerily accurate behavior, but the RELCOL+BPC print (Figure 32) will look better (if shadow detail matters) because the shadow detail will be more visible. This happens because BPC remapped Level 1 of the image file to Level 3 for the print, and also remapped all the successive values up to Level 21 to be a bit brighter (see the blue bounding box around the curves in Figure 32), maintaining a good enough slope between Levels to separate the tones, as it smoothly trends back to the reference values by Level 21. The quality of the profile and the printer both account for this desirable outcome.
The story is similar in principle, but the numbers and the corresponding picture change quite dramatically with matte papers (example, Legacy Etching, Figures 33, 34).
In Figure 33, ABSCOL rendition, the graph shows that all tones (L*) in the image file from 1 to 13 inclusive will be about the same, such that all contiguous pixels having those values will form one undifferentiated blob even if the image file contains differentiated tones below L*13. Beyond Level 13, there is lift off and the onset of contrast (note the slope of the red curve). But it’s already getting along the Luminance scale by then, so RELCOL or Perceptual Rendering Intent comes to the rescue; but this time the tonal rearrangement is considerably more dramatic than what we saw in Figure 32 for PK paper. L* 1 of the image file gets remapped to L* 16 for the print (Figure 34), and the slope of the red curve moves gradually upward from L* 16, not converging with reference values till it reaches a level beyond this graph.
This allows RELCOL or Perceptual to differentiate shadow detail (albeit on a less dense scale) that exists in the image file, here of the 30 Neutral target. Based on viewing the Roman 16 low-key B&W image printed on the new Epson Legacy Fine Art paper in RELCOL or PERCEPTUAL it’s actually quite attractive. I believe Perceptual shows an ever so slightly crisper rendition of shadow detail versus RELCOL for this image, but I had to look rather intently to appreciate a difference.
The takeaway from this little demonstration is that the Rendering Intent is doing very much what it is supposed to do and the printer/profile combination is reproducing the tonal realignment in a very smooth, linear manner. It works well whether the profile is Custom or OEM. This performance is particularly important for matte papers where there is “heavy lifting” to be done with shadow detail.
Figures 35, 36 and 37 show screen grabs of the Roman-16 low-key B&W image for No Softproof, then ABSCOL and Perceptual soft proofs respectively using Epson Legacy Etching matte fine art paper. Comparing the three images mirrors what the graphs are telling us.
Notice in particular the loss of detail rendition on the black scarf and jacket in the ABSCOL rendition of Figure 36, compared with much more distinct tonal rendition of the non-soft proofed and Perceptual versions in Figures 35 and 37. This does indeed confirm the expectations from examining the curves of Figures 33 and 34.
The printer’s rendition of highlight detail is also excellent, as evident from examining prints of the Roman 16 high-key B&W image (Figures 38 and 39), here using my own profile, but the outcomes are extremely close whether using Custom or Epson profiles.
The important observations from these images are the smoothness of tonal gradation in the smooth areas of the image and the quality of detail rendition in the model’s dress. The RELCOL + BPC version shows ever so slightly more of this detail than the Perceptual rendition, but both are very, very close.
What about ABW Mode?
No story about B&W printing in a new Epson model would be complete without addressing ABW, so here we go. First I’ll explain (in broad outline) how ABW works; then we’ll look at how it performs in this printer compared with a standard “Application Manages Color” workflow.
As most readers would know, ABW is a printer-managed printing mode accessible through Epson Color Controls in the printer driver, after having set Photoshop’s Print dialog to “Printer Manages Color”.
The ABW mode sets a Black Point and a White point and the tonality between the two is mapped depending on different “curve” shapes determined by whether the user sets tonality to Light, Normal, Dark, Darker or Darkest. Think of it as five canned contrast curve presets. It does not depend on ICC profiles under the hood. The other distinguishing feature of ABW is in its use of color inks. It does not use Cyan, Vivid Magenta, Orange or Green, but it does use Light Magenta, Light Cyan and Yellow for toning the print.
