Epson Legacy Baryta II Paper
Two Papers, Two Printers
Readers will recall the unfortunate demise of a number of inkjet papers from several paper companies, all having the common characteristics of being luster-like paper with a baryta component. Epson is one such company, which sought to replace Legacy Baryta with another similar paper. Epson does not manufacture inkjet paper. Its papers are manufactured in external paper and coating plants to Epson specifications. In late 2019, Epson announced a replacement paper to be called Legacy Baryta II. Very small amounts of it became available late in 2020 and then it disappeared, a victim of the COVID pandemic and related supply problems. In the intervening two years between then and now, Epson has been satisfying the market niche for this paper with Legacy Platine, a great, similar paper I have been using extensively.
The main differences between the discontinued Legacy Baryta product and Legacy Platine are that the former had an alpha-cellulose base and modest OBA content, while the latter has a cotton fiber base and no OBA content (OBA: Optical Brightening Agents). Hence, Legacy Platine arguably has greater permanence than the original Baryta product. Where both the Baryta II and Platine papers have a cotton fiber base and no OBA content, the immediate question is “what’s the difference” using the one or the other. Surface texture differs moderately between them, but the difference is rather subtle. For all those who weaned themselves from the original Baryta product to Legacy Platine, the question of what the new Legacy Baryta II brings to the table should be a matter of interest, explored in this article. Spoiler alert: it may just boil down to a matter of taste, but I’ll show comparisons of the technical details, regardless.
As well, being a wide-gamut PK paper, comparing how it performs in an Epson SC-P5000 printer versus an SC-P900 printer is of interest, insofar as the former has Orange and Green inks which, inter alia, provide it with the widest gamut of any desktop inkjet printer on the market, while the latter is a very popular professional printer for all those who don’t need an SC-P5000. The gamut achievable with the SC-P900 is also fully satisfactory for most photographs that most of us are likely to print. Hence, the “menu” for this article is “two papers, two printers”, the new element being Legacy Baryta II.
I’ll begin with “The Bottom Line”, and the rest of this review will show you how I got here:
The Bottom Line
When all is said and done, if you’re asking me which Legacy paper between Baryta II and Platine is preferable, I really have no answer because the dE numbers discussed further on are all so close that from an objective perspective there is little to seriously distinguish them; and, from a subjective visual perception, you’ll see in Figures 20~24 how similar the image appearances are. However, there are two physical differences between these papers that may tilt your choice one way or another: (i) the Baryta II product is slightly stiffer, which is good for protecting it from slightly rough handling, but not so good for feeding into an SC-P900 printer top feed; and (ii) the surface texture differs a bit between these papers. The difference is subtle enough that I can’t show it to you over the Internet, but it’s there. I would say it gives the Baryta II product a slight edge in apparent “brilliance”, as it’s slightly “beadier”, making some difference to reflectivity and surface gloss; however, the visible difference, I must emphasize, is rather fine. Both of these papers rank amongst the best available for archival pigment printing in inkjet printers, and I would recommend interested readers to try samples of each in order to determine a preference. I would happily print with either.
So, let us revert to the beginning, with the description of the papers, from the Epson website:
Evident from Figure 0, the most important differences are whiteness and brightness, the Baryta II product being both whiter and brighter than the Platine – though seen side by side under 4700 K Solux lighting, these differences look rather marginal. (Note: Hereafter I may drop the word “Legacy” from the paper names for sake of brevity.)
The difference of weight and thickness is numerically very little; however, in handling both papers, I find that the Baryta II seems somewhat heavier and less pliable than the Platine. I think this is advantageous, because it is perhaps less prone to damage from minor mishandling. One major characteristic of both papers I find important, especially when used in an SC-P5000 printer, is that the cotton fiber does not shed fine particulates capable of clogging the print-head. One downside of the Baryta II paper in the SC-P900 printer is that if using the top feed mechanism (because thickness is below the 0.5mm threshold between using the top feed versus the front feed), I found it necessary to help the paper engage the mechanism by holding both sides of the sheet evenly and pressing down on it till the feed engages. I hope that doing this doesn’t damage the paper feed.
Let us now turn to statistical performance, followed by an evaluation of photographs.
As I haven’t published a paper review for quite a while (because little new of interest – to me – has appeared on the market during the interval), I begin with a quick run-through of my testing procedure. While there are readers who find such technical detail boring and abstract, on the contrary I find that objective data (taste or opinion doesn’t intervene) is both interesting and a reliable predictor of print quality appearance in “real-world” photographs. Feel free to either persevere or skip to the more visual parts of the story.
The procedure works as follows:
- Profiling: I make and test my own custom profiles, in this case using Konica-Minolta’s MYIRO-1 and MYIRO Tools software. I have found the MYIRO usually provides somewhat more accurate results [measured by dE(00) of test versus reference colour values) with smaller target charts compared to the X-Rite i1Pro 3, though both are state-of-the-art performers. For custom profiling, I print the profiling target charts using the Adobe Color Print Utility, which eliminates colour management as necessary for characterising how the printer “natively” renders target colour values sent to it absent colour management. The profiles I make are calibrated to the M1 measurement condition (“M” factor explainer: https://www.xrite.com/-/media/xrite/files/literature/l7/l7-500_l7-599/l7-510_m_factor_what_does_it_mean/l7-510_m_factor_en.pdf).
I also test the manufacturer profiles (hereafter “OEM profiles”) because many people depend on these, given that unlike in the early days of inkjet printing, many OEM profiles, including those from Epson America, have become very good. At the time of writing there is no Legacy Baryta II profile for the SC-P5000 printer, but there is one for the SC-P900 printer. When Epson produces an OEM profile, I’ll analyze it and publish my findings.
- Profile Testing: Firstly, I examine the profile appearance and gamut data using the Chromix application “ColorThink Pro”. One looks for calculated gamut size and the smoothness of the gamut surface. For example, in Figure 1, the picture on the left shows a dismally failed profile, while the one on the right shows the upper portion of a well-formed, successful profile. (Yes, the one on the left was likely pilot error; it’s been trashed.)
Second, I developed a set of evaluation targets to assess the accuracy of colour rendition between the reference values of the patches in the test image file versus their measured results printed on paper. There are three tranches of this testing:
(i) a 48 patch colour and gray scale chart, in which 14 patches are gray scale and the remaining 34 multi-colour covering various shades of R,G,B,C,M,Y,K,V,O; the starting point of colour choices for the test patches came from the Barbieri control strip, but I changed and reformatted it to the needs of this testing for inkjet printing – in particular to have the colour values in gamut separately for PK and MK papers, and in whole integers. To serve the intended purpose, such charts must be printed in Absolute Colorimetric Rendering Intent (explainer: https://www.colorwiki.com/wiki/Rendering_Intent).
(ii) a 100 step grayscale (L*0 to L*100) measuring accuracy of rendition at intervals of L*1 – useful for assessing in detail the smoothness of tonal gradation from Black to White.
Only in-gamut colours are included in this testing because it is not intended to determine how well application-managed rendering intents handle out-of-gamut colours. Only the accuracy of the printing system is of interest here, and out of gamut colours will inherently reproduce inaccurately no matter the paper, profile or rendering intent.
(iii) Prints (using Relative or Perceptual Rendering Intent) of printer evaluation test pages consisting of photographs – not patches – specifically designed to illustrate various aspects of print quality, or lack thereof, including, for example, colour rendition, smoothness of tonal gradation, rendition of tone separation in shadows and highlights, rendition of detail. I find the most useful ones to be the Atkinson Printer Test Page (or the Outback version of it), the Onsight Version 9 test page and the shadow and highlight gray scale test photos from the Roman-16 package produced by the BVDM, Germany.
Below are several illustrations of some points made above:
Re Figure 4, the dots are within the profiles’ gamuts, except for the extremes of the gray scale. The P5000 gamut is wider mainly in the bright orange/yellow/green portions of the colour spectrum; a minority of photos I’ve tested require that additional gamut.
The structure of the following discussion is to proceed by printer, and within a printer section, by profile origin (Custom vs OEM) and by paper (Baryta II or Platine).
Epson SC-P5000 Printer (Custom Profile Only; OEM Profile Not Available):
Figure 5 compares printing system performance using custom profiles for both Epson Legacy Baryta II and Platine papers.
Legacy Platine exhibits about 5% more gamut volume than does Baryta II (row 3, columns C and D); however at these extremely wide gamut values, there are probably few important colours in real photos that would be in gamut for the one but not for the other.
Read on paper, Platine has a slightly brighter white point than Baryta II, which gives Platine a slight heads-up in this respect for those photos where separation of highlight detail in this extreme upper portion of the tone scale would matter. It is easy enough to discern the difference of tone between levels 95 and 97, but beyond 97 not so much. White point neutrality is very good for both papers (Rows 4-6), as seen from a* and b* values being close to zero – zero being neutral gray.
Black points are convincingly Black and similar for both papers; in both, black point neutrality is very good (Rows 7-9).
The gamut boundaries for the RGB primaries (rows 10 – 18) are fairly close in both papers. It appears that the slightly brighter white point and more intense outer boundary of red are the factors giving Platine a slightly higher overall gamut. I stress that we’re dealing in numerical ranges that would show little to no visual difference between papers.