One other important factor to know is that ABW is not an available option for the Commercial Edition SC-P5000 printer equipped with Violet ink because Violet replaces LLK and ABW needs LLK to work. Epson provided the Commercial Edition with Violet ink for prepress and graphic design usage. For photography and those who wish to print with ABW mode, as mentioned earlier, the recommendation is to use the Standard Edition of the printer, as LLK is more important than Violet, whether to enable ABW or for general photographic tonal rendition.
The perennial question about ABW is whether it delivers superior quality B&W prints than does Application Color Management with ICC profiles. Regardless of how often I’ve reviewed this matter in the past, of course with a new printer model one is obliged to review it yet again – there could be surprises or at least nuances. To remind, by “better quality” we are looking for Blacker Blacks, superior neutrality, smoother tonal gradations, and more distinct highlight detail. We can examine these questions with data and real prints; I do both here.
First with the data: Figure 40 reports on comparative Blacks for four papers and three conditions: ABW, OEM profiles, and custom profiles.
ABW has a slightly deeper Black point for the first two papers but not the second two. In all cases, the differences between OEM/ABW/Mark would be invisible to most observers. As well, the ABW Black points are not systematically and unambiguously more neutral than those generated from the ICC profiles.
Turning back to the data for the 21 Neutral test reported in Figures 25 and 26, the overall accuracy and neutrality of tonal reproduction comparing ABW with ICC color management for selected representative papers appear as follows:
- Epson Legacy Platine: OEM profile is overall more accurate than ABW; the Custom profile is overall a little less accurate; differences of neutrality are insignificant.
- Epson Legacy Etching: Differences are insignificant.
The Roman 16 prints that I made of both the low-key and high-key images indicate no striking visual advantage or disadvantage to ABW over an ICC workflow, whether for the latter using OEM or Custom profiles. For the ABW workflow, I used “Neutral”, “Dark” settings in the ABW driver. “Darker” is also usable but a bit heavier.
All of this evidence brings me back to my previous observations on this subject: these printers work so well in their normal ICC configuration that it’s largely a matter of personal taste whether one wants to use ABW for the slightly darker Black it can achieve, if you can see it; but by using it, one sacrifices the infinite degrees of control over tonality, aided by image-specific soft proofing, that one has with an ICC-managed workflow. That said, the SC-P5000 makes very fine B&W prints whether using ABW or standard Application Managed Color. Another factor about ABW that readers may find useful: Epson advised that the lightfastness testing for Black and White prints was done using the ABW Mode, which omits Vivid Magenta and Cyan. Epson has not conducted print permanence testing for black-and-white prints made by using ICC profiles.
While on the subject of print permanence, which marks an appropriate conclusion to this review, Epson states:
<Fade Resistance / Print Longevity: Color: up to 200 years; black-and-white: up to 400 years1
<1 Ink lightfastness rating based on accelerated testing of prints on specialty media, displayed indoors, under glass. Actual print stability will vary according to media, printed image, display conditions, light intensity, humidity and atmospheric conditions. Epson does not guarantee the longevity of prints. For maximum print life, display all prints under glass or lamination or properly store them. Visit www.wilhelm-research.com for the latest information.>
On the WIR website, navigation of which is best done using Search terms, there is an updated notice stating:
<Final print permanence test results for the new HDX inks printed on a variety of media will be posted on the WIR website as soon as the data become available toward the end of this year. It is one of the realities of the print permanence testing field that the more stable the print, the longer the tests take to complete. >
There is also a 3-page release under the Epson logo, dated October 22, 2015, providing further information, which has been copy-protected, making it inefficient for me to extract key passages. It is not a full test results report in the usual Wilhelm format; however, I recommend that interested readers visit the Wilhelm site, enter a search term called “HD” and navigate to the PDF.
Another highly respected print permanence authority, Aardenburg Imaging has two inkjet papers under test for the Epson SC-P600 printer, also using the Epson HD ink set (but without Orange and Green). I also recommend readers to visit the Aardenburg site, where you will find much very useful information on the subject.