Turning to the test results for colour accuracy (rows 20-22), both papers are capable of supporting highly accurate rendition of in-gamut image file colour values when the profiles are very good and the printing system from colour management through to the printer software and hardware are working within intended tolerances. An average difference of dE(00) 0.09 between two papers would be totally invisible to human visual perception. The worst performing colour had a dE(00) value of 1.34 and 1.22 for Baryta II and Platine, respectively. The inaccuracy represented by these values are very close to the boundaries of human visual perception – in other words, they hardly matter in practice.
(Explainer of dE: https://www.colorwiki.com/wiki/Delta_E:_The_Color_Difference)
Turning to the gray scale data (rows 23 onward), differences between reference and printed values are measured as simple absolute values, not dE(00). Values darker than the paper Black point and lighter than the paper White point are excluded from calculation of average dE(00) because they are out of gamut.
There’s no need to dwell too long on these results, because they are all stellar for both papers throughout the tone range. In particular, they demonstrate convincingly that the use of the Epson driver with application colour management is perfectly capable of delivering excellent B&W photographs in respect of tonal transitions and neutrality without the need for RIPs, special inks or special algorithms. I expect this statement may elicit howls of disagreement, but I find it hard to argue with these numbers; it just cannot get much better. One can drive a Black point down somewhat with alternative software and materials, but the difference between those and a 2.6 or 2.8 reported here are largely imperceptible on paper at this level of darkness. As well the extent of neutrality across the visible range is unlikely to be bettered to a visually significant degree.
The following graphics indicate the qualities portrayed in the foregoing data for the grayscale.
In the upper portion of Figure 6 dealing with tone, the Black line shows the file reference values of the gray scale, while the Red line plots the corresponding values read from the print of the gray scale patches. The closer the two lines converge, the more accurate the print of the luminance values. In the lower portion of Figure 6 dealing with hue, at a value of 0 (left-side vertical axis) the hue is completely neutral-gray. For each brightness level shown on the horizontal axis, the blue bars show departures from 0 of the blue-yellow colour range, while the red bars show departures from 0 of the green-magenta colour range. Projections above 0 portray a warm cast. Projections below 0 portray a cool cast. Any bars less than at least +/- 1.0 reflect imperceptible departures from strict neutrality.
You can readily see from the Figure 6 diagram that the accuracy of both luminance and hue neutrality rendition are excellent for the Baryta II paper/custom profile/printer combination.
The charts for the detailed 100-step gray scale analysis are interpreted identically to the above:
You can see from these detailed results the near-perfection of the luminance rendition and the imperceptible departures from strict hue neutrality for all in-gamut levels.
The result for the Platine paper is similar (Figure 8).
Turning to real-world photographs, while it’s a bit difficult to precisely replicate the look of a print on screen, I tried to do so using my Epson V-850 scanner and SilverFast Studio 8 to scan the actual prints and adjust them moderately for replicating their appearance as faithfully as I could (Figures 9, 10, 11, 12).
The key thing about properly examining these photos is that they contain memory colours and specific qualities of tonal gradation about which one should be somewhat knowledgeable in order to understand whether what one sees is an acceptable rendition of this subject matter. They are standard benchmark photos in the printing industries, where the colour management people know exactly what they are looking for when they see them. So, for example:
In the Atkinson target: many things, but principally the hue and saturation of the memory colours, the tonal separation and smoothness of tonal gradations, especially in the darker regions of the photos, the quality of skin tone rendition.
In the Onsight target: the handling of the tone scale in the deep blues of the lower right photo, the saturation and brightness of the orange-reds in the upper right photo, the quality of skin tone and hair sharpness in the photo of the young boy, the realism of red/green/blue rendition in the lower left photo, the neutrality and tonal rendition in the B&W photo, upper left.
In the Roman-16 gray-scale Dark photo: the visibility of detail in the clothing, the tonal separation of the hair and clothing from the background, the smoothness of the tone ramp constituting the background.
In the Roman-16 gray-scale Light photo: again the smoothness of the tone ramp constituting the background, this time for light tones, and the rendition of texture in the V-neck of the dress and the food in the foreground.
I find that the printed photos reflect expectations from the quality of the numbers generated by the objective measurements, such that whether printing colour or B&W, this paper combined with good profiling, colour management and printing system, will deliver fine quality prints.
SC-P900 Printer
OEM Profile
Consistently with my finding for the SC-P5000 printer, Platine supports a slightly wider gamut than does the Baryta II; however, both are below the gamut of the SC-P5000. That said, in practice I have found few photos needing more gamut than provided from the SC-P900 (see Figure 4 discussion above).
The Baryta II exhibits a very slightly shorter tonal range than the Platine (4.1 ~ 96.2 and 3.2 ~ 96.9 respectively), hardly a significant difference. The Black point of the SC-P900 can be driven a bit lower using the Carbon Black or ABW printing options (Ref. Figure 33 in my PhotoPXL review article of the SC-P900).
As this is an OEM profile not specific to the particular printer on which the test patches were printed, as usual, the accuracy will be slightly less than for a custom profile. The question of interest is whether the extent of this inaccuracy has real-world significance, and in this case for the most part it does not. The average dE(00) for the Baryta II is 1.7, and for the Platine 1.1. The worst colours have dE(00) values below 3, which while not ideal, is not usually a disturbing level of inaccuracy when viewed in real-world prints.
Gray scale accuracy with the P900 printer for both papers is good in respect of both luminance and hue neutrality (Figure 13 rows 23~34). The gray scale graphs for Legacy Baryta II, OEM profile follow:
The two figures are consistent in showing that for much of the tone scale, the test values came out just a bit brighter than the reference values, however based on this data one should not expect visible discontinuities in tonal transitions. The hue bias is toward the warm side, with many values clustered near or below 1.0, again indicating generally satisfactory hue neutrality in terms of what people would actually perceive in print.
SC-P900, Custom Profile
While the quality of the Epson profile is very good, as indicated just above, I nonetheless like to see whether it gets any better by making a custom profile which is specific to my printer. The custom profiling approach is the same for this printer as that described for the SC-P5000 above: use the Adobe Color Print Utility and the Legacy Baryta II Media Type for printing the profiling targets, again using the MYIRO 840-patch chart. Once created, the new profile is tested with four charts of patches with known reference values and four prints of real photographs as described above. The data includes a comparison with the Epson Legacy Platine for the SC-P900 printer as well.
Here again, as for the OEM profile, Platine has a slightly wider gamut volume than Baryta II. As well, Platine has a slightly broader tonal range, from 3.1~97.5 versus 4.5~95.8 respectively. These differences are not significant. In both cases, neutrality is well preserved. Profiling accuracy is very good for both papers: average dE(00) of 0.61 and 0.87 respectively. The small difference in favour of Platine is not significant. For the detailed gray scale, both papers exhibit excellent adherence to reference values and tonal neutrality, the one small difference being the accuracy of the gray scale in the mid-tones for the Baryta II paper, where the average difference from reference values is 1.09. Normally one would prefer this difference to be less than 1.0, but in this case, the printed tones within this range are about 1 L* value darker than the reference. Figure 17 shows what this would look like seen side-by-side:
As one may observe, the difference is really subtle, and this kind of portrayal maximizes the possibility of noticing it. The gray scale tonal transitions from one level to the next remain very smooth and continuous for both papers. The gray scale graphs for Legacy Baryta II, Custom profile, follow.
The mid-tone portion of Figure 19 shows the slight downward bowing of brightness rendition part-way up the mid-tones. Otherwise, adherence of brightness rendition through the in-gamut part of the tone scale is excellent. Hue neutrality is quite well preserved insofar as none of the bars deviate from zero by more than about +/- 1.0.
Turning back to the world of real photographs, in Figures 20 ~24 I compare some results of the prints made on Legacy Baryta II. The most interesting outcome of these comparisons is that they are indistinguishable between printers and profiles (Custom or OEM). This is a very good outcome given the differences between the machines and profiles, showing that proper colour management and good quality materials are doing one of the things they are supposed to do: maintain consistent colour across devices.
Mark Segal
January 2023
Mark D Segal
Toronto, ON
Mark has been making photographs for the past seven decades and started adopting a digital workflow in 1999 first with scanning film, then going fully digital in 2004. He has worked with a considerable range of software, equipment, materials and techniques over the years, accumulated substantial experience as an author, educator and communicator in several fields, was a frequent contributor to the Luminous-Landscape website and now contributes frequently with in-depth articles on the PhotoPXL website. Mark has contributed over 75 articles to the two websites up to Q1-2024, with a particular emphasis on printers and papers, given his view that a photograph printed on paper remains the epitome of fine photography, as it has been from soon after the medium was invented and started gaining momentum in the 1830s/1840s. Mark developed a particular interest in film scanning and authored the ebook “Scanning Workflows with SilverFast 8, SilverFast HDR, Adobe Photoshop Lightroom and Adobe Photoshop” (please check our Store for availability). In his “other life” (the one that pays for the photography), Mark is a retiree from the World Bank Group and was a consultant in electric power development.
